diff --git a/.nojekyll b/.nojekyll new file mode 100644 index 0000000..e69de29 diff --git a/404.html b/404.html new file mode 100644 index 0000000..b686514 --- /dev/null +++ b/404.html @@ -0,0 +1,457 @@ + + + + + + + + + + + + + + + + + + + echoSMs + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ +
+
+ +
+ + + + + + +
+ + +
+ +
+ + + + + + +
+
+ + + +
+
+
+ + + + + + + +
+
+
+ + + + +
+
+ +

404 - Not found

+ +
+
+ + + +
+ + + +
+ + + +
+
+
+
+ + + + + + + + + + \ No newline at end of file diff --git a/api_reference/index.html b/api_reference/index.html new file mode 100644 index 0000000..eb757d1 --- /dev/null +++ b/api_reference/index.html @@ -0,0 +1,10892 @@ + + + + + + + + + + + + + + + + + + + + + + + + + API reference - echoSMs + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ + + + Skip to content + + +
+
+ +
+ + + + + + +
+ + +
+ +
+ + + + + + +
+
+ + + +
+
+
+ + + + + + + +
+
+
+ + + + +
+
+ + + + + + + + + + + + + + + +

API reference

+

This is the API reference for the echoSMs package.

+

Each type of model is contained in a separate Python class (with name ending in Model), but with common calling signatures across all model classes, as defined in ScatterModelBase. There are also classes to provide ready access to the benchmark models and reference model definitions. There are also utility functions.

+

ScatterModelBase

+ + +
+ + + + +
+

+ Bases: ABC

+ + +

Base class for a class that provides a scattering model.

+

All scattering models should inherit from this class, have a name that +ends with 'Model', and provide initialisation and calculate_ts_single() functions.

+ +

Initialise.

+ + +

Attributes:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescription
long_name + str + +
+

The long name of the model.

+
+
short_name + str + +
+

A short version of the model's long name, typically an acronym.

+
+
analytical_type + str + +
+

Whether the model implements an exact or an approximate model.

+
+
boundary_types + list[str] + +
+

The types of boundary conditions that the model provides, e.g., 'fixed rigid', +'pressure release', 'fluid filled'

+
+
shapes + list[str] + +
+

The target shapes that the model can represent.

+
+
max_ka + float + +
+

An approximate maximum ka value that will result in accurate target strength results. +Note that ka is often not the only parameter that determines the accuracy of the +model (e.g., aspect ratio and incident angle can also affect the accuracy).

+
+
no_expand_parameters + list[str] + +
+

The model parameters that are not expanded into Pandas DataFrame columns or +Xarray DataArray coordinates. They will instead end up as a dict in the DataFrame or +DataArray attrs attribute.

+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
@abc.abstractmethod
+def __init__(self):
+    """Initialise.
+
+    Attributes
+    ----------
+    long_name : str
+        The long name of the model.
+    short_name : str
+        A short version of the model's long name, typically an acronym.
+    analytical_type : str
+        Whether the model implements an ``exact`` or an ``approximate`` model.
+    boundary_types : list[str]
+        The types of boundary conditions that the model provides, e.g., 'fixed rigid',
+        'pressure release', 'fluid filled'
+    shapes : list[str]
+        The target shapes that the model can represent.
+    max_ka : float
+        An approximate maximum ka value that will result in accurate target strength results.
+        Note that ka is often not the only parameter that determines the accuracy of the
+        model (e.g., aspect ratio and incident angle can also affect the accuracy).
+    no_expand_parameters : list[str]
+        The model parameters that are not expanded into Pandas DataFrame columns or
+        Xarray DataArray coordinates. They will instead end up as a dict in the DataFrame or
+        DataArray `attrs` attribute.
+    """
+    self.long_name = ''
+    self.short_name = ''
+    self.analytical_type = ''
+    self.boundary_types = []
+    self.shapes = []
+    self.max_ka = np.nan
+    self.no_expand_parameters = []
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single() + + + abstractmethod + + +

+ + +
+ +

Calculate the TS for one parameter set.

+ +
+ Source code in src/echosms/scattermodelbase.py + 
227
+228
+229
@abc.abstractmethod
+def calculate_ts_single(self):
+    """Calculate the TS for one parameter set."""
+
+
+
+ +
+ +
+ + +

+ validate_parameters(p) + + + abstractmethod + + +

+ + +
+ +

Validate the model parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
p + dict + +
+

Dict containing the model parameters.

+
+ 		+ required +
+ + +

Raises:

+ + + + + + + + + + + + + + + + + +
TypeDescription
+ ValueError + +
+

If any of the model parameters are invalid.

+
+
+ KeyError + +
+

If any required model parameters are not present.

+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
160
+161
+162
+163
+164
+165
+166
+167
+168
+169
+170
+171
+172
+173
+174
+175
@abc.abstractmethod
+def validate_parameters(self, p):
+    """Validate the model parameters.
+
+    Parameters
+    ----------
+    p : dict
+        Dict containing the model parameters.
+
+    Raises
+    ------
+    ValueError
+        If any of the model parameters are invalid.
+    KeyError
+        If any required model parameters are not present.
+    """
+
+
+
+ +
+ + + +
+ +
+ +

DCMModel

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Modal series deformed cylinder model (DCM).

+

This class contains methods to calculate acoustic scatter from finite straight cylinders with +various boundary conditions.

+ +
+ Source code in src/echosms/dcmmodel.py + 
19
+20
+21
+22
+23
+24
+25
+26
def __init__(self):
+    super().__init__()
+    self.long_name = 'deformed cylinder model'
+    self.short_name = 'dcm'
+    self.analytical_type = 'approximate analytical'
+    self.boundary_types = ['fixed rigid', 'pressure release', 'fluid filled']
+    self.shapes = ['finite cylinder']
+    self.max_ka = 20  # [1]
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(medium_c, medium_rho, a, b, theta, f, boundary_type, target_c=None, target_rho=None, validate_parameters=True, **kwargs) + +

+ + +
+ +

Calculate the scatter from a finite cylinder using the modal series deformed cylinder model.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
medium_c + float + +
+

Sound speed in the fluid medium surrounding the target [m/s].

+
+ 		+ required +
medium_rho + float + +
+

Density of the fluid medium surrounding the target [kg/m³].

+
+ 		+ required +
a + float + +
+

Radius of the cylinderical target [m].

+
+ 		+ required +
b + float + +
+

Length of the cylinderical target [m].

+
+ 		+ required +
theta + float + +
+

Pitch angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
f + float + +
+

Frequency to calculate the scattering at [Hz].

+
+ 		+ required +
boundary_type + str + +
+

The model type. Supported model types are given in the boundary_types class attribute.

+
+ 		+ required +
target_c + float + +
+

Sound speed in the fluid inside the sphere [m/s]. +Only required for boundary_type of fluid filled.

+
+ 		+ None +
target_rho + float + +
+

Density of the fluid inside the sphere [kg/m³]. +Only required for boundary_type of fluid filled.

+
+ 		+ None +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) of the target [dB].

+
+
+ + +
+ Notes +

The class implements the code in Section B.1 of Jech et al. (2015).

+
+ +
+ References +

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., +Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. +Comparisons among ten models of acoustic backscattering used in aquatic ecosystem +research. Journal of the Acoustical Society of America 138, 3742–3764. +https://doi.org/10.1121/1.4937607

+
+
+ Source code in src/echosms/dcmmodel.py + 
 42
+ 43
+ 44
+ 45
+ 46
+ 47
+ 48
+ 49
+ 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
def calculate_ts_single(self, medium_c, medium_rho, a, b, theta, f, boundary_type,
+                        target_c=None, target_rho=None, validate_parameters=True,
+                        **kwargs):
+    """
+    Calculate the scatter from a finite cylinder using the modal series deformed cylinder model.
+
+    Parameters
+    ----------
+    medium_c : float
+        Sound speed in the fluid medium surrounding the target [m/s].
+    medium_rho : float
+        Density of the fluid medium surrounding the target [kg/m³].
+    a : float
+        Radius of the cylinderical target [m].
+    b : float
+        Length of the cylinderical target [m].
+    theta : float
+        Pitch angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    f : float
+        Frequency to calculate the scattering at [Hz].
+    boundary_type : str
+        The model type. Supported model types are given in the `boundary_types` class attribute.
+    target_c : float, optional
+        Sound speed in the fluid inside the sphere [m/s].
+        Only required for `boundary_type` of ``fluid filled``.
+    target_rho : float, optional
+        Density of the fluid inside the sphere [kg/m³].
+        Only required for `boundary_type` of ``fluid filled``.
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) of the target [dB].
+
+    Notes
+    -----
+    The class implements the code in Section B.1 of Jech et al. (2015).
+
+    References
+    ----------
+    Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B.,
+    Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015.
+    Comparisons among ten models of acoustic backscattering used in aquatic ecosystem
+    research. Journal of the Acoustical Society of America 138, 3742–3764.
+    <https://doi.org/10.1121/1.4937607>
+    """
+    if validate_parameters:
+        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'b': b, 'f': f,
+             'boundary_type': boundary_type, 'target_c': target_c, 'target_rho': target_rho,
+             'theta': theta}
+        self.validate_parameters(p)
+
+    if theta == 0.0:
+        return nan
+
+    theta_rad = theta*pi/180.
+    kL = wavenumber(medium_c, f)*b
+    K = wavenumber(medium_c, f) * sin(theta_rad)
+    Ka = K*a
+
+    m = range(30)  # TODO this needs to vary with f
+
+    match boundary_type:
+        case 'fixed rigid':
+            series = map(lambda m: (-1)**m * Neumann(m)*(jvp(m, Ka) / h1vp(m, Ka)), m)
+        case 'pressure release':
+            series = map(lambda m: (-1)**m * Neumann(m)*(jv(m, Ka) / hankel1(m, Ka)), m)
+        case 'fluid filled':
+            g = target_rho/medium_rho
+            h = target_c/medium_c
+            gh = g*h
+            Kda = K/h*a
+
+            def Cm(m):
+                numer = (jvp(m, Kda)*yv(m, Ka)) / (jv(m, Kda)*jvp(m, Ka))\
+                    - gh*(yvp(m, Ka)/jvp(m, Ka))
+                denom = (jvp(m, Kda)*jv(m, Ka)) / (jv(m, Kda)*jvp(m, Ka)) - gh
+                return numer/denom
+
+            series = map(lambda m: 1j**(2*m) * Neumann(m) / (1 + 1j*Cm(m)), m)
+        case _:
+            raise ValueError(f'The {self.long_name} model does not support '
+                             f'a model type of "{boundary_type}".')
+
+    fbs = 1j*b/pi * (sin(kL*cos(theta_rad)) / (kL*cos(theta_rad))) * sum(series)
+    return 20*log10(abs(fbs))  # ts
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/dcmmodel.py + 
28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+    super()._present_and_in(p, ['boundary_type'], self.boundary_types)
+    super()._present_and_positive(p, ['medium_rho', 'medium_c', 'a', 'b', 'f'])
+
+    for bt in np.atleast_1d(p['boundary_type']):
+        if bt == 'fluid filled':
+            super()._present_and_positive(p, ['target_c', 'target_rho'])
+
+
+
+ +
+ + + +
+ +
+ +

DWBA models

+

There are several models that use the distorted-wave Born approximation, documented below:

+

DWBA

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Distorted-wave Born approximation scattering model.

+ + +
+ Note +

The DWBA model is not yet functional.

+
+
+ Source code in src/echosms/dwbamodel.py + 
15
+16
+17
+18
+19
+20
+21
+22
def __init__(self):
+    super().__init__()
+    self.long_name = 'distorted-wave Born approximation'
+    self.short_name = 'dwba'
+    self.analytical_type = 'approximate'
+    self.boundary_types = ['weakly scattering']
+    self.shapes = ['any']
+    self.max_ka = 20
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(theta, phi, f, target_rho, target_c, validate_parameters=True) + +

+ + +
+ +

Distorted-wave Born approximation scattering model.

+

Implements the distorted-wave Born approximation +model for calculating the acoustic backscatter from weakly scattering bodies.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
theta + float + +
+

Pitch angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
phi + float + +
+

Roll angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
f + float + +
+

Frequency to run the model at [Hz]

+
+ 		+ required +
target_rho + iterable[float] + +
+

Densities of each material. Must have at least the same number of entries as unique +integers in volume [kg/m³].

+
+ 		+ required +
target_c + iterable[float] + +
+

Sound speed of each material. Must have at least the same number of entries as unique +integers in volume [m/s].

+
+ 		+ required +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) [dB] of the target.

+
+
+ + +
+ Notes +

This class implements the method presented in Chu et al. (1993).

+
+ +
+ References +

Chu, D., Foote, K. G., & Stanton, T. K. (1993). Further analysis of target strength +measurements of Antarctic krill at 38 and 120 kHz: Comparison with deformed cylinder +model and inference or orientation distribution. The Journal of the Acoustical Society +of America, 93(5), 2985–2988. https://doi.org/10.1121/1.405818

+
+
+ Source code in src/echosms/dwbamodel.py + 
32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
+68
+69
+70
+71
+72
+73
+74
+75
+76
+77
+78
+79
def calculate_ts_single(self, theta, phi, f, target_rho, target_c, validate_parameters=True):
+    """Distorted-wave Born approximation scattering model.
+
+    Implements the distorted-wave Born approximation
+    model for calculating the acoustic backscatter from weakly scattering bodies.
+
+    Parameters
+    ----------
+    theta : float
+        Pitch angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    phi : float
+        Roll angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    f : float
+        Frequency to run the model at [Hz]
+    target_rho : iterable[float]
+        Densities of each material. Must have at least the same number of entries as unique
+        integers in `volume` [kg/m³].
+    target_c : iterable[float]
+        Sound speed of each material. Must have at least the same number of entries as unique
+        integers in `volume` [m/s].
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) [dB] of the target.
+
+    Notes
+    -----
+    This class implements the method presented in Chu et al. (1993).
+
+    References
+    ----------
+    Chu, D., Foote, K. G., & Stanton, T. K. (1993). Further analysis of target strength
+    measurements of Antarctic krill at 38 and 120 kHz: Comparison with deformed cylinder
+    model and inference or orientation distribution. The Journal of the Acoustical Society
+    of America, 93(5), 2985–2988. <https://doi.org/10.1121/1.405818>
+
+    """
+    if validate_parameters:
+        p = {'theta': theta, 'phi': phi, 'f': f, 'target_rho': f, 'target_c': target_c}
+        self.validate_parameters(p)
+    return None
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/dwbamodel.py + 
24
+25
+26
+27
+28
+29
+30
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+
+
+
+ +
+ + + +
+ +
+ +

PT-DWBA

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Phase-tracking distorted-wave Born approximation scattering model.

+ +
+ Source code in src/echosms/ptdwbamodel.py + 
13
+14
+15
+16
+17
+18
+19
+20
+21
def __init__(self):
+    super().__init__()
+    self.long_name = 'phase-tracking distorted-wave Born approximation'
+    self.short_name = 'pt-dwba'
+    self.analytical_type = 'approximate'
+    self.boundary_types = ['weakly scattering']
+    self.shapes = ['unrestricted voxel-based']
+    self.max_ka = 20
+    self.no_expand_parameters = ['volume', 'voxel_size', 'rho', 'c']
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(volume, theta, phi, f, voxel_size, rho, c, validate_parameters=True, **kwargs) + +

+ + +
+ +

Phase-tracking distorted-wave Born approximation scattering model.

+

Implements the phase-tracking distorted-wave Born approximation +model for calculating the acoustic backscatter from weakly scattering bodies.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
volume + Numpy ndarray[int] + +
+

The object to be modelled as a 3D volume of voxels. Array contents should be 0 +for the surrounding medium, then increasing by 1 for each additional material +type (i.e., 1, 2, 3, etc). volume should be arranged as per the echoSMs +coordinate system, where

+
    +
  • axis 0 (rows) is the x-axis
    • +
  • axis 1 (columns) is the y-axis
    • +
  • axis 2: (slices) is the z-axis
    • +
+

Increasing axes indices correspond to increasing x, y, and z values.

+
+ 		+ required +
theta + float + +
+

Pitch angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
phi + float + +
+

Roll angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
f + float + +
+

Frequency to run the model at [Hz]

+
+ 		+ required +
voxel_size + iterable[float] + +
+

The size of the voxels in volume [m], ordered (x, y, z). +This code assumes that the voxels are cubes so y and z are currently irrelevant.

+
+ 		+ required +
rho + iterable[float] + +
+

Densities of each material. Must have at least the same number of entries as unique +integers in volume [kg/m³].

+
+ 		+ required +
c + iterable[float] + +
+

Sound speed of each material. Must have at least the same number of entries as unique +integers in volume [m/s].

+
+ 		+ required +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) [dB] of the target.

+
+
+ + +
+ Notes +

This class implements the method presented in Jones et. al. (2009). The code is +based closely on the Matlab code in Jones (2006).

+
+ +
+ References +

Jones, B. A. (2006). Acoustic scattering of broadband echolocation signals +from prey of Blainville's beaked whales: Modeling and analysis. Master of Science, +Massachusetts Institute of Technology. https://doi.org/10.1575/1912/1283

+

Jones, B. A., Lavery, A. C., & Stanton, T. K. (2009). Use of the distorted +wave Born approximation to predict scattering by inhomogeneous objects: +Application to squid. The Journal of the Acoustical Society of America, +125(1), 73-88. https://doi.org/10.1121/1.3021298

+
+
+ Source code in src/echosms/ptdwbamodel.py + 
 31
+ 32
+ 33
+ 34
+ 35
+ 36
+ 37
+ 38
+ 39
+ 40
+ 41
+ 42
+ 43
+ 44
+ 45
+ 46
+ 47
+ 48
+ 49
+ 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
+153
+154
+155
+156
+157
+158
+159
+160
+161
+162
+163
+164
+165
+166
+167
+168
+169
+170
+171
+172
+173
+174
+175
+176
+177
+178
+179
+180
+181
+182
+183
+184
+185
+186
+187
+188
+189
+190
+191
+192
+193
+194
+195
def calculate_ts_single(self, volume, theta, phi, f, voxel_size, rho, c,
+                        validate_parameters=True, **kwargs):
+    """Phase-tracking distorted-wave Born approximation scattering model.
+
+    Implements the phase-tracking distorted-wave Born approximation
+    model for calculating the acoustic backscatter from weakly scattering bodies.
+
+    Parameters
+    ----------
+    volume : Numpy ndarray[int]
+        The object to be modelled as a 3D volume of voxels. Array contents should be 0
+        for the surrounding medium, then increasing by 1 for each additional material
+        type (i.e., 1, 2, 3, etc). `volume` should be arranged as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems), where
+
+        - axis 0 (rows) is the _x_-axis
+        - axis 1 (columns) is the _y_-axis
+        - axis 2: (slices) is the _z_-axis
+
+        Increasing axes indices correspond to increasing _x_, _y_, and _z_ values.
+
+    theta : float
+        Pitch angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+
+    phi : float
+        Roll angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+
+    f : float
+        Frequency to run the model at [Hz]
+
+    voxel_size : iterable[float]
+        The size of the voxels in `volume` [m], ordered (_x_, _y_, _z_).
+        This code assumes that the voxels are cubes so _y_ and _z_ are currently irrelevant.
+
+    rho : iterable[float]
+        Densities of each material. Must have at least the same number of entries as unique
+        integers in `volume` [kg/m³].
+
+    c : iterable[float]
+        Sound speed of each material. Must have at least the same number of entries as unique
+        integers in `volume` [m/s].
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) [dB] of the target.
+
+    Notes
+    -----
+    This class implements the method presented in Jones et. al. (2009). The code is
+    based closely on the Matlab code in Jones (2006).
+
+    References
+    ----------
+    Jones, B. A. (2006). Acoustic scattering of broadband echolocation signals
+    from prey of Blainville's beaked whales: Modeling and analysis. Master of Science,
+    Massachusetts Institute of Technology. <https://doi.org/10.1575/1912/1283>
+
+    Jones, B. A., Lavery, A. C., & Stanton, T. K. (2009). Use of the distorted
+    wave Born approximation to predict scattering by inhomogeneous objects:
+    Application to squid. The Journal of the Acoustical Society of America,
+    125(1), 73-88. <https://doi.org/10.1121/1.3021298>
+    """
+    if validate_parameters:
+        p = {'volume': volume, 'theta': theta, 'phi': phi, 'f': f,
+             'voxel_size': voxel_size, 'rho': rho, 'c': c}
+        self.validate_parameters(p)
+
+    # Make sure things are numpy arrays
+    rho = np.atleast_1d(rho)
+    c = np.atleast_1d(c)
+    voxel_size = np.array(voxel_size)
+
+    # volume of the voxels [m^3]
+    dv = voxel_size.prod()
+
+    # input parameter checks
+    if not len(volume.shape) == 3:
+        raise TypeError('The volume input variable must be 3-dimensional.')
+
+    if not voxel_size.shape[0] == 3:
+        raise TypeError('The voxel_size input variable must contain 3 items.')
+
+    if not np.any(voxel_size > 0):
+        raise ValueError('All voxel_size values must be positive.')
+
+    if f < 0.0:
+        raise ValueError('The f input variable must contain only positive values.')
+
+    if (theta < -0.0) or (theta > 180.0):
+        raise ValueError('The theta (pitch) angle must be between -180.0 and +180.0')
+
+    if (phi < -180.0) or (phi > 180.0):
+        raise ValueError('The phi (roll) angle must be between -180.0 and +180.0')
+
+    if volume.min() != 0:
+        raise ValueError('The volume input variable must contain zeros.')
+
+    categories = np.unique(volume)
+    if not len(categories == (volume.max() + 1)):
+        raise ValueError('The integers in volume must include all values in the series '
+                         '(0, 1, 2, ..., n), where n is the largest integer in volume.')
+
+    if not len(rho) >= len(categories):
+        raise ValueError('The target_rho variable must contain at least as many values as '
+                         'unique integers in the volume variable.')
+
+    if not len(c) >= len(categories):
+        raise ValueError('The target_c variable must contain at least as many values '
+                         'as unique integers in the volume variable.')
+
+    # density and sound speed ratios for all object materials
+    g = rho[1:] / rho[0]
+    h = c[1:] / c[0]
+
+    # Do the pitch and roll rotations
+
+    # Convert echoSMs rotation angles (which are intrinsic) into extrinsic as
+    # that is what ndimage.rotate() below uses.
+    if phi == 0.0:  # short circuit the coordinate transformation if we can
+        pitch = theta-90
+        roll = 0.0
+    else:
+        rot = R.from_euler('ZYX', (0, theta-90, -phi), degrees=True)
+        # for backscatter we don't care about yaw
+        _, pitch, roll = rot.as_euler('zyz', degrees=True)
+
+    v = ndimage.rotate(volume, pitch, axes=(0, 2), order=0)
+    v = ndimage.rotate(v, roll, axes=(1, 2), order=0)
+
+    categories = np.unique(v)  # or just take the max?
+
+    # wavenumbers in the various media
+    k = 2.0*np.pi * f / c
+
+    # DWBA coefficients
+    # amplitudes in media 1,2,...,n
+    Cb = 1.0/(g * h**2) + 1.0/g - 2.0  # gamma_kappa - gamma_rho
+    Ca = k[0]**2 * Cb / (4.0*np.pi)  # summation coefficient
+
+    # Differential phase for each voxel.
+    dph = np.zeros(v.shape)
+    masks = []
+    for i, category in enumerate(categories):
+        masks.append(np.isin(v, category))
+        dph[masks[i]] = k[i] * voxel_size[0]
+    masks.pop(0)  # don't need to keep the category[0] mask
+
+    # cumulative summation of phase along the z-direction
+    phase = dph.cumsum(axis=2) - dph/2.0
+    dA = np.zeros(phase.shape, dtype=np.complex128)
+
+    # differential phases for each voxel
+    for i, m in enumerate(masks):
+        dA[m] = Ca[i] * np.exp(2.0*1j*phase[m]) * dv
+
+    # Convert to TS
+    return 20.0 * np.log10(np.abs(dA.sum()))
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/ptdwbamodel.py + 
23
+24
+25
+26
+27
+28
+29
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+
+
+
+ +
+ + + +
+ +
+ +

SDWBA

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Stochastic distorted-wave Born approximation scattering model.

+ + +
+ Note +

The SDWBA model is not yet functional.

+
+
+ Source code in src/echosms/sdwbamodel.py + 
15
+16
+17
+18
+19
+20
+21
+22
def __init__(self):
+    super().__init__()
+    self.long_name = "stochastic distorted-wave Born approximation"
+    self.short_name = "sdwba"
+    self.analytical_type = "approximate"
+    self.boundary_types = ["weakly scattering"]
+    self.shapes = ["any"]
+    self.max_ka = 20
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(theta, phi, f, target_rho, target_c, validate_parameters=True) + +

+ + +
+ +

Stochastic distorted-wave Born approximation scattering model.

+

Implements the stochastic distorted-wave Born approximation +model for calculating the acoustic backscatter from weakly scattering bodies.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
theta + float + +
+

Pitch angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
phi + float + +
+

Roll angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
f + float + +
+

Frequency to run the model at [Hz]

+
+ 		+ required +
target_rho + iterable[float] + +
+

Densities of each material. Must have at least the same number of entries as unique +integers in volume [kg/m³].

+
+ 		+ required +
target_c + iterable[float] + +
+

Sound speed of each material. Must have at least the same number of entries as unique +integers in volume [m/s].

+
+ 		+ required +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) [dB] of the target.

+
+
+ + +
+ Notes +

This class implements the method presented in Demer & Conti (2003), Demer & Conti (2004), +and Conti & Demer (2006).

+
+ +
+ References +

Demer, D. A., & Conti, S. G. (2003). Reconciling theoretical versus empirical target +strengths of krill: Effects of phase variability on the distorted-wave Born approximation. +ICES Journal of Marine Science, 60, 429-434. +https://doi.org/10.1016/S1054-3139(03)00002-X

+

Demer, D. A., & Conti, S. G. (2004). Reconciling theoretical versus empirical +target strengths of krill: Effects of phase variability on the distorted-wave Born +approximation. ICES Journal of Marine Science, 61(1), 157-158. +https://doi.org/10.1016/j.icesjms.2003.12.003

+

Conti, S. G., & Demer, D. A. (2006). Improved parameterization of the SDWBA for estimating +krill target strength. ICES Journal of Marine Science, 63(5), 928-935. +https://doi.org/10.1016/j.icesjms.2006.02.007

+
+
+ Source code in src/echosms/sdwbamodel.py + 
32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
+68
+69
+70
+71
+72
+73
+74
+75
+76
+77
+78
+79
+80
+81
+82
+83
+84
+85
+86
+87
+88
+89
def calculate_ts_single(self, theta, phi, f, target_rho, target_c, validate_parameters=True):
+    """Stochastic distorted-wave Born approximation scattering model.
+
+    Implements the stochastic distorted-wave Born approximation
+    model for calculating the acoustic backscatter from weakly scattering bodies.
+
+    Parameters
+    ----------
+    theta : float
+        Pitch angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    phi : float
+        Roll angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    f : float
+        Frequency to run the model at [Hz]
+    target_rho : iterable[float]
+        Densities of each material. Must have at least the same number of entries as unique
+        integers in `volume` [kg/m³].
+    target_c : iterable[float]
+        Sound speed of each material. Must have at least the same number of entries as unique
+        integers in `volume` [m/s].
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) [dB] of the target.
+
+    Notes
+    -----
+    This class implements the method presented in Demer & Conti (2003), Demer & Conti (2004),
+    and Conti & Demer (2006).
+
+    References
+    ----------
+    Demer, D. A., & Conti, S. G. (2003). Reconciling theoretical versus empirical target
+    strengths of krill: Effects of phase variability on the distorted-wave Born approximation.
+    ICES Journal of Marine Science, 60, 429-434.
+    <https://doi.org/10.1016/S1054-3139(03)00002-X>
+
+    Demer, D. A., & Conti, S. G. (2004). Reconciling theoretical versus empirical
+    target strengths of krill: Effects of phase variability on the distorted-wave Born
+    approximation. ICES Journal of Marine Science, 61(1), 157-158.
+    <https://doi.org/10.1016/j.icesjms.2003.12.003>
+
+    Conti, S. G., & Demer, D. A. (2006). Improved parameterization of the SDWBA for estimating
+    krill target strength. ICES Journal of Marine Science, 63(5), 928-935.
+    <https://doi.org/10.1016/j.icesjms.2006.02.007>
+    """
+    if validate_parameters:
+        p = {'theta': theta, 'phi': phi, 'f': f, 'target_rho': f, 'target_c': target_c}
+        self.validate_parameters(p)
+
+    return None
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/sdwbamodel.py + 
24
+25
+26
+27
+28
+29
+30
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+
+
+
+ +
+ + + +
+ +
+ +

ESModel

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Elastic sphere (ES) scattering model.

+

This class calculates acoustic backscatter from elastic spheres.

+ +
+ Source code in src/echosms/esmodel.py + 
17
+18
+19
+20
+21
+22
+23
+24
def __init__(self):
+    super().__init__()
+    self.long_name = 'elastic sphere'
+    self.short_name = 'es'
+    self.analytical_type = 'exact'
+    self.boundary_types = ['elastic']
+    self.shapes = ['sphere']
+    self.max_ka = 20  # [1]
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(medium_c, medium_rho, a, f, target_longitudinal_c, target_transverse_c, target_rho, validate_parameters=True, **kwargs) + +

+ + +
+ +

Calculate the backscatter from an elastic sphere for one set of parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
medium_c + float + +
+

Sound speed in the fluid medium surrounding the sphere [m/s].

+
+ 		+ required +
medium_rho + float + +
+

Density of the fluid medium surrounding the sphere [kg/m³].

+
+ 		+ required +
a + float + +
+

Radius of the sphere [m].

+
+ 		+ required +
f + float + +
+

Frequency to calculate the scattering at [Hz].

+
+ 		+ required +
target_longitudinal_c + float + +
+

Longitudinal sound speed in the material inside the sphere [m/s].

+
+ 		+ required +
target_transverse_c + float + +
+

Transverse sound speed in the material inside the sphere [m/s].

+
+ 		+ required +
target_rho + float + +
+

Density of the material inside the sphere [kg/m³].

+
+ 		+ required +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) of the sphere [dB].

+
+
+ + +
+ Notes +

The class implements the code in MacLennan (1981).

+
+ +
+ References +

MacLennan, D. N. (1981). The Theory of Solid Spheres as Sonar Calibration Targets. +Scottish Fisheries Research Report Number 22. Department of Agriculture and Fisheries +for Scotland.

+
+
+ Source code in src/echosms/esmodel.py + 
 38
+ 39
+ 40
+ 41
+ 42
+ 43
+ 44
+ 45
+ 46
+ 47
+ 48
+ 49
+ 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
def calculate_ts_single(self, medium_c, medium_rho, a, f,
+                        target_longitudinal_c, target_transverse_c, target_rho,
+                        validate_parameters=True,
+                        **kwargs) -> float:
+    """
+    Calculate the backscatter from an elastic sphere for one set of parameters.
+
+    Parameters
+    ----------
+    medium_c : float
+        Sound speed in the fluid medium surrounding the sphere [m/s].
+    medium_rho : float
+        Density of the fluid medium surrounding the sphere [kg/m³].
+    a : float
+        Radius of the sphere [m].
+    f : float
+        Frequency to calculate the scattering at [Hz].
+    target_longitudinal_c : float
+        Longitudinal sound speed in the material inside the sphere [m/s].
+    target_transverse_c : float
+        Transverse sound speed in the material inside the sphere [m/s].
+    target_rho : float
+        Density of the material inside the sphere [kg/m³].
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) of the sphere [dB].
+
+    Notes
+    -----
+    The class implements the code in MacLennan (1981).
+
+    References
+    ----------
+    MacLennan, D. N. (1981). The Theory of Solid Spheres as Sonar Calibration Targets.
+    Scottish Fisheries Research Report Number 22. Department of Agriculture and Fisheries
+    for Scotland.
+    """
+    if validate_parameters:
+        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'f': f,
+             'target_longitudinal_c': target_longitudinal_c,
+             'target_transverse_c': target_transverse_c,
+             'target_rho': target_rho}
+        self.validate_parameters(p)
+
+    q = wavenumber(medium_c, f)*a
+    q1 = q*medium_c/target_longitudinal_c
+    q2 = q*medium_c/target_transverse_c
+    alpha = 2. * (target_rho/medium_rho) * (target_transverse_c/medium_c)**2
+    beta = (target_rho/medium_rho) * (target_longitudinal_c/medium_c)**2 - alpha
+
+    # Use n instead of l (ell) because l looks like 1.
+    def S(n):
+        A2 = (n**2 + n-2) * spherical_jn(n, q2) + q2**2 * spherical_jnpp(n, q2)
+        A1 = 2*n*(n+1) * (q1*spherical_jn(n, q1, True) - spherical_jn(n, q1))
+        B2 = A2*q1**2 * (beta*spherical_jn(n, q1) - alpha*spherical_jnpp(n, q1))\
+            - A1*alpha * (spherical_jn(n, q2) - q2*spherical_jn(n, q2, True))
+        B1 = q * (A2*q1*spherical_jn(n, q1, True) - A1*spherical_jn(n, q2))
+        eta_n = atan(-(B2*spherical_jn(n, q, True) - B1*spherical_jn(n, q))
+                     / (B2*spherical_yn(n, q, True) - B1*spherical_yn(n, q)))
+
+        return (-1)**n * (2*n+1) * sin(eta_n) * exp(1j*eta_n)
+
+    # Estimate the number of terms to use in the summation
+    n_max = round(q+10)
+    tol = 1e-10  # somewhat arbitrary
+    while abs(S(n_max)) > tol:
+        n_max += 10
+
+    if n_max > 200:
+        warn('TS results may be inaccurate because the modal series required a large '
+             f'number ({n_max}) of terms to converge.')
+
+    n = range(n_max)
+
+    f_inf = -2.0/q * sum(map(S, n))
+
+    return 10*log10(a**2 * abs(f_inf)**2 / 4.0)
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/esmodel.py + 
26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+    super()._present_and_in(p, ['boundary_type'], self.boundary_types)
+    super()._present_and_positive(p, ['medium_rho', 'medium_c', 'a', 'f',
+                                      'target_longitudinal_c',
+                                      'target_transverse_c', 'target_rho'])
+
+
+
+ +
+ + + +
+ +
+ +

KAModel

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Kirchhoff approximation (KA) scattering model.

+

This class calculates acoustic scatter from arbitrary surfaces.

+ +
+ Source code in src/echosms/kamodel.py + 
16
+17
+18
+19
+20
+21
+22
+23
+24
def __init__(self):
+    super().__init__()
+    self.long_name = 'Kirchhoff approximation'
+    self.short_name = 'ka'
+    self.analytical_type = 'approximate'
+    self.boundary_types = ['pressure release']
+    self.shapes = ['closed surfaces']
+    self.max_ka = 20  # [1]
+    self.no_expand_parameters = ['mesh']
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(medium_c, theta, phi, f, mesh, boundary_type, validate_parameters=True, **kwargs) + +

+ + +
+ +

Calculate the scatter using the ka model for one set of parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
medium_c + float + +
+

Sound speed in the fluid medium surrounding the target [m/s].

+
+ 		+ required +
theta + float + +
+

Pitch angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
phi + float + +
+

Roll angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
f + float + +
+

Frequency to calculate the scattering at [Hz].

+
+ 		+ required +
mesh + Any + +
+

The triangular mesh that defines the scattering surface. This parameter must provide +attributes with names of:

+
    +
  • triangles_center (the position of the centre of each triangular face [m]),
    • +
  • face_normals (the outward-pointing unit normals for each triangular face),
    • +
  • area_faces (the area of each triangular face [m²]).
    • +
+

A suitable library for creating and manipulating triangular meshes +is trimesh.

+
+ 		+ required +
boundary_type + str + +
+

The boundary type. Supported types are given in the boundary_types class variable.

+
+ 		+ required +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) of the target [dB].

+
+
+ + +
+ Notes +

The class implements the code in Foote (1985).

+
+ +
+ References +

Foote, K. G. (1985). Rather-high-frequency sound scattering of swimbladdered fish. +The Journal of the Acoustical Society of America, 78(2), 688–700. +https://doi.org/10.1121/1.392438

+
+
+ Source code in src/echosms/kamodel.py + 
 36
+ 37
+ 38
+ 39
+ 40
+ 41
+ 42
+ 43
+ 44
+ 45
+ 46
+ 47
+ 48
+ 49
+ 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
def calculate_ts_single(self, medium_c, theta, phi, f, mesh,
+                        boundary_type, validate_parameters=True, **kwargs) -> float:
+    """
+    Calculate the scatter using the ka model for one set of parameters.
+
+    Parameters
+    ----------
+    medium_c : float
+        Sound speed in the fluid medium surrounding the target [m/s].
+    theta : float
+        Pitch angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    phi : float
+        Roll angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    f : float
+        Frequency to calculate the scattering at [Hz].
+    mesh : Any
+        The triangular mesh that defines the scattering surface. This parameter must provide
+        attributes with names of:
+
+        - `triangles_center` (the position of the centre of each triangular face [m]),
+        - `face_normals` (the outward-pointing unit normals for each triangular face),
+        - `area_faces` (the area of each triangular face [m²]).
+
+        A suitable library for creating and manipulating triangular meshes
+        is [trimesh](https://trimesh.org).
+    boundary_type : str
+        The boundary type. Supported types are given in the `boundary_types` class variable.
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) of the target [dB].
+
+    Notes
+    -----
+    The class implements the code in Foote (1985).
+
+    References
+    ----------
+    Foote, K. G. (1985). Rather-high-frequency sound scattering of swimbladdered fish.
+    The Journal of the Acoustical Society of America, 78(2), 688–700.
+    <https://doi.org/10.1121/1.392438>
+
+    """
+    if validate_parameters:
+        p = {'medium_c': medium_c, 'theta': theta, 'phi': phi, 'f': f, 'mesh': mesh}
+        self.validate_parameters(p)
+
+    if boundary_type not in self.boundary_types:
+        raise ValueError(f'The {self.long_name} model does not support '
+                         f'a model type of "{boundary_type}".')
+
+    # This model keeps the organism fixed and varies the incident wave vector. So need
+    # to convert the theta and phi echoSMs coordinate sytem Tait-Bryan angles
+    # into an (x,y,z) vector.
+
+    # Acoustic wave incident vector and its' norm
+    rot = R.from_euler('ZYX', (0, theta-90, -phi), degrees=True)
+    k_norm = rot.as_matrix() @ np.array([[0, 0, 1]]).T
+    k = k_norm * wavenumber(medium_c, f)
+
+    r = mesh.triangles_center  # position vector of each surface element
+    dS = mesh.area_faces.reshape((-1, 1))  # [m^2]
+
+    kn_nn = mesh.face_normals @ k_norm
+
+    fbs = 1./wavelength(medium_c, f)\
+        * np.sum(np.exp(2j*r @ k) * np.heaviside(kn_nn, 0.5) * kn_nn * dS)
+
+    return 10*log10(abs(fbs)**2)  # ts
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/kamodel.py + 
26
+27
+28
+29
+30
+31
+32
+33
+34
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+    super()._present_and_in(p, ['boundary_type'], self.boundary_types)
+    super()._present_and_positive(p, ['medium_c', 'f'])
+
+
+
+ +
+ + + +
+ +
+ +

MSSModel

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Modal series solution (MSS) scattering model.

+

This class calculates acoustic scatter from spheres and shells with various +boundary conditions, as listed in the boundary_types class attribute.

+ +
+ Source code in src/echosms/mssmodel.py + 
19
+20
+21
+22
+23
+24
+25
+26
+27
+28
def __init__(self):
+    super().__init__()
+    self.long_name = 'modal series solution'
+    self.short_name = 'mss'
+    self.analytical_type = 'exact'
+    self.boundary_types = ['fixed rigid', 'pressure release', 'fluid filled',
+                           'fluid shell fluid interior',
+                           'fluid shell pressure release interior']
+    self.shapes = ['sphere']
+    self.max_ka = 20  # [1]
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(medium_c, medium_rho, a, f, boundary_type, target_c=None, target_rho=None, shell_c=None, shell_rho=None, shell_thickness=None, validate_parameters=True, **kwargs) + +

+ + +
+ +

Calculate the scatter using the mss model for one set of parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
medium_c + float + +
+

Sound speed in the fluid medium surrounding the target [m/s].

+
+ 		+ required +
medium_rho + float + +
+

Density of the fluid medium surrounding the target [kg/m³].

+
+ 		+ required +
a + float + +
+

Radius of the spherical target [m].

+
+ 		+ required +
f + float + +
+

Frequency to calculate the scattering at [Hz].

+
+ 		+ required +
boundary_type + str + +
+

The boundary type. Supported types are given in the boundary_types class variable.

+
+ 		+ required +
target_c + float + +
+

Sound speed in the fluid inside the sphere [m/s]. +Only required for boundary_type of fluid filled.

+
+ 		+ None +
target_rho + float + +
+

Density of the fluid inside the sphere [kg/m³]. +Only required for boundary_type of fluid filled.

+
+ 		+ None +
shell_c + float + +
+

Sound speed in the spherical shell [m/s]. +Only required for boundary_types that include a fluid shell.

+
+ 		+ None +
shell_rho + float + +
+

Density in the spherical shell [kg/m³]. +Only required for boundary_types that include a fluid shell.

+
+ 		+ None +
shell_thickness + float + +
+

Thickness of the spherical shell [m]. This value is subtracted from a to give +the radius of the interior sphere. +Only required for boundary_types that include a fluid shell.

+
+ 		+ None +
validate_parameters + bool + +
+

Whether to validate the model parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) of the target [dB].

+
+
+ + +
+ Notes +

The class implements the code in Section A.1 of Jech et al. (2015).

+
+ +
+ References +

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., +Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. +Comparisons among ten models of acoustic backscattering used in aquatic ecosystem +research. Journal of the Acoustical Society of America 138, 3742–3764. +https://doi.org/10.1121/1.4937607

+
+
+ Source code in src/echosms/mssmodel.py + 
 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
+153
+154
+155
+156
+157
+158
+159
+160
+161
+162
+163
+164
+165
+166
+167
+168
+169
+170
+171
+172
+173
def calculate_ts_single(self, medium_c, medium_rho, a, f, boundary_type,
+                        target_c=None, target_rho=None,
+                        shell_c=None, shell_rho=None, shell_thickness=None,
+                        validate_parameters=True,
+                        **kwargs) -> float:
+    """
+    Calculate the scatter using the mss model for one set of parameters.
+
+    Parameters
+    ----------
+    medium_c : float
+        Sound speed in the fluid medium surrounding the target [m/s].
+    medium_rho : float
+        Density of the fluid medium surrounding the target [kg/m³].
+    a : float
+        Radius of the spherical target [m].
+    f : float
+        Frequency to calculate the scattering at [Hz].
+    boundary_type : str
+        The boundary type. Supported types are given in the `boundary_types` class variable.
+    target_c : float, optional
+        Sound speed in the fluid inside the sphere [m/s].
+        Only required for `boundary_type` of ``fluid filled``.
+    target_rho : float, optional
+        Density of the fluid inside the sphere [kg/m³].
+        Only required for `boundary_type` of ``fluid filled``.
+    shell_c : float, optional
+        Sound speed in the spherical shell [m/s].
+        Only required for `boundary_type`s that include a fluid shell.
+    shell_rho : float, optional
+        Density in the spherical shell [kg/m³].
+        Only required for `boundary_type`s that include a fluid shell.
+    shell_thickness : float, optional
+        Thickness of the spherical shell [m]. This value is subtracted from ``a`` to give
+        the radius of the interior sphere.
+        Only required for `boundary_type`s that include a fluid shell.
+    validate_parameters : bool
+        Whether to validate the model parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) of the target [dB].
+
+    Notes
+    -----
+    The class implements the code in Section A.1 of Jech et al. (2015).
+
+    References
+    ----------
+    Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N.,
+    Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015.
+    Comparisons among ten models of acoustic backscattering used in aquatic ecosystem
+    research. Journal of the Acoustical Society of America 138, 3742–3764.
+    <https://doi.org/10.1121/1.4937607>
+    """
+    if validate_parameters:
+        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'f': f,
+             'boundary_type': boundary_type, 'target_c': target_c, 'target_rho': target_rho,
+             'shell_c': shell_c, 'shell_rho': shell_rho, 'shell_thickness': shell_thickness}
+        self.validate_parameters(p)
+
+    k0 = wavenumber(medium_c, f)
+    ka = k0*a
+    n = np.arange(0, round(ka+20))
+
+    match boundary_type:
+        case 'fixed rigid':
+            A = list(map(lambda x: -spherical_jn(x, ka, True) / h1(x, ka, True), n))
+        case 'pressure release':
+            A = list(map(lambda x: -spherical_jn(x, ka) / h1(x, ka), n))
+        case 'fluid filled':
+            k1a = wavenumber(target_c, f)*a
+            gh = target_rho/medium_rho * target_c/medium_c
+
+            def Cn_fr(n):
+                return\
+                    ((spherical_jn(n, k1a, True)*spherical_yn(n, ka))
+                        / (spherical_jn(n, k1a)*spherical_jn(n, ka, True))
+                        - gh*(spherical_yn(n, ka, True)/spherical_jn(n, ka, True)))\
+                    / ((spherical_jn(n, k1a, True)*spherical_jn(n, ka))
+                       / (spherical_jn(n, k1a)*spherical_jn(n, ka, True))-gh)
+
+            A = -1/(1 + 1j*np.asarray(list(map(Cn_fr, n)), dtype=complex))
+        case 'fluid shell fluid interior':
+            b = a - shell_thickness
+
+            g21 = shell_rho / medium_rho
+            h21 = shell_c / medium_c
+            g32 = target_rho / shell_rho
+            h32 = target_c / shell_c
+
+            k1a = wavenumber(medium_c, f) * a
+            k2 = wavenumber(shell_c, f)
+            k3b = wavenumber(target_c, f) * b
+
+            def Cn_fsfi(n):
+                (b1, b2, a11, a21, a12, a22, a32, a13, a23, a33) =\
+                    MSSModel.__eqn9(n, k1a, g21, h21, k2*a, k2*b, k3b, h32, g32)
+                return (b1*a22*a33 + a13*b2*a32 - a12*b2*a33 - b1*a23*a32)\
+                    / (a11*a22*a33 + a13*a21*a32 - a12*a21*a33 - a11*a23*a32)
+
+            A = list(map(Cn_fsfi, n))
+        case 'fluid shell pressure release interior':
+            b = a - shell_thickness
+
+            g21 = shell_rho / medium_rho
+            h21 = shell_c / medium_c
+
+            k1a = wavenumber(medium_c, f) * a
+            k2 = wavenumber(shell_c, f)
+            ksa = k2 * a  # ksa is used in the paper, but isn't that the same as k2a?
+
+            def Cn_fspri(n):
+                (b1, b2, d1, d2, a11, a21) = MSSModel.__eqn10(n, k1a, g21, h21, ksa, k2*a, k2*b)
+                return (b1*d2-d1*b2) / (a11*d2-d1*a21)
+
+            A = list(map(Cn_fspri, n))
+        case _:
+            raise ValueError(f'The {self.long_name} model does not support '
+                             f'a model type of "{boundary_type}".')
+
+    fbs = -1j/k0 * np.sum((-1)**n * (2*n+1) * A)
+    return 20*log10(abs(fbs))  # ts
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/mssmodel.py + 
30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+
+    """
+    p = as_dict(params)
+    super()._present_and_in(p, ['boundary_type'], self.boundary_types)
+    super()._present_and_positive(p, ['medium_rho', 'a', 'f'])
+
+    for bt in np.atleast_1d(p['boundary_type']):
+        match bt:
+            case 'fluid filled':
+                super()._present_and_positive(p, ['target_c', 'target_rho'])
+            case 'fluid shell fluid interior':
+                super()._present_and_positive(p, ['target_c', 'target_rho', 'shell_c',
+                                                  'shell_rho', 'shell_thickness'])
+            case 'fluid shell pressure release interior':
+                super()._present_and_positive(p, ['shell_c', 'shell_rho', 'shell_thickness'])
+
+
+
+ +
+ + + +
+ +
+ +

PSMSModel

+ + +
+ + + + +
+

+ Bases: ScatterModelBase

+ + +

Prolate spheroidal modal series (PSMS) scattering model.

+ + +
+ Note +

The fluid filled boundary type implementation is currently only accurate +for weakly scattering interiors. Support for strongly scattering +(e.g., gas-filled) will come later.

+
+
+ Source code in src/echosms/psmsmodel.py + 
19
+20
+21
+22
+23
+24
+25
+26
def __init__(self):
+    super().__init__()
+    self.long_name = 'prolate spheroidal modal series'
+    self.short_name = 'psms'
+    self.analytical_type = 'exact'
+    self.boundary_types = ['fixed rigid', 'pressure release', 'fluid filled']
+    self.shapes = ['prolate spheroid']
+    self.max_ka = 10  # [1]
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ calculate_ts(data, expand=False, inplace=False, multiprocess=False) + +

+ + +
+ +

Calculate the target strength (TS) for many parameters.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
data + Pandas DataFrame, Xarray DataArray or dict + +
+

Requirements for the different input data types are:

+
    +
  • DataFrame: column names must match the function parameter names in + calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
    • +
  • DataArray: dimension names must match the function parameter names in + calculate_ts_single(). TS values will be calculated for all combinations of the + coordinate variables.
    • +
  • dict: keys must match the function parameters in calculate_ts_single(). + TS values will be calculated for all combinations of the dict values.
    • +
+
+ 		+ required +
multiprocess + bool + +
+

Split the ts calculation across CPU cores. Multiprocessing is currently provided by +mapply with little customisation. For more +sophisticated uses it may be preferred to use a multiprocessing package of your choice +directly on the calculate_ts_single() method. See the code in this method +(calculate_ts()) for an example.

+
+ 		+ False +
expand + bool + +
+

Only applicable if data is a dict. If True, will use +as_dataframe() +to expand the dict into a DataFrame with one column per dict key +and return that, adding a column named ts for the results.

+
+ 		+ False +
inplace + bool + +
+

Only applicable if data is a DataFrame. If True, the results +will be added to the input DataFrame in a column named ts. If a ts column +already exists, it is overwritten.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ None, list[float], Series, or DataFrame + +
+

The return type and value are determined by the type of the input variable (data) and +the expand and inplace parameters:

+
    +
  • dict input and expand=False returns a list of floats.
    • +
  • dict input and expand=True returns a DataFrame.
    • +
  • DataFrame input and inplace=False returns a Series.
    • +
  • DataFrame input and inplace=True modifies data and returns None.
    • +
  • DataArray input always modifies data and returns None.
    • +
+
+
+ +
+ Source code in src/echosms/scattermodelbase.py + 
 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):
+    """Calculate the target strength (TS) for many parameters.
+
+    Parameters
+    ----------
+    data : Pandas DataFrame, Xarray DataArray or dict
+        Requirements for the different input data types are:
+
+        - **DataFrame**: column names must match the function parameter names in
+          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
+        - **DataArray**: dimension names must match the function parameter names in
+          calculate_ts_single(). TS values will be calculated for all combinations of the
+          coordinate variables.
+        - **dict**: keys must match the function parameters in calculate_ts_single().
+          TS values will be calculated for all combinations of the dict values.
+
+    multiprocess : bool
+        Split the ts calculation across CPU cores. Multiprocessing is currently provided by
+        [mapply](https://github.com/ddelange/mapply) with little customisation. For more
+        sophisticated uses it may be preferred to use a multiprocessing package of your choice
+        directly on the `calculate_ts_single()` method. See the code in this method
+        (`calculate_ts()`) for an example.
+
+    expand : bool
+        Only applicable if `data` is a dict. If `True`, will use
+        [`as_dataframe()`][echosms.utils.as_dataframe]
+        to expand the dict into a DataFrame with one column per dict key
+        and return that, adding a column named `ts` for the results.
+
+    inplace : bool
+        Only applicable if `data` is a DataFrame. If `True`, the results
+        will be added to the input DataFrame in a column named `ts`. If a `ts` column
+        already exists, it is overwritten.
+
+    Returns
+    -------
+    : None, list[float], Series, or DataFrame
+        The return type and value are determined by the type of the input variable (`data`) and
+        the `expand` and `inplace` parameters:
+
+        - dict input and `expand=False` returns a list of floats.
+        - dict input and `expand=True` returns a DataFrame.
+        - DataFrame input and `inplace=False` returns a Series.
+        - DataFrame input and `inplace=True` modifies `data` and returns `None`.
+        - DataArray input always modifies `data` and returns `None`.
+
+    """
+    match data:
+        case dict():
+            data_df = as_dataframe(data, self.no_expand_parameters)
+        case pd.DataFrame():
+            data_df = data
+        case xr.DataArray():
+            data_df = data.to_dataframe().reset_index()
+            data_df.attrs = data.attrs
+        case _:
+            raise ValueError(f'Data type of {type(data)} is not supported'
+                             ' (only dictionaries, Pandas DataFrames and'
+                             ' Xarray DataArrays are).')
+
+    self.validate_parameters(data_df)
+
+    # Get the non-expandable model parameters
+    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}
+
+    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a
+    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma
+    # and parenthesis instead causes the tuple to have one entry of the dict.
+
+    if multiprocess:
+        from mapply.mapply import mapply
+        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)
+    else:  # this uses just one CPU
+        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)
+
+    match data:
+        case dict() if expand:
+            data_df['ts'] = ts
+            return data_df
+        case dict():
+            return ts.to_list()
+        case pd.DataFrame() if inplace:
+            data_df['ts'] = ts
+            return None
+        case pd.DataFrame():
+            return ts.rename('ts', inplace=True)
+        case xr.DataArray():
+            data.values = ts.to_numpy().reshape(data.shape)
+            return None
+        case _:
+            raise AssertionError('This code should never be reached - unsupported input data '
+                                 f'type of {type(data)}.')
+
+
+
+ +
+ +
+ + +

+ calculate_ts_single(medium_c, medium_rho, a, b, theta, f, boundary_type, target_c=None, target_rho=None, validate_parameters=True) + +

+ + +
+ +

Prolate spheroid modal series (PSMS) solution model.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
medium_c + float + +
+

Sound speed in the fluid medium surrounding the target [m/s].

+
+ 		+ required +
medium_rho + float + +
+

Density of the fluid medium surrounding the target [kg/m³].

+
+ 		+ required +
a + float + +
+

Prolate spheroid major axis radius [m].

+
+ 		+ required +
b + float + +
+

Prolate spheroid minor axis radius [m].

+
+ 		+ required +
theta + float + +
+

Pitch angle to calculate the scattering as per the echoSMs +coordinate system [°].

+
+ 		+ required +
f + float + +
+

Frequency to calculate the scattering at [Hz].

+
+ 		+ required +
boundary_type + str + +
+

The model type. Supported model types are given in the boundary_types class variable.

+
+ 		+ required +
target_c + float + +
+

Sound speed in the fluid inside the target [m/s]. +Only required for boundary_type of fluid filled.

+
+ 		+ None +
target_rho + float + +
+

Density of the fluid inside the target [kg/m³]. +Only required for boundary_type of fluid filled.

+
+ 		+ None +
validate_parameters + bool + +
+

Whether to validate the input parameters.

+
+ 		+ True +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The target strength (re 1 m²) of the target [dB].

+
+
+ + +
+ Notes +

The backscattered target strength of a pressure release or fluid-filled prolate spheroid +is calculated using the PSMS method of Furusawa (1988) and corrections in +Furusawa et al. (1994).

+
+ +
+ References +

Furusawa, M. (1988). "Prolate spheroidal models for predicting general + trends of fish target strength," J. Acoust. Soc. Jpn. 9, 13-24. +Furusawa, M., Miyanohana, Y., Ariji, M., and Sawada, Y. (1994). + “Prediction of krill target strength by liquid prolate spheroid + model,” Fish. Sci., 60, 261-265.

+
+
+ Source code in src/echosms/psmsmodel.py + 
 43
+ 44
+ 45
+ 46
+ 47
+ 48
+ 49
+ 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
+153
+154
+155
+156
+157
def calculate_ts_single(self, medium_c, medium_rho, a, b, theta, f, boundary_type,
+                        target_c=None, target_rho=None, validate_parameters=True):
+    """Prolate spheroid modal series (PSMS) solution model.
+
+    Parameters
+    ----------
+    medium_c : float
+        Sound speed in the fluid medium surrounding the target [m/s].
+    medium_rho : float
+        Density of the fluid medium surrounding the target [kg/m³].
+    a : float
+        Prolate spheroid major axis radius [m].
+    b : float
+        Prolate spheroid minor axis radius [m].
+    theta : float
+        Pitch angle to calculate the scattering as per the echoSMs
+        [coordinate system](https://ices-tools-dev.github.io/echoSMs/
+        conventions/#coordinate-systems) [°].
+    f : float
+        Frequency to calculate the scattering at [Hz].
+    boundary_type : str
+        The model type. Supported model types are given in the `boundary_types` class variable.
+    target_c : float
+        Sound speed in the fluid inside the target [m/s].
+        Only required for `boundary_type` of ``fluid filled``.
+    target_rho : float
+        Density of the fluid inside the target [kg/m³].
+        Only required for `boundary_type` of ``fluid filled``.
+    validate_parameters : bool
+        Whether to validate the input parameters.
+
+    Returns
+    -------
+    : float
+        The target strength (re 1 m²) of the target [dB].
+
+    Notes
+    -----
+    The backscattered target strength of a pressure release or fluid-filled prolate spheroid
+    is calculated using the PSMS method of Furusawa (1988) and corrections in
+    Furusawa et al. (1994).
+
+    References
+    ----------
+    Furusawa, M. (1988). "Prolate spheroidal models for predicting general
+        trends of fish target strength," J. Acoust. Soc. Jpn. 9, 13-24.
+    Furusawa, M., Miyanohana, Y., Ariji, M., and Sawada, Y. (1994).
+        “Prediction of krill target strength by liquid prolate spheroid
+        model,” Fish. Sci., 60, 261-265.
+    """
+    if validate_parameters:
+        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'b': b,
+             'theta': theta, 'f': f, 'boundary_type': boundary_type,
+             'target_c': target_c, 'target_rho': target_rho}
+        self.validate_parameters(p)
+
+    if boundary_type not in self.boundary_types:
+        raise ValueError(f'The {self.long_name} model does not support '
+                         f'a model type of "{boundary_type}".')
+
+    xim = (1.0 - (b/a)**2)**(-.5)
+    q = a/xim  # semi-focal length
+
+    km = wavenumber(medium_c, f)
+    hm = km*q
+
+    if boundary_type == 'fluid filled':
+        g = target_rho / medium_rho
+        ht = wavenumber(target_c, f)*q
+
+    # Phi, the port/starboard angle is fixed for this code
+    phi_inc = np.pi  # incident direction
+    phi_sca = np.pi + phi_inc  # scattered direction
+
+    theta_inc = np.deg2rad(theta)  # incident direction
+    theta_sca = np.pi - theta_inc  # scattered direction
+
+    # Approximate limits on the summations
+    m_max = int(np.ceil(2*km*b))
+    n_max = int(m_max + np.ceil(hm/2))
+
+    f_sca = 0.0
+    for m in range(m_max+1):
+        epsilon_m = Neumann(m)
+        cos_term = np.cos(m*(phi_sca - phi_inc))
+        for n in range(m, n_max+1):
+            Smn_inc, _ = pro_ang1(m, n, hm, np.cos(theta_inc))
+            Smn_sca, _ = pro_ang1(m, n, hm, np.cos(theta_sca))
+            # The Meixner-Schäfke normalisation scheme for the angular function of the first
+            # kind. Refer to eqn 21.7.11 in Abramowitz, M., and Stegun, I. A. (1964).
+            # Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables
+            # (Dover, New York), 10th ed.
+            N_mn = 2/(2*n+1) * factorial(n+m) / factorial(n-m)
+
+            R1m, dR1m = pro_rad1(m, n, hm, xim)
+            R2m, dR2m = pro_rad2(m, n, hm, xim)
+
+            match boundary_type:
+                case 'fluid filled':
+                    # Note: we can implement the simpler equations if impedances are
+                    # similar between the medium and the target. The gas-filled
+                    # condition does not meet that, so we have two paths here. The simplified
+                    # equations are quicker, so it is worth to do.
+                    if (abs(1.0-target_c/medium_c) <= 0.01) and (abs(1.0-g) <= 0.01):
+                        Amn = PSMSModel._fluidfilled_approx(m, n, hm, ht, xim, g)
+                    else:
+                        Amn = PSMSModel._fluidfilled_exact(m, n, hm, ht, xim, g, theta_inc)
+                case 'pressure release':
+                    Amn = -R1m/(R1m + 1j*R2m)
+                case 'fixed rigid':
+                    Amn = -dR1m/(dR1m + 1j*dR2m)
+
+            f_sca += epsilon_m * (Smn_inc / N_mn) * Smn_sca * Amn * cos_term
+
+    return 20*np.log10(np.abs(-2j / km * f_sca))
+
+
+
+ +
+ +
+ + +

+ validate_parameters(params) + +

+ + +
+ +

Validate the model parameters.

+

See here for calling details.

+ +
+ Source code in src/echosms/psmsmodel.py + 
28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
def validate_parameters(self, params):
+    """Validate the model parameters.
+
+    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.
+    """
+
+    p = as_dict(params)
+    super()._present_and_in(p, ['boundary_type'], self.boundary_types)
+    super()._present_and_positive(p, ['medium_c', 'medium_rho', 'a', 'b', 'f'])
+
+    for bt in np.atleast_1d(p['boundary_type']):
+        match bt:
+            case 'fluid filled':
+                super()._present_and_positive(p, ['target_c', 'target_rho'])
+
+
+
+ +
+ + + +
+ +
+ +

ReferenceModels

+ + +
+ + + + +
+ + +

Provide access to reference scattering model parameters.

+

Reference models are the models and parameters defined in Jech et al. (2015). +The parameters are stored in a TOML-formatted file in the echoSMs repository +and this class provides easy access to the data in that file. Additional reference +models may be defined in the future and added to the TOML file (for example, entries +have been added for all known calibration sphere sizes).

+ + +

Attributes:

+ + + + + + + + + + + + + + + +
NameTypeDescription
definitions + dict + +
+

A dict representation of the target definitions.toml file.

+
+
+ + +

Raises:

+ + + + + + + + + + + + + + + + + +
TypeDescription
+ TOMLDecodeError + +
+

If the target definitions.toml file is not valid TOML.

+
+
+ KeyError + +
+

If the target definitions.toml file has multiple target entries with the same name.

+
+
+ + +
+ References +

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., +Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. +Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. +Journal of the Acoustical Society of America 138, 3742–3764. https://doi.org/10.1121/1.4937607

+
+
+ Source code in src/echosms/referencemodels.py + 
42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
def __init__(self):
+    self.defs_filename = Path(__file__).parent/Path('resources')/Path('target definitions.toml')
+
+    self.definitions = []
+
+    with open(self.defs_filename, 'rb') as f:
+        try:
+            self.definitions = tomllib.load(f)
+        except tomllib.TOMLDecodeError as e:
+            raise SyntaxError(f'Error while parsing file "{self.defs_filename.name}"') from e
+
+    # Flag duplicate target names
+    pda = pd.Series(self.names())
+    duplicates = list(pda[pda.duplicated()])
+    if duplicates:
+        raise KeyError(f'The "{self.defs_filename.name}" file has multiple targets '
+                       f'with the same name: '+', '.join(duplicates))
+
+    # Substitute parameters names in the target section by the values of those
+    # parameters.
+    for t in self.definitions['target']:
+        for k, v in t.items():
+            try:
+                t[k] = self.definitions['parameters'][v]
+            except (KeyError, TypeError):
+                pass
+
+
+ + + +
+ + + + + + + + + +
+ + +

+ names() + +

+ + +
+ +

Names of all model definitions.

+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ iterable of str + +
+

All model names in the target definitions.toml file.

+
+
+ +
+ Source code in src/echosms/referencemodels.py + 
69
+70
+71
+72
+73
+74
+75
+76
+77
def names(self):
+    """Names of all model definitions.
+
+    Returns
+    -------
+    : iterable of str
+        All model names in the ``target definitions.toml`` file.
+    """
+    return [n['name'] for n in self.definitions['target']]
+
+
+
+ +
+ +
+ + +

+ parameters(name) + +

+ + +
+ +

Model parameters for a particular model.

+

Model parameters are a subset of the model specification where the metadata items have +been removed.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
name + str + +
+

The name of a model in the target definitions.toml file.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ dict + +
+

The model parameters for the requested model or an empty set if no model with that name.

+
+
+ +
+ Source code in src/echosms/referencemodels.py + 
 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
def parameters(self, name):
+    """Model parameters for a particular model.
+
+    Model parameters are a subset of the model specification where the metadata items have
+    been removed.
+
+    Parameters
+    ----------
+    name : str
+        The name of a model in the ``target definitions.toml`` file.
+
+    Returns
+    -------
+    : dict
+        The model parameters for the requested model or an empty set if no model with that name.
+
+    """
+    s = self.specification(name)
+
+    if not s:
+        return []
+
+    # Remove the entries that are not parameters
+    p = s.copy()
+    for k in ['name', 'shape', 'description', 'source', 'benchmark_model']:
+        p.pop(k, None)
+    return p
+
+
+
+ +
+ +
+ + +

+ specification(name) + +

+ + +
+ +

Model definitions for a particular model.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
name + str + +
+

The name of a model in the target definitions.toml file.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ dict + +
+

The model definitions for the requested model or an empty set if no model +with that name.

+
+
+ +
+ Source code in src/echosms/referencemodels.py + 
79
+80
+81
+82
+83
+84
+85
+86
+87
+88
+89
+90
+91
+92
+93
+94
+95
+96
+97
def specification(self, name):
+    """Model definitions for a particular model.
+
+    Parameters
+    ----------
+    name : str
+        The name of a model in the ``target definitions.toml`` file.
+
+    Returns
+    -------
+    : dict
+        The model definitions for the requested model or an empty set if no model
+        with that name.
+    """
+    s = [t for t in self.definitions['target'] if t['name'] == name]
+    if not s:
+        return s
+
+    return s[0]
+
+
+
+ +
+ + + +
+ +
+ +

BenchmarkData

+ + +
+ + + + +
+ + +

Convenient interface to the benchmark dataset.

+

This dataset contains the TS results from Jech et al. (2015).

+ + +

Attributes:

+ + + + + + + + + + + + + + + + + + + + +
NameTypeDescription
angle_dataset + Pandas DataFrame + +
+

The angle dataset from the benchmark model runs.

+
+
freq_dataset + Pandas DataFrame + +
+

The frequency dataset from the benchmark model runs.

+
+
+ + +
+ Notes +

The column names in the source benchmark files have been changed to be the same as those used +in the ReferenceModels model definitions.

+
+ +
+ References +

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., +Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. +Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. +Journal of the Acoustical Society of America 138, 3742-3764. https://doi.org/10.1121/1.4937607

+
+
+ Source code in src/echosms/benchmarkdata.py + 
59
+60
+61
+62
+63
+64
+65
+66
+67
+68
+69
+70
+71
def __init__(self):
+
+    data_directory = Path(__file__).parent/Path('resources')/Path('BenchMark_Data')
+
+    angle_data_file = data_directory/'Benchmark_Angle_TS.csv'
+    freq_data_file = data_directory/'Benchmark_Frequency_TS.csv'
+
+    self.angle_dataset = pd.read_csv(angle_data_file)
+    self.freq_dataset = pd.read_csv(freq_data_file)
+
+    # Change the column names to match the reference model names used in ReferenceModels
+    self.angle_dataset.rename(columns=BenchmarkData.a_rename, inplace=True)
+    self.freq_dataset.rename(columns=BenchmarkData.f_rename, inplace=True)
+
+
+ + + +
+ + + + + + + + + + + +
+ +
+ +

Utilities

+ + +
+ + + + +
+ +

Miscellaneous utility functions.

+ + + +
+ + + + + + + + + +
+ + +

+ Neumann(m) + +

+ + +
+ +

Neumann number.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
m + int + +
+

The input integer.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ int + +
+

The Neumann number.

+
+
+ +
+ Source code in src/echosms/utils.py + 
14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
def Neumann(m: int) -> int:
+    """Neumann number.
+
+    Parameters
+    ----------
+    m :
+        The input integer.
+
+    Returns
+    -------
+    :
+        The Neumann number.
+    """
+    if m == 0:
+        return 1
+    return 2
+
+
+
+ +
+ +
+ + +

+ as_dataarray(params, no_expand=[]) + +

+ + +
+ +

Convert model parameters from dict form to a Xarray DataArray.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
params + dict + +
+

The model parameters.

+
+ 		+ required +
no_expand + list + +
+

Key values of the non-expandable model parameters in params.

+
+ 		+ [] +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ DataArray + +
+

Returns a multi-dimensional DataArray generated from the Cartesian product of all +expandable items in the input dict. Non-expandable items are added to the DataArray +attrs property. Expandable items are those that can be sensibly expanded into +DataArray coordinates. Not all models have non-expandable items. +The array is named ts, the values are initialised to nan, the +dimension names are the dict keys, and the coordinate variables are the dict values.

+
+
+ +
+ Source code in src/echosms/utils.py + 
153
+154
+155
+156
+157
+158
+159
+160
+161
+162
+163
+164
+165
+166
+167
+168
+169
+170
+171
+172
+173
+174
+175
+176
+177
+178
+179
+180
+181
+182
+183
+184
+185
def as_dataarray(params: dict, no_expand: list = []) -> xr.DataArray:
+    """Convert model parameters from dict form to a Xarray DataArray.
+
+    Parameters
+    ----------
+    params :
+        The model parameters.
+
+    no_expand :
+        Key values of the non-expandable model parameters in `params`.
+
+    Returns
+    -------
+    :
+        Returns a multi-dimensional DataArray generated from the Cartesian product of all
+        expandable items in the input dict. Non-expandable items are added to the DataArray
+        attrs property. Expandable items are those that can be sensibly expanded into
+        DataArray coordinates. Not all models have non-expandable items.
+        The array is named `ts`, the values are initialised to `nan`, the
+        dimension names are the dict keys, and the coordinate variables are the dict values.
+
+    """
+    expand, nexpand = split_dict(params, no_expand)
+
+    # Convert scalars to iterables so xarray is happy
+    for k, v in expand.items():
+        if not isinstance(v, Iterable) or isinstance(v, str):
+            expand[k] = [v]
+
+    sz = [len(v) for k, v in expand.items()]
+    return xr.DataArray(data=np.full(sz, np.nan), coords=expand, name='ts',
+                        attrs={'units': 'dB', 'dB_reference': '1 m^2',
+                               'parameters': nexpand})
+
+
+
+ +
+ +
+ + +

+ as_dataframe(params, no_expand=[]) + +

+ + +
+ +

Convert model parameters from dict form to a Pandas DataFrame.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
params + dict + +
+

The model parameters.

+
+ 		+ required +
no_expand + list + +
+

Key values of the non-expandable model parameters in params.

+
+ 		+ [] +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ DataFrame + +
+

Returns a Pandas DataFrame generated from the Cartesian product of all expandable +items in the input dict. DataFrame column names are obtained from the dict keys. +Non-expandable items are added to the DataFrame attrs property. Expandable items are +those that can be sensibly expanded into DataFrame columns. Not all models have +non-expandable items.

+
+
+ +
+ Source code in src/echosms/utils.py + 
188
+189
+190
+191
+192
+193
+194
+195
+196
+197
+198
+199
+200
+201
+202
+203
+204
+205
+206
+207
+208
+209
+210
+211
+212
+213
+214
+215
+216
+217
+218
def as_dataframe(params: dict, no_expand: list = []) -> pd.DataFrame:
+    """Convert model parameters from dict form to a Pandas DataFrame.
+
+    Parameters
+    ----------
+    params :
+        The model parameters.
+
+    no_expand :
+        Key values of the non-expandable model parameters in `params`.
+
+    Returns
+    -------
+    :
+        Returns a Pandas DataFrame generated from the Cartesian product of all expandable
+        items in the input dict. DataFrame column names are obtained from the dict keys.
+        Non-expandable items are added to the DataFrame attrs property. Expandable items are
+        those that can be sensibly expanded into DataFrame columns. Not all models have
+        non-expandable items.
+
+    """
+    expand, nexpand = split_dict(params, no_expand)
+
+    # Use meshgrid to do the Cartesian product then create a Pandas DataFrame from that, having
+    # flattened the multidimensional arrays and using a dict to provide column names.
+    # This preserves the differing dtypes in each column compared to other ways of
+    # constructing the DataFrame).
+    df = pd.DataFrame({k: t.flatten()
+                       for k, t in zip(expand.keys(), np.meshgrid(*tuple(expand.values())))})
+    df.attrs = {'parameters': nexpand}
+    return df
+
+
+
+ +
+ +
+ + +

+ as_dict(params) + +

+ + +
+ +

Convert model parameters from DataFrame or DataArray to dict.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
params + dict | DataFrame | DataArray + +
+

The model parameters

+
+ 		+ required +
+ + +

Raises:

+ + + + + + + + + + + + + +
TypeDescription
+ TypeError: + +
+

If the input data type is not supported.

+
+
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ dict + +
+

A dict containing the model parameters.

+
+
+ +
+ Source code in src/echosms/utils.py + 
221
+222
+223
+224
+225
+226
+227
+228
+229
+230
+231
+232
+233
+234
+235
+236
+237
+238
+239
+240
+241
+242
+243
+244
+245
+246
+247
+248
+249
+250
def as_dict(params: dict | pd.DataFrame | xr.DataArray) -> dict:
+    """Convert model parameters from DataFrame or DataArray to dict.
+
+    Parameters
+    ----------
+    params:
+        The model parameters
+
+    Raises
+    ------
+    TypeError:
+        If the input data type is not supported.
+
+    Returns
+    -------
+    :
+        A dict containing the model parameters.
+    """
+    if isinstance(params, dict):
+        return params
+
+    # Get the non-expandable model parameters
+    p = params.attrs['parameters'] if 'parameters' in params.attrs else {}
+
+    if isinstance(params, xr.DataArray):
+        return dict(zip(params.coords, params.indexes.values())) | p
+    elif isinstance(params, pd.DataFrame):
+        return params.to_dict(orient='list') | p
+
+    raise TypeError('Only dict, DataFrame, or DataArray are accepted.')
+
+
+
+ +
+ +
+ + +

+ h1(n, z, derivative=False) + +

+ + +
+ +

Spherical Hankel function of the first kind or its' derivative.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
n + int + +
+

Order (n ≥ 0).

+
+ 		+ required +
z + float + +
+

Argument of the Hankel function.

+
+ 		+ required +
derivative + +
+

if True, the value of the derivative (rather than the function itself) is returned.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ complex + +
+

Value of the spherical Hankel function

+
+
+ + +

Raises:

+ + + + + + + + + + + + + +
TypeDescription
+ ValueError + +
+

For negative n values.

+
+
+ + +
+ Notes +

The value of the Hankel function is calculated from spherical Bessel functions [1].

+

The derivative is computed from spherical Hankel functions [2].

+
+ +
+ References +

[1] https://dlmf.nist.gov/10.47.E10

+

[2] https://dlmf.nist.gov/10.51.E2

+
+
+ Source code in src/echosms/utils.py + 
 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
def h1(n: int, z: float, derivative=False) -> complex:
+    """Spherical Hankel function of the first kind or its' derivative.
+
+    Parameters
+    ----------
+    n :
+        Order (n ≥ 0).
+    z :
+        Argument of the Hankel function.
+    derivative :
+        if True, the value of the derivative (rather than the function itself) is returned.
+
+    Returns
+    -------
+    :
+        Value of the spherical Hankel function
+
+    Raises
+    ------
+    ValueError
+        For negative n values.
+
+    Notes
+    -----
+    The value of the Hankel function is calculated from spherical Bessel functions [1].
+
+    The derivative is computed from spherical Hankel functions [2].
+
+    References
+    ----------
+    [1] <https://dlmf.nist.gov/10.47.E10>
+
+    [2] <https://dlmf.nist.gov/10.51.E2>
+    """
+    if n < 0:
+        raise ValueError('Negative n values are not supported for spherical Hankel functions.')
+
+    if not derivative:
+        return spherical_jn(n, z) + 1j*spherical_yn(n, z)
+    return -h1(n+1, z) + (n/z) * h1(n, z)
+
+
+
+ +
+ +
+ + +

+ pro_ang1(m, n, c, eta, norm=False) + +

+ + +
+ +

Prolate spheroidal angular function of the first kind and derivative.

+

Calculates the prolate spheroidal angular function of the first kind and its' +derivative with respect to eta.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
m + int + +
+

The order parameter (≥ 0)

+
+ 		+ required +
n + int + +
+

The degree parameter (≥ m).

+
+ 		+ required +
c + float + +
+

The size parameter.

+
+ 		+ required +
eta + float + +
+

The angular coordinate, η, where |η| ≤ 1.

+
+ 		+ required +
norm + +
+

If False, returned values are not normalised (i.e., the Meixner-Schäfke normlalisation +scheme is used). For large m this norm becomes very large. If True the returned values +are scaled by the square root of the normalisation of the corresponding Legendre function. +This avoids the large values that occur when norm is False.

+
+ 		+ False +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ tuple[float, float] + +
+

The value of the prolate spheroidal angular function and its' derivative.

+
+
+ + +
+ Notes +

This method uses the prolate spheroidal wave function code for non complex +arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on +github. This code is in Fortran90 +and was interfaced to Python using numpy.f2py then wrapped with the current method to +provide a similar calling convention as the Scipy function of the same name.

+
+ +
+ References +

Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal +radial functions of the first kind and their first derivatives. Quarterly of Applied +Mathematics, 60(3), 589-599. https://doi.org/10.1090/qam/1914443

+

Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal +Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied +Mathematics, 62(3), 493-507. https://doi.org/10.1090/qam/2086042

+
+
+ Source code in src/echosms/utils.py + 
253
+254
+255
+256
+257
+258
+259
+260
+261
+262
+263
+264
+265
+266
+267
+268
+269
+270
+271
+272
+273
+274
+275
+276
+277
+278
+279
+280
+281
+282
+283
+284
+285
+286
+287
+288
+289
+290
+291
+292
+293
+294
+295
+296
+297
+298
+299
+300
+301
+302
+303
+304
+305
+306
+307
+308
+309
+310
+311
def pro_ang1(m: int, n: int, c: float, eta: float, norm=False) -> tuple[float, float]:
+    """Prolate spheroidal angular function of the first kind and derivative.
+
+    Calculates the prolate spheroidal angular function of the first kind and its'
+    derivative with respect to `eta`.
+
+    Parameters
+    ----------
+    m :
+        The order parameter (≥ 0)
+    n :
+        The degree parameter (≥ `m`).
+    c :
+        The size parameter.
+    eta :
+        The angular coordinate, η, where |η| ≤ 1.
+    norm :
+        If `False`, returned values are not normalised (i.e., the Meixner-Schäfke normlalisation
+        scheme is used). For large `m` this norm becomes very large. If `True` the returned values
+        are scaled by the square root of the normalisation of the corresponding Legendre function.
+        This avoids the large values that occur when `norm` is `False`.
+
+    Returns
+    -------
+    :
+        The value of the prolate spheroidal angular function and its' derivative.
+
+    Notes
+    -----
+    This method uses the prolate spheroidal wave function code for non complex
+    arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on
+    [github](https://github.com/MathieuandSpheroidalWaveFunctions). This code is in Fortran90
+    and was interfaced to Python using `numpy.f2py` then wrapped with the current method to
+    provide a similar calling convention as the Scipy function of the same name.
+
+    References
+    ----------
+    Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal
+    radial functions of the first kind and their first derivatives. Quarterly of Applied
+    Mathematics, 60(3), 589-599. <https://doi.org/10.1090/qam/1914443>
+
+    Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal
+    Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied
+    Mathematics, 62(3), 493-507. <https://doi.org/10.1090/qam/2086042>
+    """
+    if m < 0:
+        raise ValueError('The m parameter must be positive.')
+    if n < m:
+        raise ValueError('The n parameter must be greater than or equal to the m parameter.')
+    if abs(eta) > 1.0:
+        raise ValueError('The eta parameter must be less than or equal to 1')
+
+    a = prolate_swf.profcn(c=c, m=m, lnum=n-m+2, x1=0.0, ioprad=0, iopang=2,
+                           iopnorm=int(norm), arg=[eta])
+    p = swf_t._make(a)
+    s = p.s1c * np.float_power(10.0, p.is1e)
+    sp = p.s1dc * np.float_power(10.0, p.is1de)
+
+    return s[n-m][0], sp[n-m][0]
+
+
+
+ +
+ +
+ + +

+ pro_rad1(m, n, c, xi) + +

+ + +
+ +

Prolate spheroidal radial function of the first kind and derivative.

+

Calculates the prolate spheroidal radial function of the first kind and its' +derivative with respect to xi.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
m + int + +
+

The order parameter (≥ 0).

+
+ 		+ required +
n + int + +
+

The degree parameter (≥ m).

+
+ 		+ required +
c + float + +
+

The size parameter.

+
+ 		+ required +
xi + float + +
+

The radial coordinate, ξ, where ξ ≥ 1.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ tuple[float, float] + +
+

The value of the prolate spheroidal radial function and its' derivative.

+
+
+ + +
+ Notes +

This method uses the prolate spheroidal wave function code for non complex +arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on +github. This code is in Fortran90 +and was interfaced to Python using numpy.f2py then wrapped with the current method to +provide a similar calling convention as the Scipy function of the same name.

+
+ +
+ References +

Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal +radial functions of the first kind and their first derivatives. Quarterly of Applied +Mathematics, 60(3), 589-599. https://doi.org/10.1090/qam/1914443

+

Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal +Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied +Mathematics, 62(3), 493-507. https://doi.org/10.1090/qam/2086042

+
+
+ Source code in src/echosms/utils.py + 
314
+315
+316
+317
+318
+319
+320
+321
+322
+323
+324
+325
+326
+327
+328
+329
+330
+331
+332
+333
+334
+335
+336
+337
+338
+339
+340
+341
+342
+343
+344
+345
+346
+347
+348
+349
+350
+351
+352
+353
+354
+355
+356
+357
+358
+359
+360
+361
+362
+363
+364
+365
+366
def pro_rad1(m: int, n: int, c: float, xi: float) -> tuple[float, float]:
+    """Prolate spheroidal radial function of the first kind and derivative.
+
+    Calculates the prolate spheroidal radial function of the first kind and its'
+    derivative with respect to `xi`.
+
+    Parameters
+    ----------
+    m :
+        The order parameter (≥ 0).
+    n :
+        The degree parameter (≥ `m`).
+    c :
+        The size parameter.
+    xi :
+        The radial coordinate, ξ, where ξ ≥ 1.
+
+    Returns
+    -------
+    :
+        The value of the prolate spheroidal radial function and its' derivative.
+
+    Notes
+    -----
+    This method uses the prolate spheroidal wave function code for non complex
+    arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on
+    [github](https://github.com/MathieuandSpheroidalWaveFunctions). This code is in Fortran90
+    and was interfaced to Python using `numpy.f2py` then wrapped with the current method to
+    provide a similar calling convention as the Scipy function of the same name.
+
+    References
+    ----------
+    Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal
+    radial functions of the first kind and their first derivatives. Quarterly of Applied
+    Mathematics, 60(3), 589-599. <https://doi.org/10.1090/qam/1914443>
+
+    Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal
+    Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied
+    Mathematics, 62(3), 493-507. <https://doi.org/10.1090/qam/2086042>
+    """
+    if m < 0:
+        raise ValueError('The m parameter must be positive.')
+    if n < m:
+        raise ValueError('The n parameter must be greater than or equal to the m parameter.')
+    if xi < 1.0:
+        raise ValueError('The xi parameter must be greater than or equal to 1')
+
+    a = prolate_swf.profcn(c=c, m=m, lnum=n-m+2, x1=xi-1.0, ioprad=1, iopang=0, iopnorm=0, arg=[0])
+    p = swf_t._make(a)
+    s = p.r1c * np.float_power(10.0, p.ir1e)
+    sp = p.r1dc * np.float_power(10.0, p.ir1de)
+
+    return s[n-m], sp[n-m]
+
+
+
+ +
+ +
+ + +

+ pro_rad2(m, n, c, xi) + +

+ + +
+ +

Prolate spheroidal radial function of the second kind and derivative.

+

Calculates the prolate spheroidal radial function of the second kind and its' +derivative with respect to xi.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
m + int + +
+

The order parameter (≥ 0).

+
+ 		+ required +
n + int + +
+

The degree parameter (≥ m).

+
+ 		+ required +
c + float + +
+

The size parameter.

+
+ 		+ required +
xi + float + +
+

The radial coordinate, ξ, where ξ ≥ 1.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ tuple[float, float] + +
+

The value of the prolate spheroidal radial function and its' derivative.

+
+
+ + +
+ Notes +

This method uses the prolate spheroidal wave function code for non complex +arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on +github. This code is in Fortran90 +and was interfaced to Python using numpy.f2py then wrapped with the current method to +provide a similar calling convention as the Scipy function of the same name.

+
+ +
+ References +

Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal +radial functions of the first kind and their first derivatives. Quarterly of Applied +Mathematics, 60(3), 589-599. https://doi.org/10.1090/qam/1914443

+

Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal +Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied +Mathematics, 62(3), 493-507. https://doi.org/10.1090/qam/2086042

+
+
+ Source code in src/echosms/utils.py + 
369
+370
+371
+372
+373
+374
+375
+376
+377
+378
+379
+380
+381
+382
+383
+384
+385
+386
+387
+388
+389
+390
+391
+392
+393
+394
+395
+396
+397
+398
+399
+400
+401
+402
+403
+404
+405
+406
+407
+408
+409
+410
+411
+412
+413
+414
+415
+416
+417
+418
+419
+420
+421
+422
+423
+424
+425
+426
+427
+428
+429
+430
+431
def pro_rad2(m: int, n: int, c: float, xi: float) -> tuple[float, float]:
+    """Prolate spheroidal radial function of the second kind and derivative.
+
+    Calculates the prolate spheroidal radial function of the second kind and its'
+    derivative with respect to `xi`.
+
+    Parameters
+    ----------
+    m :
+        The order parameter (≥ 0).
+    n :
+        The degree parameter (≥ `m`).
+    c :
+        The size parameter.
+    xi :
+        The radial coordinate, ξ, where ξ ≥ 1.
+
+    Returns
+    -------
+    :
+        The value of the prolate spheroidal radial function and its' derivative.
+
+    Notes
+    -----
+    This method uses the prolate spheroidal wave function code for non complex
+    arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on
+    [github](https://github.com/MathieuandSpheroidalWaveFunctions). This code is in Fortran90
+    and was interfaced to Python using `numpy.f2py` then wrapped with the current method to
+    provide a similar calling convention as the Scipy function of the same name.
+
+    References
+    ----------
+    Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal
+    radial functions of the first kind and their first derivatives. Quarterly of Applied
+    Mathematics, 60(3), 589-599. <https://doi.org/10.1090/qam/1914443>
+
+    Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal
+    Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied
+    Mathematics, 62(3), 493-507. <https://doi.org/10.1090/qam/2086042>
+    """
+    if m < 0:
+        raise ValueError('The m parameter must be positive.')
+    if n < m:
+        raise ValueError('The n parameter must be greater than or equal to the m parameter.')
+    if xi < 1.0:
+        raise ValueError('The xi parameter must be greater than or equal to 1')
+
+    ioprad = 1 if xi-1.0 < 1e-10 else 2
+
+    # Add +2 to lnum instead of +1 as it exposes a bug in the Fortran code - if n = 0, zeros
+    # are returned instead of the correct value.
+    a = prolate_swf.profcn(c=c, m=m, lnum=n-m+2, x1=xi-1.0,
+                           ioprad=ioprad, iopang=0, iopnorm=0, arg=[0])
+    p = swf_t._make(a)
+
+    if ioprad == 1:
+        s = np.inf
+        sp = np.inf
+    else:
+        s = p.r2c * np.float_power(10.0, p.ir2e)
+        sp = p.r2dc * np.float_power(10.0, p.ir2de)
+
+    return s[n-m], sp[n-m]
+
+
+
+ +
+ +
+ + +

+ spherical_jnpp(n, z) + +

+ + +
+ +

Second derivative of the spherical Bessel function.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
n + int + +
+

Order (n ≥ 0)

+
+ 		+ required +
z + float + +
+

Argument of the Bessel function.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The second derivative of the spherical Bessel function.

+
+
+ +
+ Source code in src/echosms/utils.py + 
112
+113
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
def spherical_jnpp(n: int, z: float) -> float:
+    """Second derivative of the spherical Bessel function.
+
+    Parameters
+    ----------
+    n :
+        Order (n ≥ 0)
+    z :
+        Argument of the Bessel function.
+
+    Returns
+    -------
+    :
+        The second derivative of the spherical Bessel function.
+
+    """
+    return 1./z**2 * ((n**2-n-z**2)*spherical_jn(n, z) + 2.*z*spherical_jn(n+1, z))
+
+
+
+ +
+ +
+ + +

+ split_dict(d, s) + +

+ + +
+ +

Split a dict into two dicts based on a list of keys.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
d + dict + +
+

Dict to be split.

+
+ 		+ required +
s + list + +
+

List of dict keys to use for splitting d.

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ tuple(dict, dict) + +
+

The input dict split into two dicts based on the keys in s. The first tuple item +contains the items that do not have keys in s.

+
+
+ +
+ Source code in src/echosms/utils.py + 
131
+132
+133
+134
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
def split_dict(d: dict, s: list) -> tuple[dict, dict]:
+    """Split a dict into two dicts based on a list of keys.
+
+    Parameters
+    ----------
+    d : dict
+        Dict to be split.
+
+    s: list
+        List of dict keys to use for splitting `d`.
+
+    Returns
+    -------
+    : tuple(dict, dict)
+        The `input` dict split into two dicts based on the keys in `s`. The first tuple item
+        contains the items that do not have keys in `s`.
+    """
+    contains = {k: v for k, v in d.items() if k in s}
+    ncontains = {k: v for k, v in d.items() if k not in s}
+    return ncontains, contains
+
+
+
+ +
+ +
+ + +

+ wavelength(c, f) + +

+ + +
+ +

Calculate the acoustic wavelength.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
c + float + +
+

Sound speed [m/s]

+
+ 		+ required +
f + float + +
+

Frequency [Hz]

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The acoustic wavelength [m].

+
+
+ +
+ Source code in src/echosms/utils.py + 
51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
def wavelength(c: float, f: float) -> float:
+    """Calculate the acoustic wavelength.
+
+    Parameters
+    ----------
+    c :
+        Sound speed [m/s]
+
+    f :
+        Frequency [Hz]
+
+    Returns
+    -------
+    :
+        The acoustic wavelength [m].
+    """
+    return c/f
+
+
+
+ +
+ +
+ + +

+ wavenumber(c, f) + +

+ + +
+ +

Calculate the acoustic wavenumber.

+ + +

Parameters:

+ + + + + + + + + + + + + + + + + + + + + + + +
NameTypeDescriptionDefault
c + float + +
+

Sound speed [m/s]

+
+ 		+ required +
f + float + +
+

Frequency [Hz]

+
+ 		+ required +
+ + +

Returns:

+ + + + + + + + + + + + + +
TypeDescription
+ float + +
+

The acoustic wavenumber [m⁻¹].

+
+
+ +
+ Source code in src/echosms/utils.py + 
32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
def wavenumber(c: float, f: float) -> float:
+    """Calculate the acoustic wavenumber.
+
+    Parameters
+    ----------
+    c :
+        Sound speed [m/s]
+
+    f :
+        Frequency [Hz]
+
+    Returns
+    -------
+    :
+        The acoustic wavenumber [m⁻¹].
+    """
+    return 2*np.pi*f/c
+
+
+
+ +
+ + + +
+ +
+ +
+ + + + + + + + + + + + + +
+
+ + + +
+ + + +
+ + + +
+
+
+
+ + + + + + + + + + \ No newline at end of file diff --git a/assets/_mkdocstrings.css b/assets/_mkdocstrings.css new file mode 100644 index 0000000..85449ec --- /dev/null +++ b/assets/_mkdocstrings.css @@ -0,0 +1,119 @@ + +/* Avoid breaking parameter names, etc. in table cells. */ +.doc-contents td code { + word-break: normal !important; +} + +/* No line break before first paragraph of descriptions. */ +.doc-md-description, +.doc-md-description>p:first-child { + display: inline; +} + +/* Max width for docstring sections tables. */ +.doc .md-typeset__table, +.doc .md-typeset__table table { + display: table !important; + width: 100%; +} + +.doc .md-typeset__table tr { + display: table-row; +} + +/* Defaults in Spacy table style. */ +.doc-param-default { + float: right; +} + +/* Backward-compatibility: docstring section titles in bold. */ +.doc-section-title { + font-weight: bold; +} + +/* Symbols in Navigation and ToC. */ +:root, +[data-md-color-scheme="default"] { + --doc-symbol-attribute-fg-color: #953800; + --doc-symbol-function-fg-color: #8250df; + --doc-symbol-method-fg-color: #8250df; + --doc-symbol-class-fg-color: #0550ae; + --doc-symbol-module-fg-color: #5cad0f; + + --doc-symbol-attribute-bg-color: #9538001a; + --doc-symbol-function-bg-color: #8250df1a; + --doc-symbol-method-bg-color: #8250df1a; + --doc-symbol-class-bg-color: #0550ae1a; + --doc-symbol-module-bg-color: #5cad0f1a; +} + +[data-md-color-scheme="slate"] { + --doc-symbol-attribute-fg-color: #ffa657; + --doc-symbol-function-fg-color: #d2a8ff; + --doc-symbol-method-fg-color: #d2a8ff; + --doc-symbol-class-fg-color: #79c0ff; + --doc-symbol-module-fg-color: #baff79; + + --doc-symbol-attribute-bg-color: #ffa6571a; + --doc-symbol-function-bg-color: #d2a8ff1a; + --doc-symbol-method-bg-color: #d2a8ff1a; + --doc-symbol-class-bg-color: #79c0ff1a; + --doc-symbol-module-bg-color: #baff791a; +} + +code.doc-symbol { + border-radius: .1rem; + font-size: .85em; + padding: 0 .3em; + font-weight: bold; +} + +code.doc-symbol-attribute { + color: var(--doc-symbol-attribute-fg-color); + background-color: var(--doc-symbol-attribute-bg-color); +} + +code.doc-symbol-attribute::after { + content: "attr"; +} + +code.doc-symbol-function { + color: var(--doc-symbol-function-fg-color); + background-color: var(--doc-symbol-function-bg-color); +} + +code.doc-symbol-function::after { + content: "func"; +} + +code.doc-symbol-method { + color: var(--doc-symbol-method-fg-color); + background-color: var(--doc-symbol-method-bg-color); +} + +code.doc-symbol-method::after { + content: "meth"; +} + +code.doc-symbol-class { + color: var(--doc-symbol-class-fg-color); + background-color: var(--doc-symbol-class-bg-color); +} + +code.doc-symbol-class::after { + content: "class"; +} + +code.doc-symbol-module { + color: var(--doc-symbol-module-fg-color); + background-color: var(--doc-symbol-module-bg-color); +} + +code.doc-symbol-module::after { + content: "mod"; +} + +.doc-signature .autorefs { + color: inherit; + border-bottom: 1px dotted currentcolor; +} diff --git a/assets/images/favicon.png b/assets/images/favicon.png new file mode 100644 index 0000000..1cf13b9 Binary files /dev/null and b/assets/images/favicon.png differ diff --git a/assets/javascripts/bundle.525ec568.min.js b/assets/javascripts/bundle.525ec568.min.js new file mode 100644 index 0000000..4b08eae --- /dev/null +++ b/assets/javascripts/bundle.525ec568.min.js @@ -0,0 +1,16 @@ +"use strict";(()=>{var Wi=Object.create;var gr=Object.defineProperty;var Di=Object.getOwnPropertyDescriptor;var Vi=Object.getOwnPropertyNames,Vt=Object.getOwnPropertySymbols,Ni=Object.getPrototypeOf,yr=Object.prototype.hasOwnProperty,ao=Object.prototype.propertyIsEnumerable;var io=(e,t,r)=>t in e?gr(e,t,{enumerable:!0,configurable:!0,writable:!0,value:r}):e[t]=r,$=(e,t)=>{for(var r in t||(t={}))yr.call(t,r)&&io(e,r,t[r]);if(Vt)for(var r of Vt(t))ao.call(t,r)&&io(e,r,t[r]);return e};var so=(e,t)=>{var r={};for(var o in e)yr.call(e,o)&&t.indexOf(o)<0&&(r[o]=e[o]);if(e!=null&&Vt)for(var o of Vt(e))t.indexOf(o)<0&&ao.call(e,o)&&(r[o]=e[o]);return r};var xr=(e,t)=>()=>(t||e((t={exports:{}}).exports,t),t.exports);var zi=(e,t,r,o)=>{if(t&&typeof t=="object"||typeof t=="function")for(let n of Vi(t))!yr.call(e,n)&&n!==r&&gr(e,n,{get:()=>t[n],enumerable:!(o=Di(t,n))||o.enumerable});return e};var Mt=(e,t,r)=>(r=e!=null?Wi(Ni(e)):{},zi(t||!e||!e.__esModule?gr(r,"default",{value:e,enumerable:!0}):r,e));var co=(e,t,r)=>new Promise((o,n)=>{var i=p=>{try{s(r.next(p))}catch(c){n(c)}},a=p=>{try{s(r.throw(p))}catch(c){n(c)}},s=p=>p.done?o(p.value):Promise.resolve(p.value).then(i,a);s((r=r.apply(e,t)).next())});var lo=xr((Er,po)=>{(function(e,t){typeof Er=="object"&&typeof po!="undefined"?t():typeof define=="function"&&define.amd?define(t):t()})(Er,function(){"use strict";function e(r){var o=!0,n=!1,i=null,a={text:!0,search:!0,url:!0,tel:!0,email:!0,password:!0,number:!0,date:!0,month:!0,week:!0,time:!0,datetime:!0,"datetime-local":!0};function s(k){return!!(k&&k!==document&&k.nodeName!=="HTML"&&k.nodeName!=="BODY"&&"classList"in k&&"contains"in k.classList)}function p(k){var ft=k.type,qe=k.tagName;return!!(qe==="INPUT"&&a[ft]&&!k.readOnly||qe==="TEXTAREA"&&!k.readOnly||k.isContentEditable)}function c(k){k.classList.contains("focus-visible")||(k.classList.add("focus-visible"),k.setAttribute("data-focus-visible-added",""))}function l(k){k.hasAttribute("data-focus-visible-added")&&(k.classList.remove("focus-visible"),k.removeAttribute("data-focus-visible-added"))}function f(k){k.metaKey||k.altKey||k.ctrlKey||(s(r.activeElement)&&c(r.activeElement),o=!0)}function u(k){o=!1}function d(k){s(k.target)&&(o||p(k.target))&&c(k.target)}function y(k){s(k.target)&&(k.target.classList.contains("focus-visible")||k.target.hasAttribute("data-focus-visible-added"))&&(n=!0,window.clearTimeout(i),i=window.setTimeout(function(){n=!1},100),l(k.target))}function L(k){document.visibilityState==="hidden"&&(n&&(o=!0),X())}function X(){document.addEventListener("mousemove",J),document.addEventListener("mousedown",J),document.addEventListener("mouseup",J),document.addEventListener("pointermove",J),document.addEventListener("pointerdown",J),document.addEventListener("pointerup",J),document.addEventListener("touchmove",J),document.addEventListener("touchstart",J),document.addEventListener("touchend",J)}function te(){document.removeEventListener("mousemove",J),document.removeEventListener("mousedown",J),document.removeEventListener("mouseup",J),document.removeEventListener("pointermove",J),document.removeEventListener("pointerdown",J),document.removeEventListener("pointerup",J),document.removeEventListener("touchmove",J),document.removeEventListener("touchstart",J),document.removeEventListener("touchend",J)}function J(k){k.target.nodeName&&k.target.nodeName.toLowerCase()==="html"||(o=!1,te())}document.addEventListener("keydown",f,!0),document.addEventListener("mousedown",u,!0),document.addEventListener("pointerdown",u,!0),document.addEventListener("touchstart",u,!0),document.addEventListener("visibilitychange",L,!0),X(),r.addEventListener("focus",d,!0),r.addEventListener("blur",y,!0),r.nodeType===Node.DOCUMENT_FRAGMENT_NODE&&r.host?r.host.setAttribute("data-js-focus-visible",""):r.nodeType===Node.DOCUMENT_NODE&&(document.documentElement.classList.add("js-focus-visible"),document.documentElement.setAttribute("data-js-focus-visible",""))}if(typeof window!="undefined"&&typeof document!="undefined"){window.applyFocusVisiblePolyfill=e;var t;try{t=new CustomEvent("focus-visible-polyfill-ready")}catch(r){t=document.createEvent("CustomEvent"),t.initCustomEvent("focus-visible-polyfill-ready",!1,!1,{})}window.dispatchEvent(t)}typeof document!="undefined"&&e(document)})});var qr=xr((hy,On)=>{"use strict";/*! + * escape-html + * Copyright(c) 2012-2013 TJ Holowaychuk + * Copyright(c) 2015 Andreas Lubbe + * Copyright(c) 2015 Tiancheng "Timothy" Gu + * MIT Licensed + */var $a=/["'&<>]/;On.exports=Pa;function Pa(e){var t=""+e,r=$a.exec(t);if(!r)return t;var o,n="",i=0,a=0;for(i=r.index;i{/*! + * clipboard.js v2.0.11 + * https://clipboardjs.com/ + * + * Licensed MIT © Zeno Rocha + */(function(t,r){typeof It=="object"&&typeof Yr=="object"?Yr.exports=r():typeof define=="function"&&define.amd?define([],r):typeof It=="object"?It.ClipboardJS=r():t.ClipboardJS=r()})(It,function(){return function(){var e={686:function(o,n,i){"use strict";i.d(n,{default:function(){return Ui}});var a=i(279),s=i.n(a),p=i(370),c=i.n(p),l=i(817),f=i.n(l);function u(V){try{return document.execCommand(V)}catch(A){return!1}}var d=function(A){var M=f()(A);return u("cut"),M},y=d;function L(V){var A=document.documentElement.getAttribute("dir")==="rtl",M=document.createElement("textarea");M.style.fontSize="12pt",M.style.border="0",M.style.padding="0",M.style.margin="0",M.style.position="absolute",M.style[A?"right":"left"]="-9999px";var F=window.pageYOffset||document.documentElement.scrollTop;return M.style.top="".concat(F,"px"),M.setAttribute("readonly",""),M.value=V,M}var X=function(A,M){var F=L(A);M.container.appendChild(F);var D=f()(F);return u("copy"),F.remove(),D},te=function(A){var M=arguments.length>1&&arguments[1]!==void 0?arguments[1]:{container:document.body},F="";return typeof A=="string"?F=X(A,M):A instanceof HTMLInputElement&&!["text","search","url","tel","password"].includes(A==null?void 0:A.type)?F=X(A.value,M):(F=f()(A),u("copy")),F},J=te;function k(V){"@babel/helpers - typeof";return typeof Symbol=="function"&&typeof Symbol.iterator=="symbol"?k=function(M){return typeof M}:k=function(M){return M&&typeof Symbol=="function"&&M.constructor===Symbol&&M!==Symbol.prototype?"symbol":typeof M},k(V)}var ft=function(){var A=arguments.length>0&&arguments[0]!==void 0?arguments[0]:{},M=A.action,F=M===void 0?"copy":M,D=A.container,Y=A.target,$e=A.text;if(F!=="copy"&&F!=="cut")throw new Error('Invalid "action" value, use either "copy" or "cut"');if(Y!==void 0)if(Y&&k(Y)==="object"&&Y.nodeType===1){if(F==="copy"&&Y.hasAttribute("disabled"))throw new Error('Invalid "target" attribute. Please use "readonly" instead of "disabled" attribute');if(F==="cut"&&(Y.hasAttribute("readonly")||Y.hasAttribute("disabled")))throw new Error(`Invalid "target" attribute. You can't cut text from elements with "readonly" or "disabled" attributes`)}else throw new Error('Invalid "target" value, use a valid Element');if($e)return J($e,{container:D});if(Y)return F==="cut"?y(Y):J(Y,{container:D})},qe=ft;function Fe(V){"@babel/helpers - typeof";return typeof Symbol=="function"&&typeof Symbol.iterator=="symbol"?Fe=function(M){return typeof M}:Fe=function(M){return M&&typeof Symbol=="function"&&M.constructor===Symbol&&M!==Symbol.prototype?"symbol":typeof M},Fe(V)}function ki(V,A){if(!(V instanceof A))throw new TypeError("Cannot call a class as a function")}function no(V,A){for(var M=0;M0&&arguments[0]!==void 0?arguments[0]:{};this.action=typeof D.action=="function"?D.action:this.defaultAction,this.target=typeof D.target=="function"?D.target:this.defaultTarget,this.text=typeof D.text=="function"?D.text:this.defaultText,this.container=Fe(D.container)==="object"?D.container:document.body}},{key:"listenClick",value:function(D){var Y=this;this.listener=c()(D,"click",function($e){return Y.onClick($e)})}},{key:"onClick",value:function(D){var Y=D.delegateTarget||D.currentTarget,$e=this.action(Y)||"copy",Dt=qe({action:$e,container:this.container,target:this.target(Y),text:this.text(Y)});this.emit(Dt?"success":"error",{action:$e,text:Dt,trigger:Y,clearSelection:function(){Y&&Y.focus(),window.getSelection().removeAllRanges()}})}},{key:"defaultAction",value:function(D){return vr("action",D)}},{key:"defaultTarget",value:function(D){var Y=vr("target",D);if(Y)return document.querySelector(Y)}},{key:"defaultText",value:function(D){return vr("text",D)}},{key:"destroy",value:function(){this.listener.destroy()}}],[{key:"copy",value:function(D){var Y=arguments.length>1&&arguments[1]!==void 0?arguments[1]:{container:document.body};return J(D,Y)}},{key:"cut",value:function(D){return y(D)}},{key:"isSupported",value:function(){var D=arguments.length>0&&arguments[0]!==void 0?arguments[0]:["copy","cut"],Y=typeof D=="string"?[D]:D,$e=!!document.queryCommandSupported;return Y.forEach(function(Dt){$e=$e&&!!document.queryCommandSupported(Dt)}),$e}}]),M}(s()),Ui=Fi},828:function(o){var n=9;if(typeof Element!="undefined"&&!Element.prototype.matches){var i=Element.prototype;i.matches=i.matchesSelector||i.mozMatchesSelector||i.msMatchesSelector||i.oMatchesSelector||i.webkitMatchesSelector}function a(s,p){for(;s&&s.nodeType!==n;){if(typeof s.matches=="function"&&s.matches(p))return s;s=s.parentNode}}o.exports=a},438:function(o,n,i){var a=i(828);function s(l,f,u,d,y){var L=c.apply(this,arguments);return l.addEventListener(u,L,y),{destroy:function(){l.removeEventListener(u,L,y)}}}function p(l,f,u,d,y){return typeof l.addEventListener=="function"?s.apply(null,arguments):typeof u=="function"?s.bind(null,document).apply(null,arguments):(typeof l=="string"&&(l=document.querySelectorAll(l)),Array.prototype.map.call(l,function(L){return s(L,f,u,d,y)}))}function c(l,f,u,d){return function(y){y.delegateTarget=a(y.target,f),y.delegateTarget&&d.call(l,y)}}o.exports=p},879:function(o,n){n.node=function(i){return i!==void 0&&i instanceof HTMLElement&&i.nodeType===1},n.nodeList=function(i){var a=Object.prototype.toString.call(i);return i!==void 0&&(a==="[object NodeList]"||a==="[object HTMLCollection]")&&"length"in i&&(i.length===0||n.node(i[0]))},n.string=function(i){return typeof i=="string"||i instanceof String},n.fn=function(i){var a=Object.prototype.toString.call(i);return a==="[object Function]"}},370:function(o,n,i){var a=i(879),s=i(438);function p(u,d,y){if(!u&&!d&&!y)throw new Error("Missing required arguments");if(!a.string(d))throw new TypeError("Second argument must be a String");if(!a.fn(y))throw new TypeError("Third argument must be a Function");if(a.node(u))return c(u,d,y);if(a.nodeList(u))return l(u,d,y);if(a.string(u))return f(u,d,y);throw new TypeError("First argument must be a String, HTMLElement, HTMLCollection, or NodeList")}function c(u,d,y){return u.addEventListener(d,y),{destroy:function(){u.removeEventListener(d,y)}}}function l(u,d,y){return Array.prototype.forEach.call(u,function(L){L.addEventListener(d,y)}),{destroy:function(){Array.prototype.forEach.call(u,function(L){L.removeEventListener(d,y)})}}}function f(u,d,y){return s(document.body,u,d,y)}o.exports=p},817:function(o){function n(i){var a;if(i.nodeName==="SELECT")i.focus(),a=i.value;else if(i.nodeName==="INPUT"||i.nodeName==="TEXTAREA"){var s=i.hasAttribute("readonly");s||i.setAttribute("readonly",""),i.select(),i.setSelectionRange(0,i.value.length),s||i.removeAttribute("readonly"),a=i.value}else{i.hasAttribute("contenteditable")&&i.focus();var p=window.getSelection(),c=document.createRange();c.selectNodeContents(i),p.removeAllRanges(),p.addRange(c),a=p.toString()}return a}o.exports=n},279:function(o){function n(){}n.prototype={on:function(i,a,s){var p=this.e||(this.e={});return(p[i]||(p[i]=[])).push({fn:a,ctx:s}),this},once:function(i,a,s){var p=this;function c(){p.off(i,c),a.apply(s,arguments)}return c._=a,this.on(i,c,s)},emit:function(i){var a=[].slice.call(arguments,1),s=((this.e||(this.e={}))[i]||[]).slice(),p=0,c=s.length;for(p;p0&&i[i.length-1])&&(c[0]===6||c[0]===2)){r=0;continue}if(c[0]===3&&(!i||c[1]>i[0]&&c[1]=e.length&&(e=void 0),{value:e&&e[o++],done:!e}}};throw new TypeError(t?"Object is not iterable.":"Symbol.iterator is not defined.")}function N(e,t){var r=typeof Symbol=="function"&&e[Symbol.iterator];if(!r)return e;var o=r.call(e),n,i=[],a;try{for(;(t===void 0||t-- >0)&&!(n=o.next()).done;)i.push(n.value)}catch(s){a={error:s}}finally{try{n&&!n.done&&(r=o.return)&&r.call(o)}finally{if(a)throw a.error}}return i}function q(e,t,r){if(r||arguments.length===2)for(var o=0,n=t.length,i;o1||p(d,L)})},y&&(n[d]=y(n[d])))}function p(d,y){try{c(o[d](y))}catch(L){u(i[0][3],L)}}function c(d){d.value instanceof nt?Promise.resolve(d.value.v).then(l,f):u(i[0][2],d)}function l(d){p("next",d)}function f(d){p("throw",d)}function u(d,y){d(y),i.shift(),i.length&&p(i[0][0],i[0][1])}}function uo(e){if(!Symbol.asyncIterator)throw new TypeError("Symbol.asyncIterator is not defined.");var t=e[Symbol.asyncIterator],r;return t?t.call(e):(e=typeof he=="function"?he(e):e[Symbol.iterator](),r={},o("next"),o("throw"),o("return"),r[Symbol.asyncIterator]=function(){return this},r);function o(i){r[i]=e[i]&&function(a){return new Promise(function(s,p){a=e[i](a),n(s,p,a.done,a.value)})}}function n(i,a,s,p){Promise.resolve(p).then(function(c){i({value:c,done:s})},a)}}function H(e){return typeof e=="function"}function ut(e){var t=function(o){Error.call(o),o.stack=new Error().stack},r=e(t);return r.prototype=Object.create(Error.prototype),r.prototype.constructor=r,r}var zt=ut(function(e){return function(r){e(this),this.message=r?r.length+` errors occurred during unsubscription: +`+r.map(function(o,n){return n+1+") "+o.toString()}).join(` + `):"",this.name="UnsubscriptionError",this.errors=r}});function Qe(e,t){if(e){var r=e.indexOf(t);0<=r&&e.splice(r,1)}}var Ue=function(){function e(t){this.initialTeardown=t,this.closed=!1,this._parentage=null,this._finalizers=null}return e.prototype.unsubscribe=function(){var t,r,o,n,i;if(!this.closed){this.closed=!0;var a=this._parentage;if(a)if(this._parentage=null,Array.isArray(a))try{for(var s=he(a),p=s.next();!p.done;p=s.next()){var c=p.value;c.remove(this)}}catch(L){t={error:L}}finally{try{p&&!p.done&&(r=s.return)&&r.call(s)}finally{if(t)throw t.error}}else a.remove(this);var l=this.initialTeardown;if(H(l))try{l()}catch(L){i=L instanceof zt?L.errors:[L]}var f=this._finalizers;if(f){this._finalizers=null;try{for(var u=he(f),d=u.next();!d.done;d=u.next()){var y=d.value;try{ho(y)}catch(L){i=i!=null?i:[],L instanceof zt?i=q(q([],N(i)),N(L.errors)):i.push(L)}}}catch(L){o={error:L}}finally{try{d&&!d.done&&(n=u.return)&&n.call(u)}finally{if(o)throw o.error}}}if(i)throw new zt(i)}},e.prototype.add=function(t){var r;if(t&&t!==this)if(this.closed)ho(t);else{if(t instanceof e){if(t.closed||t._hasParent(this))return;t._addParent(this)}(this._finalizers=(r=this._finalizers)!==null&&r!==void 0?r:[]).push(t)}},e.prototype._hasParent=function(t){var r=this._parentage;return r===t||Array.isArray(r)&&r.includes(t)},e.prototype._addParent=function(t){var r=this._parentage;this._parentage=Array.isArray(r)?(r.push(t),r):r?[r,t]:t},e.prototype._removeParent=function(t){var r=this._parentage;r===t?this._parentage=null:Array.isArray(r)&&Qe(r,t)},e.prototype.remove=function(t){var r=this._finalizers;r&&Qe(r,t),t instanceof e&&t._removeParent(this)},e.EMPTY=function(){var t=new e;return t.closed=!0,t}(),e}();var Tr=Ue.EMPTY;function qt(e){return e instanceof Ue||e&&"closed"in e&&H(e.remove)&&H(e.add)&&H(e.unsubscribe)}function ho(e){H(e)?e():e.unsubscribe()}var Pe={onUnhandledError:null,onStoppedNotification:null,Promise:void 0,useDeprecatedSynchronousErrorHandling:!1,useDeprecatedNextContext:!1};var dt={setTimeout:function(e,t){for(var r=[],o=2;o0},enumerable:!1,configurable:!0}),t.prototype._trySubscribe=function(r){return this._throwIfClosed(),e.prototype._trySubscribe.call(this,r)},t.prototype._subscribe=function(r){return this._throwIfClosed(),this._checkFinalizedStatuses(r),this._innerSubscribe(r)},t.prototype._innerSubscribe=function(r){var o=this,n=this,i=n.hasError,a=n.isStopped,s=n.observers;return i||a?Tr:(this.currentObservers=null,s.push(r),new Ue(function(){o.currentObservers=null,Qe(s,r)}))},t.prototype._checkFinalizedStatuses=function(r){var o=this,n=o.hasError,i=o.thrownError,a=o.isStopped;n?r.error(i):a&&r.complete()},t.prototype.asObservable=function(){var r=new j;return r.source=this,r},t.create=function(r,o){return new To(r,o)},t}(j);var To=function(e){oe(t,e);function t(r,o){var n=e.call(this)||this;return n.destination=r,n.source=o,n}return t.prototype.next=function(r){var o,n;(n=(o=this.destination)===null||o===void 0?void 0:o.next)===null||n===void 0||n.call(o,r)},t.prototype.error=function(r){var o,n;(n=(o=this.destination)===null||o===void 0?void 0:o.error)===null||n===void 0||n.call(o,r)},t.prototype.complete=function(){var r,o;(o=(r=this.destination)===null||r===void 0?void 0:r.complete)===null||o===void 0||o.call(r)},t.prototype._subscribe=function(r){var o,n;return(n=(o=this.source)===null||o===void 0?void 0:o.subscribe(r))!==null&&n!==void 0?n:Tr},t}(g);var _r=function(e){oe(t,e);function t(r){var o=e.call(this)||this;return o._value=r,o}return Object.defineProperty(t.prototype,"value",{get:function(){return this.getValue()},enumerable:!1,configurable:!0}),t.prototype._subscribe=function(r){var o=e.prototype._subscribe.call(this,r);return!o.closed&&r.next(this._value),o},t.prototype.getValue=function(){var r=this,o=r.hasError,n=r.thrownError,i=r._value;if(o)throw n;return this._throwIfClosed(),i},t.prototype.next=function(r){e.prototype.next.call(this,this._value=r)},t}(g);var At={now:function(){return(At.delegate||Date).now()},delegate:void 0};var Ct=function(e){oe(t,e);function t(r,o,n){r===void 0&&(r=1/0),o===void 0&&(o=1/0),n===void 0&&(n=At);var i=e.call(this)||this;return i._bufferSize=r,i._windowTime=o,i._timestampProvider=n,i._buffer=[],i._infiniteTimeWindow=!0,i._infiniteTimeWindow=o===1/0,i._bufferSize=Math.max(1,r),i._windowTime=Math.max(1,o),i}return t.prototype.next=function(r){var o=this,n=o.isStopped,i=o._buffer,a=o._infiniteTimeWindow,s=o._timestampProvider,p=o._windowTime;n||(i.push(r),!a&&i.push(s.now()+p)),this._trimBuffer(),e.prototype.next.call(this,r)},t.prototype._subscribe=function(r){this._throwIfClosed(),this._trimBuffer();for(var o=this._innerSubscribe(r),n=this,i=n._infiniteTimeWindow,a=n._buffer,s=a.slice(),p=0;p0?e.prototype.schedule.call(this,r,o):(this.delay=o,this.state=r,this.scheduler.flush(this),this)},t.prototype.execute=function(r,o){return o>0||this.closed?e.prototype.execute.call(this,r,o):this._execute(r,o)},t.prototype.requestAsyncId=function(r,o,n){return n===void 0&&(n=0),n!=null&&n>0||n==null&&this.delay>0?e.prototype.requestAsyncId.call(this,r,o,n):(r.flush(this),0)},t}(gt);var Lo=function(e){oe(t,e);function t(){return e!==null&&e.apply(this,arguments)||this}return t}(yt);var kr=new Lo(Oo);var Mo=function(e){oe(t,e);function t(r,o){var n=e.call(this,r,o)||this;return n.scheduler=r,n.work=o,n}return t.prototype.requestAsyncId=function(r,o,n){return n===void 0&&(n=0),n!==null&&n>0?e.prototype.requestAsyncId.call(this,r,o,n):(r.actions.push(this),r._scheduled||(r._scheduled=vt.requestAnimationFrame(function(){return r.flush(void 0)})))},t.prototype.recycleAsyncId=function(r,o,n){var i;if(n===void 0&&(n=0),n!=null?n>0:this.delay>0)return e.prototype.recycleAsyncId.call(this,r,o,n);var a=r.actions;o!=null&&((i=a[a.length-1])===null||i===void 0?void 0:i.id)!==o&&(vt.cancelAnimationFrame(o),r._scheduled=void 0)},t}(gt);var _o=function(e){oe(t,e);function t(){return e!==null&&e.apply(this,arguments)||this}return t.prototype.flush=function(r){this._active=!0;var o=this._scheduled;this._scheduled=void 0;var n=this.actions,i;r=r||n.shift();do if(i=r.execute(r.state,r.delay))break;while((r=n[0])&&r.id===o&&n.shift());if(this._active=!1,i){for(;(r=n[0])&&r.id===o&&n.shift();)r.unsubscribe();throw i}},t}(yt);var me=new _o(Mo);var S=new j(function(e){return e.complete()});function Yt(e){return e&&H(e.schedule)}function Hr(e){return e[e.length-1]}function Xe(e){return H(Hr(e))?e.pop():void 0}function ke(e){return Yt(Hr(e))?e.pop():void 0}function Bt(e,t){return typeof Hr(e)=="number"?e.pop():t}var xt=function(e){return e&&typeof e.length=="number"&&typeof e!="function"};function Gt(e){return H(e==null?void 0:e.then)}function Jt(e){return H(e[bt])}function Xt(e){return Symbol.asyncIterator&&H(e==null?void 0:e[Symbol.asyncIterator])}function Zt(e){return new TypeError("You provided "+(e!==null&&typeof e=="object"?"an invalid object":"'"+e+"'")+" where a stream was expected. You can provide an Observable, Promise, ReadableStream, Array, AsyncIterable, or Iterable.")}function Zi(){return typeof Symbol!="function"||!Symbol.iterator?"@@iterator":Symbol.iterator}var er=Zi();function tr(e){return H(e==null?void 0:e[er])}function rr(e){return fo(this,arguments,function(){var r,o,n,i;return Nt(this,function(a){switch(a.label){case 0:r=e.getReader(),a.label=1;case 1:a.trys.push([1,,9,10]),a.label=2;case 2:return[4,nt(r.read())];case 3:return o=a.sent(),n=o.value,i=o.done,i?[4,nt(void 0)]:[3,5];case 4:return[2,a.sent()];case 5:return[4,nt(n)];case 6:return[4,a.sent()];case 7:return a.sent(),[3,2];case 8:return[3,10];case 9:return r.releaseLock(),[7];case 10:return[2]}})})}function or(e){return H(e==null?void 0:e.getReader)}function U(e){if(e instanceof j)return e;if(e!=null){if(Jt(e))return ea(e);if(xt(e))return ta(e);if(Gt(e))return ra(e);if(Xt(e))return Ao(e);if(tr(e))return oa(e);if(or(e))return na(e)}throw Zt(e)}function ea(e){return new j(function(t){var r=e[bt]();if(H(r.subscribe))return r.subscribe(t);throw new TypeError("Provided object does not correctly implement Symbol.observable")})}function ta(e){return new j(function(t){for(var r=0;r=2;return function(o){return o.pipe(e?b(function(n,i){return e(n,i,o)}):le,Te(1),r?De(t):Qo(function(){return new ir}))}}function jr(e){return e<=0?function(){return S}:E(function(t,r){var o=[];t.subscribe(T(r,function(n){o.push(n),e=2,!0))}function pe(e){e===void 0&&(e={});var t=e.connector,r=t===void 0?function(){return new g}:t,o=e.resetOnError,n=o===void 0?!0:o,i=e.resetOnComplete,a=i===void 0?!0:i,s=e.resetOnRefCountZero,p=s===void 0?!0:s;return function(c){var l,f,u,d=0,y=!1,L=!1,X=function(){f==null||f.unsubscribe(),f=void 0},te=function(){X(),l=u=void 0,y=L=!1},J=function(){var k=l;te(),k==null||k.unsubscribe()};return E(function(k,ft){d++,!L&&!y&&X();var qe=u=u!=null?u:r();ft.add(function(){d--,d===0&&!L&&!y&&(f=Ur(J,p))}),qe.subscribe(ft),!l&&d>0&&(l=new at({next:function(Fe){return qe.next(Fe)},error:function(Fe){L=!0,X(),f=Ur(te,n,Fe),qe.error(Fe)},complete:function(){y=!0,X(),f=Ur(te,a),qe.complete()}}),U(k).subscribe(l))})(c)}}function Ur(e,t){for(var r=[],o=2;oe.next(document)),e}function P(e,t=document){return Array.from(t.querySelectorAll(e))}function R(e,t=document){let r=fe(e,t);if(typeof r=="undefined")throw new ReferenceError(`Missing element: expected "${e}" to be present`);return r}function fe(e,t=document){return t.querySelector(e)||void 0}function Ie(){var e,t,r,o;return(o=(r=(t=(e=document.activeElement)==null?void 0:e.shadowRoot)==null?void 0:t.activeElement)!=null?r:document.activeElement)!=null?o:void 0}var wa=O(h(document.body,"focusin"),h(document.body,"focusout")).pipe(_e(1),Q(void 0),m(()=>Ie()||document.body),G(1));function et(e){return wa.pipe(m(t=>e.contains(t)),K())}function $t(e,t){return C(()=>O(h(e,"mouseenter").pipe(m(()=>!0)),h(e,"mouseleave").pipe(m(()=>!1))).pipe(t?Ht(r=>Le(+!r*t)):le,Q(e.matches(":hover"))))}function Jo(e,t){if(typeof t=="string"||typeof t=="number")e.innerHTML+=t.toString();else if(t instanceof Node)e.appendChild(t);else if(Array.isArray(t))for(let r of t)Jo(e,r)}function x(e,t,...r){let o=document.createElement(e);if(t)for(let n of Object.keys(t))typeof t[n]!="undefined"&&(typeof t[n]!="boolean"?o.setAttribute(n,t[n]):o.setAttribute(n,""));for(let n of r)Jo(o,n);return o}function sr(e){if(e>999){let t=+((e-950)%1e3>99);return`${((e+1e-6)/1e3).toFixed(t)}k`}else return e.toString()}function Tt(e){let t=x("script",{src:e});return C(()=>(document.head.appendChild(t),O(h(t,"load"),h(t,"error").pipe(v(()=>$r(()=>new ReferenceError(`Invalid script: ${e}`))))).pipe(m(()=>{}),_(()=>document.head.removeChild(t)),Te(1))))}var Xo=new g,Ta=C(()=>typeof ResizeObserver=="undefined"?Tt("https://unpkg.com/resize-observer-polyfill"):I(void 0)).pipe(m(()=>new ResizeObserver(e=>e.forEach(t=>Xo.next(t)))),v(e=>O(Ye,I(e)).pipe(_(()=>e.disconnect()))),G(1));function ce(e){return{width:e.offsetWidth,height:e.offsetHeight}}function ge(e){let t=e;for(;t.clientWidth===0&&t.parentElement;)t=t.parentElement;return Ta.pipe(w(r=>r.observe(t)),v(r=>Xo.pipe(b(o=>o.target===t),_(()=>r.unobserve(t)))),m(()=>ce(e)),Q(ce(e)))}function St(e){return{width:e.scrollWidth,height:e.scrollHeight}}function cr(e){let t=e.parentElement;for(;t&&(e.scrollWidth<=t.scrollWidth&&e.scrollHeight<=t.scrollHeight);)t=(e=t).parentElement;return t?e:void 0}function Zo(e){let t=[],r=e.parentElement;for(;r;)(e.clientWidth>r.clientWidth||e.clientHeight>r.clientHeight)&&t.push(r),r=(e=r).parentElement;return t.length===0&&t.push(document.documentElement),t}function Ve(e){return{x:e.offsetLeft,y:e.offsetTop}}function en(e){let t=e.getBoundingClientRect();return{x:t.x+window.scrollX,y:t.y+window.scrollY}}function tn(e){return O(h(window,"load"),h(window,"resize")).pipe(Me(0,me),m(()=>Ve(e)),Q(Ve(e)))}function pr(e){return{x:e.scrollLeft,y:e.scrollTop}}function Ne(e){return O(h(e,"scroll"),h(window,"scroll"),h(window,"resize")).pipe(Me(0,me),m(()=>pr(e)),Q(pr(e)))}var rn=new g,Sa=C(()=>I(new IntersectionObserver(e=>{for(let t of e)rn.next(t)},{threshold:0}))).pipe(v(e=>O(Ye,I(e)).pipe(_(()=>e.disconnect()))),G(1));function tt(e){return Sa.pipe(w(t=>t.observe(e)),v(t=>rn.pipe(b(({target:r})=>r===e),_(()=>t.unobserve(e)),m(({isIntersecting:r})=>r))))}function on(e,t=16){return Ne(e).pipe(m(({y:r})=>{let o=ce(e),n=St(e);return r>=n.height-o.height-t}),K())}var lr={drawer:R("[data-md-toggle=drawer]"),search:R("[data-md-toggle=search]")};function nn(e){return lr[e].checked}function Je(e,t){lr[e].checked!==t&&lr[e].click()}function ze(e){let t=lr[e];return h(t,"change").pipe(m(()=>t.checked),Q(t.checked))}function Oa(e,t){switch(e.constructor){case HTMLInputElement:return e.type==="radio"?/^Arrow/.test(t):!0;case HTMLSelectElement:case HTMLTextAreaElement:return!0;default:return e.isContentEditable}}function La(){return O(h(window,"compositionstart").pipe(m(()=>!0)),h(window,"compositionend").pipe(m(()=>!1))).pipe(Q(!1))}function an(){let e=h(window,"keydown").pipe(b(t=>!(t.metaKey||t.ctrlKey)),m(t=>({mode:nn("search")?"search":"global",type:t.key,claim(){t.preventDefault(),t.stopPropagation()}})),b(({mode:t,type:r})=>{if(t==="global"){let o=Ie();if(typeof o!="undefined")return!Oa(o,r)}return!0}),pe());return La().pipe(v(t=>t?S:e))}function ye(){return new URL(location.href)}function lt(e,t=!1){if(B("navigation.instant")&&!t){let r=x("a",{href:e.href});document.body.appendChild(r),r.click(),r.remove()}else location.href=e.href}function sn(){return new g}function cn(){return location.hash.slice(1)}function pn(e){let t=x("a",{href:e});t.addEventListener("click",r=>r.stopPropagation()),t.click()}function Ma(e){return O(h(window,"hashchange"),e).pipe(m(cn),Q(cn()),b(t=>t.length>0),G(1))}function ln(e){return Ma(e).pipe(m(t=>fe(`[id="${t}"]`)),b(t=>typeof t!="undefined"))}function Pt(e){let t=matchMedia(e);return ar(r=>t.addListener(()=>r(t.matches))).pipe(Q(t.matches))}function mn(){let e=matchMedia("print");return O(h(window,"beforeprint").pipe(m(()=>!0)),h(window,"afterprint").pipe(m(()=>!1))).pipe(Q(e.matches))}function Nr(e,t){return e.pipe(v(r=>r?t():S))}function zr(e,t){return new j(r=>{let o=new XMLHttpRequest;return o.open("GET",`${e}`),o.responseType="blob",o.addEventListener("load",()=>{o.status>=200&&o.status<300?(r.next(o.response),r.complete()):r.error(new Error(o.statusText))}),o.addEventListener("error",()=>{r.error(new Error("Network error"))}),o.addEventListener("abort",()=>{r.complete()}),typeof(t==null?void 0:t.progress$)!="undefined"&&(o.addEventListener("progress",n=>{var i;if(n.lengthComputable)t.progress$.next(n.loaded/n.total*100);else{let a=(i=o.getResponseHeader("Content-Length"))!=null?i:0;t.progress$.next(n.loaded/+a*100)}}),t.progress$.next(5)),o.send(),()=>o.abort()})}function je(e,t){return zr(e,t).pipe(v(r=>r.text()),m(r=>JSON.parse(r)),G(1))}function fn(e,t){let r=new DOMParser;return zr(e,t).pipe(v(o=>o.text()),m(o=>r.parseFromString(o,"text/html")),G(1))}function un(e,t){let r=new DOMParser;return zr(e,t).pipe(v(o=>o.text()),m(o=>r.parseFromString(o,"text/xml")),G(1))}function dn(){return{x:Math.max(0,scrollX),y:Math.max(0,scrollY)}}function hn(){return O(h(window,"scroll",{passive:!0}),h(window,"resize",{passive:!0})).pipe(m(dn),Q(dn()))}function bn(){return{width:innerWidth,height:innerHeight}}function vn(){return h(window,"resize",{passive:!0}).pipe(m(bn),Q(bn()))}function gn(){return z([hn(),vn()]).pipe(m(([e,t])=>({offset:e,size:t})),G(1))}function mr(e,{viewport$:t,header$:r}){let o=t.pipe(ee("size")),n=z([o,r]).pipe(m(()=>Ve(e)));return z([r,t,n]).pipe(m(([{height:i},{offset:a,size:s},{x:p,y:c}])=>({offset:{x:a.x-p,y:a.y-c+i},size:s})))}function _a(e){return h(e,"message",t=>t.data)}function Aa(e){let t=new g;return t.subscribe(r=>e.postMessage(r)),t}function yn(e,t=new Worker(e)){let r=_a(t),o=Aa(t),n=new g;n.subscribe(o);let i=o.pipe(Z(),ie(!0));return n.pipe(Z(),Re(r.pipe(W(i))),pe())}var Ca=R("#__config"),Ot=JSON.parse(Ca.textContent);Ot.base=`${new URL(Ot.base,ye())}`;function xe(){return Ot}function B(e){return Ot.features.includes(e)}function Ee(e,t){return typeof t!="undefined"?Ot.translations[e].replace("#",t.toString()):Ot.translations[e]}function Se(e,t=document){return R(`[data-md-component=${e}]`,t)}function ae(e,t=document){return P(`[data-md-component=${e}]`,t)}function ka(e){let t=R(".md-typeset > :first-child",e);return h(t,"click",{once:!0}).pipe(m(()=>R(".md-typeset",e)),m(r=>({hash:__md_hash(r.innerHTML)})))}function xn(e){if(!B("announce.dismiss")||!e.childElementCount)return S;if(!e.hidden){let t=R(".md-typeset",e);__md_hash(t.innerHTML)===__md_get("__announce")&&(e.hidden=!0)}return C(()=>{let t=new g;return t.subscribe(({hash:r})=>{e.hidden=!0,__md_set("__announce",r)}),ka(e).pipe(w(r=>t.next(r)),_(()=>t.complete()),m(r=>$({ref:e},r)))})}function Ha(e,{target$:t}){return t.pipe(m(r=>({hidden:r!==e})))}function En(e,t){let r=new g;return r.subscribe(({hidden:o})=>{e.hidden=o}),Ha(e,t).pipe(w(o=>r.next(o)),_(()=>r.complete()),m(o=>$({ref:e},o)))}function Rt(e,t){return t==="inline"?x("div",{class:"md-tooltip md-tooltip--inline",id:e,role:"tooltip"},x("div",{class:"md-tooltip__inner md-typeset"})):x("div",{class:"md-tooltip",id:e,role:"tooltip"},x("div",{class:"md-tooltip__inner md-typeset"}))}function wn(...e){return x("div",{class:"md-tooltip2",role:"tooltip"},x("div",{class:"md-tooltip2__inner md-typeset"},e))}function Tn(e,t){if(t=t?`${t}_annotation_${e}`:void 0,t){let r=t?`#${t}`:void 0;return x("aside",{class:"md-annotation",tabIndex:0},Rt(t),x("a",{href:r,class:"md-annotation__index",tabIndex:-1},x("span",{"data-md-annotation-id":e})))}else return x("aside",{class:"md-annotation",tabIndex:0},Rt(t),x("span",{class:"md-annotation__index",tabIndex:-1},x("span",{"data-md-annotation-id":e})))}function Sn(e){return x("button",{class:"md-clipboard md-icon",title:Ee("clipboard.copy"),"data-clipboard-target":`#${e} > code`})}var Ln=Mt(qr());function Qr(e,t){let r=t&2,o=t&1,n=Object.keys(e.terms).filter(p=>!e.terms[p]).reduce((p,c)=>[...p,x("del",null,(0,Ln.default)(c))," "],[]).slice(0,-1),i=xe(),a=new URL(e.location,i.base);B("search.highlight")&&a.searchParams.set("h",Object.entries(e.terms).filter(([,p])=>p).reduce((p,[c])=>`${p} ${c}`.trim(),""));let{tags:s}=xe();return x("a",{href:`${a}`,class:"md-search-result__link",tabIndex:-1},x("article",{class:"md-search-result__article md-typeset","data-md-score":e.score.toFixed(2)},r>0&&x("div",{class:"md-search-result__icon md-icon"}),r>0&&x("h1",null,e.title),r<=0&&x("h2",null,e.title),o>0&&e.text.length>0&&e.text,e.tags&&x("nav",{class:"md-tags"},e.tags.map(p=>{let c=s?p in s?`md-tag-icon md-tag--${s[p]}`:"md-tag-icon":"";return x("span",{class:`md-tag ${c}`},p)})),o>0&&n.length>0&&x("p",{class:"md-search-result__terms"},Ee("search.result.term.missing"),": ",...n)))}function Mn(e){let t=e[0].score,r=[...e],o=xe(),n=r.findIndex(l=>!`${new URL(l.location,o.base)}`.includes("#")),[i]=r.splice(n,1),a=r.findIndex(l=>l.scoreQr(l,1)),...p.length?[x("details",{class:"md-search-result__more"},x("summary",{tabIndex:-1},x("div",null,p.length>0&&p.length===1?Ee("search.result.more.one"):Ee("search.result.more.other",p.length))),...p.map(l=>Qr(l,1)))]:[]];return x("li",{class:"md-search-result__item"},c)}function _n(e){return x("ul",{class:"md-source__facts"},Object.entries(e).map(([t,r])=>x("li",{class:`md-source__fact md-source__fact--${t}`},typeof r=="number"?sr(r):r)))}function Kr(e){let t=`tabbed-control tabbed-control--${e}`;return x("div",{class:t,hidden:!0},x("button",{class:"tabbed-button",tabIndex:-1,"aria-hidden":"true"}))}function An(e){return x("div",{class:"md-typeset__scrollwrap"},x("div",{class:"md-typeset__table"},e))}function Ra(e){var o;let t=xe(),r=new URL(`../${e.version}/`,t.base);return x("li",{class:"md-version__item"},x("a",{href:`${r}`,class:"md-version__link"},e.title,((o=t.version)==null?void 0:o.alias)&&e.aliases.length>0&&x("span",{class:"md-version__alias"},e.aliases[0])))}function Cn(e,t){var o;let r=xe();return e=e.filter(n=>{var i;return!((i=n.properties)!=null&&i.hidden)}),x("div",{class:"md-version"},x("button",{class:"md-version__current","aria-label":Ee("select.version")},t.title,((o=r.version)==null?void 0:o.alias)&&t.aliases.length>0&&x("span",{class:"md-version__alias"},t.aliases[0])),x("ul",{class:"md-version__list"},e.map(Ra)))}var Ia=0;function ja(e){let t=z([et(e),$t(e)]).pipe(m(([o,n])=>o||n),K()),r=C(()=>Zo(e)).pipe(ne(Ne),pt(1),He(t),m(()=>en(e)));return t.pipe(Ae(o=>o),v(()=>z([t,r])),m(([o,n])=>({active:o,offset:n})),pe())}function Fa(e,t){let{content$:r,viewport$:o}=t,n=`__tooltip2_${Ia++}`;return C(()=>{let i=new g,a=new _r(!1);i.pipe(Z(),ie(!1)).subscribe(a);let s=a.pipe(Ht(c=>Le(+!c*250,kr)),K(),v(c=>c?r:S),w(c=>c.id=n),pe());z([i.pipe(m(({active:c})=>c)),s.pipe(v(c=>$t(c,250)),Q(!1))]).pipe(m(c=>c.some(l=>l))).subscribe(a);let p=a.pipe(b(c=>c),re(s,o),m(([c,l,{size:f}])=>{let u=e.getBoundingClientRect(),d=u.width/2;if(l.role==="tooltip")return{x:d,y:8+u.height};if(u.y>=f.height/2){let{height:y}=ce(l);return{x:d,y:-16-y}}else return{x:d,y:16+u.height}}));return z([s,i,p]).subscribe(([c,{offset:l},f])=>{c.style.setProperty("--md-tooltip-host-x",`${l.x}px`),c.style.setProperty("--md-tooltip-host-y",`${l.y}px`),c.style.setProperty("--md-tooltip-x",`${f.x}px`),c.style.setProperty("--md-tooltip-y",`${f.y}px`),c.classList.toggle("md-tooltip2--top",f.y<0),c.classList.toggle("md-tooltip2--bottom",f.y>=0)}),a.pipe(b(c=>c),re(s,(c,l)=>l),b(c=>c.role==="tooltip")).subscribe(c=>{let l=ce(R(":scope > *",c));c.style.setProperty("--md-tooltip-width",`${l.width}px`),c.style.setProperty("--md-tooltip-tail","0px")}),a.pipe(K(),ve(me),re(s)).subscribe(([c,l])=>{l.classList.toggle("md-tooltip2--active",c)}),z([a.pipe(b(c=>c)),s]).subscribe(([c,l])=>{l.role==="dialog"?(e.setAttribute("aria-controls",n),e.setAttribute("aria-haspopup","dialog")):e.setAttribute("aria-describedby",n)}),a.pipe(b(c=>!c)).subscribe(()=>{e.removeAttribute("aria-controls"),e.removeAttribute("aria-describedby"),e.removeAttribute("aria-haspopup")}),ja(e).pipe(w(c=>i.next(c)),_(()=>i.complete()),m(c=>$({ref:e},c)))})}function mt(e,{viewport$:t},r=document.body){return Fa(e,{content$:new j(o=>{let n=e.title,i=wn(n);return o.next(i),e.removeAttribute("title"),r.append(i),()=>{i.remove(),e.setAttribute("title",n)}}),viewport$:t})}function Ua(e,t){let r=C(()=>z([tn(e),Ne(t)])).pipe(m(([{x:o,y:n},i])=>{let{width:a,height:s}=ce(e);return{x:o-i.x+a/2,y:n-i.y+s/2}}));return et(e).pipe(v(o=>r.pipe(m(n=>({active:o,offset:n})),Te(+!o||1/0))))}function kn(e,t,{target$:r}){let[o,n]=Array.from(e.children);return C(()=>{let i=new g,a=i.pipe(Z(),ie(!0));return i.subscribe({next({offset:s}){e.style.setProperty("--md-tooltip-x",`${s.x}px`),e.style.setProperty("--md-tooltip-y",`${s.y}px`)},complete(){e.style.removeProperty("--md-tooltip-x"),e.style.removeProperty("--md-tooltip-y")}}),tt(e).pipe(W(a)).subscribe(s=>{e.toggleAttribute("data-md-visible",s)}),O(i.pipe(b(({active:s})=>s)),i.pipe(_e(250),b(({active:s})=>!s))).subscribe({next({active:s}){s?e.prepend(o):o.remove()},complete(){e.prepend(o)}}),i.pipe(Me(16,me)).subscribe(({active:s})=>{o.classList.toggle("md-tooltip--active",s)}),i.pipe(pt(125,me),b(()=>!!e.offsetParent),m(()=>e.offsetParent.getBoundingClientRect()),m(({x:s})=>s)).subscribe({next(s){s?e.style.setProperty("--md-tooltip-0",`${-s}px`):e.style.removeProperty("--md-tooltip-0")},complete(){e.style.removeProperty("--md-tooltip-0")}}),h(n,"click").pipe(W(a),b(s=>!(s.metaKey||s.ctrlKey))).subscribe(s=>{s.stopPropagation(),s.preventDefault()}),h(n,"mousedown").pipe(W(a),re(i)).subscribe(([s,{active:p}])=>{var c;if(s.button!==0||s.metaKey||s.ctrlKey)s.preventDefault();else if(p){s.preventDefault();let l=e.parentElement.closest(".md-annotation");l instanceof HTMLElement?l.focus():(c=Ie())==null||c.blur()}}),r.pipe(W(a),b(s=>s===o),Ge(125)).subscribe(()=>e.focus()),Ua(e,t).pipe(w(s=>i.next(s)),_(()=>i.complete()),m(s=>$({ref:e},s)))})}function Wa(e){return e.tagName==="CODE"?P(".c, .c1, .cm",e):[e]}function Da(e){let t=[];for(let r of Wa(e)){let o=[],n=document.createNodeIterator(r,NodeFilter.SHOW_TEXT);for(let i=n.nextNode();i;i=n.nextNode())o.push(i);for(let i of o){let a;for(;a=/(\(\d+\))(!)?/.exec(i.textContent);){let[,s,p]=a;if(typeof p=="undefined"){let c=i.splitText(a.index);i=c.splitText(s.length),t.push(c)}else{i.textContent=s,t.push(i);break}}}}return t}function Hn(e,t){t.append(...Array.from(e.childNodes))}function fr(e,t,{target$:r,print$:o}){let n=t.closest("[id]"),i=n==null?void 0:n.id,a=new Map;for(let s of Da(t)){let[,p]=s.textContent.match(/\((\d+)\)/);fe(`:scope > li:nth-child(${p})`,e)&&(a.set(p,Tn(p,i)),s.replaceWith(a.get(p)))}return a.size===0?S:C(()=>{let s=new g,p=s.pipe(Z(),ie(!0)),c=[];for(let[l,f]of a)c.push([R(".md-typeset",f),R(`:scope > li:nth-child(${l})`,e)]);return o.pipe(W(p)).subscribe(l=>{e.hidden=!l,e.classList.toggle("md-annotation-list",l);for(let[f,u]of c)l?Hn(f,u):Hn(u,f)}),O(...[...a].map(([,l])=>kn(l,t,{target$:r}))).pipe(_(()=>s.complete()),pe())})}function $n(e){if(e.nextElementSibling){let t=e.nextElementSibling;if(t.tagName==="OL")return t;if(t.tagName==="P"&&!t.children.length)return $n(t)}}function Pn(e,t){return C(()=>{let r=$n(e);return typeof r!="undefined"?fr(r,e,t):S})}var Rn=Mt(Br());var Va=0;function In(e){if(e.nextElementSibling){let t=e.nextElementSibling;if(t.tagName==="OL")return t;if(t.tagName==="P"&&!t.children.length)return In(t)}}function Na(e){return ge(e).pipe(m(({width:t})=>({scrollable:St(e).width>t})),ee("scrollable"))}function jn(e,t){let{matches:r}=matchMedia("(hover)"),o=C(()=>{let n=new g,i=n.pipe(jr(1));n.subscribe(({scrollable:c})=>{c&&r?e.setAttribute("tabindex","0"):e.removeAttribute("tabindex")});let a=[];if(Rn.default.isSupported()&&(e.closest(".copy")||B("content.code.copy")&&!e.closest(".no-copy"))){let c=e.closest("pre");c.id=`__code_${Va++}`;let l=Sn(c.id);c.insertBefore(l,e),B("content.tooltips")&&a.push(mt(l,{viewport$}))}let s=e.closest(".highlight");if(s instanceof HTMLElement){let c=In(s);if(typeof c!="undefined"&&(s.classList.contains("annotate")||B("content.code.annotate"))){let l=fr(c,e,t);a.push(ge(s).pipe(W(i),m(({width:f,height:u})=>f&&u),K(),v(f=>f?l:S)))}}return P(":scope > span[id]",e).length&&e.classList.add("md-code__content"),Na(e).pipe(w(c=>n.next(c)),_(()=>n.complete()),m(c=>$({ref:e},c)),Re(...a))});return B("content.lazy")?tt(e).pipe(b(n=>n),Te(1),v(()=>o)):o}function za(e,{target$:t,print$:r}){let o=!0;return O(t.pipe(m(n=>n.closest("details:not([open])")),b(n=>e===n),m(()=>({action:"open",reveal:!0}))),r.pipe(b(n=>n||!o),w(()=>o=e.open),m(n=>({action:n?"open":"close"}))))}function Fn(e,t){return C(()=>{let r=new g;return r.subscribe(({action:o,reveal:n})=>{e.toggleAttribute("open",o==="open"),n&&e.scrollIntoView()}),za(e,t).pipe(w(o=>r.next(o)),_(()=>r.complete()),m(o=>$({ref:e},o)))})}var Un=".node circle,.node ellipse,.node path,.node polygon,.node rect{fill:var(--md-mermaid-node-bg-color);stroke:var(--md-mermaid-node-fg-color)}marker{fill:var(--md-mermaid-edge-color)!important}.edgeLabel .label rect{fill:#0000}.label{color:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}.label foreignObject{line-height:normal;overflow:visible}.label div .edgeLabel{color:var(--md-mermaid-label-fg-color)}.edgeLabel,.edgeLabel p,.label div .edgeLabel{background-color:var(--md-mermaid-label-bg-color)}.edgeLabel,.edgeLabel p{fill:var(--md-mermaid-label-bg-color);color:var(--md-mermaid-edge-color)}.edgePath .path,.flowchart-link{stroke:var(--md-mermaid-edge-color);stroke-width:.05rem}.edgePath .arrowheadPath{fill:var(--md-mermaid-edge-color);stroke:none}.cluster rect{fill:var(--md-default-fg-color--lightest);stroke:var(--md-default-fg-color--lighter)}.cluster span{color:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}g #flowchart-circleEnd,g #flowchart-circleStart,g #flowchart-crossEnd,g #flowchart-crossStart,g #flowchart-pointEnd,g #flowchart-pointStart{stroke:none}g.classGroup line,g.classGroup rect{fill:var(--md-mermaid-node-bg-color);stroke:var(--md-mermaid-node-fg-color)}g.classGroup text{fill:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}.classLabel .box{fill:var(--md-mermaid-label-bg-color);background-color:var(--md-mermaid-label-bg-color);opacity:1}.classLabel .label{fill:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}.node .divider{stroke:var(--md-mermaid-node-fg-color)}.relation{stroke:var(--md-mermaid-edge-color)}.cardinality{fill:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}.cardinality text{fill:inherit!important}defs #classDiagram-compositionEnd,defs #classDiagram-compositionStart,defs #classDiagram-dependencyEnd,defs #classDiagram-dependencyStart,defs #classDiagram-extensionEnd,defs #classDiagram-extensionStart{fill:var(--md-mermaid-edge-color)!important;stroke:var(--md-mermaid-edge-color)!important}defs #classDiagram-aggregationEnd,defs #classDiagram-aggregationStart{fill:var(--md-mermaid-label-bg-color)!important;stroke:var(--md-mermaid-edge-color)!important}g.stateGroup rect{fill:var(--md-mermaid-node-bg-color);stroke:var(--md-mermaid-node-fg-color)}g.stateGroup .state-title{fill:var(--md-mermaid-label-fg-color)!important;font-family:var(--md-mermaid-font-family)}g.stateGroup .composit{fill:var(--md-mermaid-label-bg-color)}.nodeLabel,.nodeLabel p{color:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}a .nodeLabel{text-decoration:underline}.node circle.state-end,.node circle.state-start,.start-state{fill:var(--md-mermaid-edge-color);stroke:none}.end-state-inner,.end-state-outer{fill:var(--md-mermaid-edge-color)}.end-state-inner,.node circle.state-end{stroke:var(--md-mermaid-label-bg-color)}.transition{stroke:var(--md-mermaid-edge-color)}[id^=state-fork] rect,[id^=state-join] rect{fill:var(--md-mermaid-edge-color)!important;stroke:none!important}.statediagram-cluster.statediagram-cluster .inner{fill:var(--md-default-bg-color)}.statediagram-cluster rect{fill:var(--md-mermaid-node-bg-color);stroke:var(--md-mermaid-node-fg-color)}.statediagram-state rect.divider{fill:var(--md-default-fg-color--lightest);stroke:var(--md-default-fg-color--lighter)}defs #statediagram-barbEnd{stroke:var(--md-mermaid-edge-color)}.attributeBoxEven,.attributeBoxOdd{fill:var(--md-mermaid-node-bg-color);stroke:var(--md-mermaid-node-fg-color)}.entityBox{fill:var(--md-mermaid-label-bg-color);stroke:var(--md-mermaid-node-fg-color)}.entityLabel{fill:var(--md-mermaid-label-fg-color);font-family:var(--md-mermaid-font-family)}.relationshipLabelBox{fill:var(--md-mermaid-label-bg-color);fill-opacity:1;background-color:var(--md-mermaid-label-bg-color);opacity:1}.relationshipLabel{fill:var(--md-mermaid-label-fg-color)}.relationshipLine{stroke:var(--md-mermaid-edge-color)}defs #ONE_OR_MORE_END *,defs #ONE_OR_MORE_START *,defs #ONLY_ONE_END *,defs #ONLY_ONE_START *,defs #ZERO_OR_MORE_END *,defs #ZERO_OR_MORE_START *,defs #ZERO_OR_ONE_END *,defs #ZERO_OR_ONE_START *{stroke:var(--md-mermaid-edge-color)!important}defs #ZERO_OR_MORE_END circle,defs #ZERO_OR_MORE_START circle{fill:var(--md-mermaid-label-bg-color)}.actor{fill:var(--md-mermaid-sequence-actor-bg-color);stroke:var(--md-mermaid-sequence-actor-border-color)}text.actor>tspan{fill:var(--md-mermaid-sequence-actor-fg-color);font-family:var(--md-mermaid-font-family)}line{stroke:var(--md-mermaid-sequence-actor-line-color)}.actor-man circle,.actor-man line{fill:var(--md-mermaid-sequence-actorman-bg-color);stroke:var(--md-mermaid-sequence-actorman-line-color)}.messageLine0,.messageLine1{stroke:var(--md-mermaid-sequence-message-line-color)}.note{fill:var(--md-mermaid-sequence-note-bg-color);stroke:var(--md-mermaid-sequence-note-border-color)}.loopText,.loopText>tspan,.messageText,.noteText>tspan{stroke:none;font-family:var(--md-mermaid-font-family)!important}.messageText{fill:var(--md-mermaid-sequence-message-fg-color)}.loopText,.loopText>tspan{fill:var(--md-mermaid-sequence-loop-fg-color)}.noteText>tspan{fill:var(--md-mermaid-sequence-note-fg-color)}#arrowhead path{fill:var(--md-mermaid-sequence-message-line-color);stroke:none}.loopLine{fill:var(--md-mermaid-sequence-loop-bg-color);stroke:var(--md-mermaid-sequence-loop-border-color)}.labelBox{fill:var(--md-mermaid-sequence-label-bg-color);stroke:none}.labelText,.labelText>span{fill:var(--md-mermaid-sequence-label-fg-color);font-family:var(--md-mermaid-font-family)}.sequenceNumber{fill:var(--md-mermaid-sequence-number-fg-color)}rect.rect{fill:var(--md-mermaid-sequence-box-bg-color);stroke:none}rect.rect+text.text{fill:var(--md-mermaid-sequence-box-fg-color)}defs #sequencenumber{fill:var(--md-mermaid-sequence-number-bg-color)!important}";var Gr,Qa=0;function Ka(){return typeof mermaid=="undefined"||mermaid instanceof Element?Tt("https://unpkg.com/mermaid@11/dist/mermaid.min.js"):I(void 0)}function Wn(e){return e.classList.remove("mermaid"),Gr||(Gr=Ka().pipe(w(()=>mermaid.initialize({startOnLoad:!1,themeCSS:Un,sequence:{actorFontSize:"16px",messageFontSize:"16px",noteFontSize:"16px"}})),m(()=>{}),G(1))),Gr.subscribe(()=>co(this,null,function*(){e.classList.add("mermaid");let t=`__mermaid_${Qa++}`,r=x("div",{class:"mermaid"}),o=e.textContent,{svg:n,fn:i}=yield mermaid.render(t,o),a=r.attachShadow({mode:"closed"});a.innerHTML=n,e.replaceWith(r),i==null||i(a)})),Gr.pipe(m(()=>({ref:e})))}var Dn=x("table");function Vn(e){return e.replaceWith(Dn),Dn.replaceWith(An(e)),I({ref:e})}function Ya(e){let t=e.find(r=>r.checked)||e[0];return O(...e.map(r=>h(r,"change").pipe(m(()=>R(`label[for="${r.id}"]`))))).pipe(Q(R(`label[for="${t.id}"]`)),m(r=>({active:r})))}function Nn(e,{viewport$:t,target$:r}){let o=R(".tabbed-labels",e),n=P(":scope > input",e),i=Kr("prev");e.append(i);let a=Kr("next");return e.append(a),C(()=>{let s=new g,p=s.pipe(Z(),ie(!0));z([s,ge(e),tt(e)]).pipe(W(p),Me(1,me)).subscribe({next([{active:c},l]){let f=Ve(c),{width:u}=ce(c);e.style.setProperty("--md-indicator-x",`${f.x}px`),e.style.setProperty("--md-indicator-width",`${u}px`);let d=pr(o);(f.xd.x+l.width)&&o.scrollTo({left:Math.max(0,f.x-16),behavior:"smooth"})},complete(){e.style.removeProperty("--md-indicator-x"),e.style.removeProperty("--md-indicator-width")}}),z([Ne(o),ge(o)]).pipe(W(p)).subscribe(([c,l])=>{let f=St(o);i.hidden=c.x<16,a.hidden=c.x>f.width-l.width-16}),O(h(i,"click").pipe(m(()=>-1)),h(a,"click").pipe(m(()=>1))).pipe(W(p)).subscribe(c=>{let{width:l}=ce(o);o.scrollBy({left:l*c,behavior:"smooth"})}),r.pipe(W(p),b(c=>n.includes(c))).subscribe(c=>c.click()),o.classList.add("tabbed-labels--linked");for(let c of n){let l=R(`label[for="${c.id}"]`);l.replaceChildren(x("a",{href:`#${l.htmlFor}`,tabIndex:-1},...Array.from(l.childNodes))),h(l.firstElementChild,"click").pipe(W(p),b(f=>!(f.metaKey||f.ctrlKey)),w(f=>{f.preventDefault(),f.stopPropagation()})).subscribe(()=>{history.replaceState({},"",`#${l.htmlFor}`),l.click()})}return B("content.tabs.link")&&s.pipe(Ce(1),re(t)).subscribe(([{active:c},{offset:l}])=>{let f=c.innerText.trim();if(c.hasAttribute("data-md-switching"))c.removeAttribute("data-md-switching");else{let u=e.offsetTop-l.y;for(let y of P("[data-tabs]"))for(let L of P(":scope > input",y)){let X=R(`label[for="${L.id}"]`);if(X!==c&&X.innerText.trim()===f){X.setAttribute("data-md-switching",""),L.click();break}}window.scrollTo({top:e.offsetTop-u});let d=__md_get("__tabs")||[];__md_set("__tabs",[...new Set([f,...d])])}}),s.pipe(W(p)).subscribe(()=>{for(let c of P("audio, video",e))c.pause()}),Ya(n).pipe(w(c=>s.next(c)),_(()=>s.complete()),m(c=>$({ref:e},c)))}).pipe(Ke(se))}function zn(e,{viewport$:t,target$:r,print$:o}){return O(...P(".annotate:not(.highlight)",e).map(n=>Pn(n,{target$:r,print$:o})),...P("pre:not(.mermaid) > code",e).map(n=>jn(n,{target$:r,print$:o})),...P("pre.mermaid",e).map(n=>Wn(n)),...P("table:not([class])",e).map(n=>Vn(n)),...P("details",e).map(n=>Fn(n,{target$:r,print$:o})),...P("[data-tabs]",e).map(n=>Nn(n,{viewport$:t,target$:r})),...P("[title]",e).filter(()=>B("content.tooltips")).map(n=>mt(n,{viewport$:t})))}function Ba(e,{alert$:t}){return t.pipe(v(r=>O(I(!0),I(!1).pipe(Ge(2e3))).pipe(m(o=>({message:r,active:o})))))}function qn(e,t){let r=R(".md-typeset",e);return C(()=>{let o=new g;return o.subscribe(({message:n,active:i})=>{e.classList.toggle("md-dialog--active",i),r.textContent=n}),Ba(e,t).pipe(w(n=>o.next(n)),_(()=>o.complete()),m(n=>$({ref:e},n)))})}var Ga=0;function Ja(e,t){document.body.append(e);let{width:r}=ce(e);e.style.setProperty("--md-tooltip-width",`${r}px`),e.remove();let o=cr(t),n=typeof o!="undefined"?Ne(o):I({x:0,y:0}),i=O(et(t),$t(t)).pipe(K());return z([i,n]).pipe(m(([a,s])=>{let{x:p,y:c}=Ve(t),l=ce(t),f=t.closest("table");return f&&t.parentElement&&(p+=f.offsetLeft+t.parentElement.offsetLeft,c+=f.offsetTop+t.parentElement.offsetTop),{active:a,offset:{x:p-s.x+l.width/2-r/2,y:c-s.y+l.height+8}}}))}function Qn(e){let t=e.title;if(!t.length)return S;let r=`__tooltip_${Ga++}`,o=Rt(r,"inline"),n=R(".md-typeset",o);return n.innerHTML=t,C(()=>{let i=new g;return i.subscribe({next({offset:a}){o.style.setProperty("--md-tooltip-x",`${a.x}px`),o.style.setProperty("--md-tooltip-y",`${a.y}px`)},complete(){o.style.removeProperty("--md-tooltip-x"),o.style.removeProperty("--md-tooltip-y")}}),O(i.pipe(b(({active:a})=>a)),i.pipe(_e(250),b(({active:a})=>!a))).subscribe({next({active:a}){a?(e.insertAdjacentElement("afterend",o),e.setAttribute("aria-describedby",r),e.removeAttribute("title")):(o.remove(),e.removeAttribute("aria-describedby"),e.setAttribute("title",t))},complete(){o.remove(),e.removeAttribute("aria-describedby"),e.setAttribute("title",t)}}),i.pipe(Me(16,me)).subscribe(({active:a})=>{o.classList.toggle("md-tooltip--active",a)}),i.pipe(pt(125,me),b(()=>!!e.offsetParent),m(()=>e.offsetParent.getBoundingClientRect()),m(({x:a})=>a)).subscribe({next(a){a?o.style.setProperty("--md-tooltip-0",`${-a}px`):o.style.removeProperty("--md-tooltip-0")},complete(){o.style.removeProperty("--md-tooltip-0")}}),Ja(o,e).pipe(w(a=>i.next(a)),_(()=>i.complete()),m(a=>$({ref:e},a)))}).pipe(Ke(se))}function Xa({viewport$:e}){if(!B("header.autohide"))return I(!1);let t=e.pipe(m(({offset:{y:n}})=>n),Be(2,1),m(([n,i])=>[nMath.abs(i-n.y)>100),m(([,[n]])=>n),K()),o=ze("search");return z([e,o]).pipe(m(([{offset:n},i])=>n.y>400&&!i),K(),v(n=>n?r:I(!1)),Q(!1))}function Kn(e,t){return C(()=>z([ge(e),Xa(t)])).pipe(m(([{height:r},o])=>({height:r,hidden:o})),K((r,o)=>r.height===o.height&&r.hidden===o.hidden),G(1))}function Yn(e,{header$:t,main$:r}){return C(()=>{let o=new g,n=o.pipe(Z(),ie(!0));o.pipe(ee("active"),He(t)).subscribe(([{active:a},{hidden:s}])=>{e.classList.toggle("md-header--shadow",a&&!s),e.hidden=s});let i=ue(P("[title]",e)).pipe(b(()=>B("content.tooltips")),ne(a=>Qn(a)));return r.subscribe(o),t.pipe(W(n),m(a=>$({ref:e},a)),Re(i.pipe(W(n))))})}function Za(e,{viewport$:t,header$:r}){return mr(e,{viewport$:t,header$:r}).pipe(m(({offset:{y:o}})=>{let{height:n}=ce(e);return{active:o>=n}}),ee("active"))}function Bn(e,t){return C(()=>{let r=new g;r.subscribe({next({active:n}){e.classList.toggle("md-header__title--active",n)},complete(){e.classList.remove("md-header__title--active")}});let o=fe(".md-content h1");return typeof o=="undefined"?S:Za(o,t).pipe(w(n=>r.next(n)),_(()=>r.complete()),m(n=>$({ref:e},n)))})}function Gn(e,{viewport$:t,header$:r}){let o=r.pipe(m(({height:i})=>i),K()),n=o.pipe(v(()=>ge(e).pipe(m(({height:i})=>({top:e.offsetTop,bottom:e.offsetTop+i})),ee("bottom"))));return z([o,n,t]).pipe(m(([i,{top:a,bottom:s},{offset:{y:p},size:{height:c}}])=>(c=Math.max(0,c-Math.max(0,a-p,i)-Math.max(0,c+p-s)),{offset:a-i,height:c,active:a-i<=p})),K((i,a)=>i.offset===a.offset&&i.height===a.height&&i.active===a.active))}function es(e){let t=__md_get("__palette")||{index:e.findIndex(o=>matchMedia(o.getAttribute("data-md-color-media")).matches)},r=Math.max(0,Math.min(t.index,e.length-1));return I(...e).pipe(ne(o=>h(o,"change").pipe(m(()=>o))),Q(e[r]),m(o=>({index:e.indexOf(o),color:{media:o.getAttribute("data-md-color-media"),scheme:o.getAttribute("data-md-color-scheme"),primary:o.getAttribute("data-md-color-primary"),accent:o.getAttribute("data-md-color-accent")}})),G(1))}function Jn(e){let t=P("input",e),r=x("meta",{name:"theme-color"});document.head.appendChild(r);let o=x("meta",{name:"color-scheme"});document.head.appendChild(o);let n=Pt("(prefers-color-scheme: light)");return C(()=>{let i=new g;return i.subscribe(a=>{if(document.body.setAttribute("data-md-color-switching",""),a.color.media==="(prefers-color-scheme)"){let s=matchMedia("(prefers-color-scheme: light)"),p=document.querySelector(s.matches?"[data-md-color-media='(prefers-color-scheme: light)']":"[data-md-color-media='(prefers-color-scheme: dark)']");a.color.scheme=p.getAttribute("data-md-color-scheme"),a.color.primary=p.getAttribute("data-md-color-primary"),a.color.accent=p.getAttribute("data-md-color-accent")}for(let[s,p]of Object.entries(a.color))document.body.setAttribute(`data-md-color-${s}`,p);for(let s=0;sa.key==="Enter"),re(i,(a,s)=>s)).subscribe(({index:a})=>{a=(a+1)%t.length,t[a].click(),t[a].focus()}),i.pipe(m(()=>{let a=Se("header"),s=window.getComputedStyle(a);return o.content=s.colorScheme,s.backgroundColor.match(/\d+/g).map(p=>(+p).toString(16).padStart(2,"0")).join("")})).subscribe(a=>r.content=`#${a}`),i.pipe(ve(se)).subscribe(()=>{document.body.removeAttribute("data-md-color-switching")}),es(t).pipe(W(n.pipe(Ce(1))),ct(),w(a=>i.next(a)),_(()=>i.complete()),m(a=>$({ref:e},a)))})}function Xn(e,{progress$:t}){return C(()=>{let r=new g;return r.subscribe(({value:o})=>{e.style.setProperty("--md-progress-value",`${o}`)}),t.pipe(w(o=>r.next({value:o})),_(()=>r.complete()),m(o=>({ref:e,value:o})))})}var Jr=Mt(Br());function ts(e){e.setAttribute("data-md-copying","");let t=e.closest("[data-copy]"),r=t?t.getAttribute("data-copy"):e.innerText;return e.removeAttribute("data-md-copying"),r.trimEnd()}function Zn({alert$:e}){Jr.default.isSupported()&&new j(t=>{new Jr.default("[data-clipboard-target], [data-clipboard-text]",{text:r=>r.getAttribute("data-clipboard-text")||ts(R(r.getAttribute("data-clipboard-target")))}).on("success",r=>t.next(r))}).pipe(w(t=>{t.trigger.focus()}),m(()=>Ee("clipboard.copied"))).subscribe(e)}function ei(e,t){return e.protocol=t.protocol,e.hostname=t.hostname,e}function rs(e,t){let r=new Map;for(let o of P("url",e)){let n=R("loc",o),i=[ei(new URL(n.textContent),t)];r.set(`${i[0]}`,i);for(let a of P("[rel=alternate]",o)){let s=a.getAttribute("href");s!=null&&i.push(ei(new URL(s),t))}}return r}function ur(e){return un(new URL("sitemap.xml",e)).pipe(m(t=>rs(t,new URL(e))),de(()=>I(new Map)))}function os(e,t){if(!(e.target instanceof Element))return S;let r=e.target.closest("a");if(r===null)return S;if(r.target||e.metaKey||e.ctrlKey)return S;let o=new URL(r.href);return o.search=o.hash="",t.has(`${o}`)?(e.preventDefault(),I(new URL(r.href))):S}function ti(e){let t=new Map;for(let r of P(":scope > *",e.head))t.set(r.outerHTML,r);return t}function ri(e){for(let t of P("[href], [src]",e))for(let r of["href","src"]){let o=t.getAttribute(r);if(o&&!/^(?:[a-z]+:)?\/\//i.test(o)){t[r]=t[r];break}}return I(e)}function ns(e){for(let o of["[data-md-component=announce]","[data-md-component=container]","[data-md-component=header-topic]","[data-md-component=outdated]","[data-md-component=logo]","[data-md-component=skip]",...B("navigation.tabs.sticky")?["[data-md-component=tabs]"]:[]]){let n=fe(o),i=fe(o,e);typeof n!="undefined"&&typeof i!="undefined"&&n.replaceWith(i)}let t=ti(document);for(let[o,n]of ti(e))t.has(o)?t.delete(o):document.head.appendChild(n);for(let o of t.values()){let n=o.getAttribute("name");n!=="theme-color"&&n!=="color-scheme"&&o.remove()}let r=Se("container");return We(P("script",r)).pipe(v(o=>{let n=e.createElement("script");if(o.src){for(let i of o.getAttributeNames())n.setAttribute(i,o.getAttribute(i));return o.replaceWith(n),new j(i=>{n.onload=()=>i.complete()})}else return n.textContent=o.textContent,o.replaceWith(n),S}),Z(),ie(document))}function oi({location$:e,viewport$:t,progress$:r}){let o=xe();if(location.protocol==="file:")return S;let n=ur(o.base);I(document).subscribe(ri);let i=h(document.body,"click").pipe(He(n),v(([p,c])=>os(p,c)),pe()),a=h(window,"popstate").pipe(m(ye),pe());i.pipe(re(t)).subscribe(([p,{offset:c}])=>{history.replaceState(c,""),history.pushState(null,"",p)}),O(i,a).subscribe(e);let s=e.pipe(ee("pathname"),v(p=>fn(p,{progress$:r}).pipe(de(()=>(lt(p,!0),S)))),v(ri),v(ns),pe());return O(s.pipe(re(e,(p,c)=>c)),s.pipe(v(()=>e),ee("pathname"),v(()=>e),ee("hash")),e.pipe(K((p,c)=>p.pathname===c.pathname&&p.hash===c.hash),v(()=>i),w(()=>history.back()))).subscribe(p=>{var c,l;history.state!==null||!p.hash?window.scrollTo(0,(l=(c=history.state)==null?void 0:c.y)!=null?l:0):(history.scrollRestoration="auto",pn(p.hash),history.scrollRestoration="manual")}),e.subscribe(()=>{history.scrollRestoration="manual"}),h(window,"beforeunload").subscribe(()=>{history.scrollRestoration="auto"}),t.pipe(ee("offset"),_e(100)).subscribe(({offset:p})=>{history.replaceState(p,"")}),s}var ni=Mt(qr());function ii(e){let t=e.separator.split("|").map(n=>n.replace(/(\(\?[!=<][^)]+\))/g,"").length===0?"\uFFFD":n).join("|"),r=new RegExp(t,"img"),o=(n,i,a)=>`${i}${a}`;return n=>{n=n.replace(/[\s*+\-:~^]+/g," ").trim();let i=new RegExp(`(^|${e.separator}|)(${n.replace(/[|\\{}()[\]^$+*?.-]/g,"\\$&").replace(r,"|")})`,"img");return a=>(0,ni.default)(a).replace(i,o).replace(/<\/mark>(\s+)]*>/img,"$1")}}function jt(e){return e.type===1}function dr(e){return e.type===3}function ai(e,t){let r=yn(e);return O(I(location.protocol!=="file:"),ze("search")).pipe(Ae(o=>o),v(()=>t)).subscribe(({config:o,docs:n})=>r.next({type:0,data:{config:o,docs:n,options:{suggest:B("search.suggest")}}})),r}function si(e){var l;let{selectedVersionSitemap:t,selectedVersionBaseURL:r,currentLocation:o,currentBaseURL:n}=e,i=(l=Xr(n))==null?void 0:l.pathname;if(i===void 0)return;let a=ss(o.pathname,i);if(a===void 0)return;let s=ps(t.keys());if(!t.has(s))return;let p=Xr(a,s);if(!p||!t.has(p.href))return;let c=Xr(a,r);if(c)return c.hash=o.hash,c.search=o.search,c}function Xr(e,t){try{return new URL(e,t)}catch(r){return}}function ss(e,t){if(e.startsWith(t))return e.slice(t.length)}function cs(e,t){let r=Math.min(e.length,t.length),o;for(o=0;oS)),o=r.pipe(m(n=>{let[,i]=t.base.match(/([^/]+)\/?$/);return n.find(({version:a,aliases:s})=>a===i||s.includes(i))||n[0]}));r.pipe(m(n=>new Map(n.map(i=>[`${new URL(`../${i.version}/`,t.base)}`,i]))),v(n=>h(document.body,"click").pipe(b(i=>!i.metaKey&&!i.ctrlKey),re(o),v(([i,a])=>{if(i.target instanceof Element){let s=i.target.closest("a");if(s&&!s.target&&n.has(s.href)){let p=s.href;return!i.target.closest(".md-version")&&n.get(p)===a?S:(i.preventDefault(),I(new URL(p)))}}return S}),v(i=>ur(i).pipe(m(a=>{var s;return(s=si({selectedVersionSitemap:a,selectedVersionBaseURL:i,currentLocation:ye(),currentBaseURL:t.base}))!=null?s:i})))))).subscribe(n=>lt(n,!0)),z([r,o]).subscribe(([n,i])=>{R(".md-header__topic").appendChild(Cn(n,i))}),e.pipe(v(()=>o)).subscribe(n=>{var a;let i=__md_get("__outdated",sessionStorage);if(i===null){i=!0;let s=((a=t.version)==null?void 0:a.default)||"latest";Array.isArray(s)||(s=[s]);e:for(let p of s)for(let c of n.aliases.concat(n.version))if(new RegExp(p,"i").test(c)){i=!1;break e}__md_set("__outdated",i,sessionStorage)}if(i)for(let s of ae("outdated"))s.hidden=!1})}function ls(e,{worker$:t}){let{searchParams:r}=ye();r.has("q")&&(Je("search",!0),e.value=r.get("q"),e.focus(),ze("search").pipe(Ae(i=>!i)).subscribe(()=>{let i=ye();i.searchParams.delete("q"),history.replaceState({},"",`${i}`)}));let o=et(e),n=O(t.pipe(Ae(jt)),h(e,"keyup"),o).pipe(m(()=>e.value),K());return z([n,o]).pipe(m(([i,a])=>({value:i,focus:a})),G(1))}function pi(e,{worker$:t}){let r=new g,o=r.pipe(Z(),ie(!0));z([t.pipe(Ae(jt)),r],(i,a)=>a).pipe(ee("value")).subscribe(({value:i})=>t.next({type:2,data:i})),r.pipe(ee("focus")).subscribe(({focus:i})=>{i&&Je("search",i)}),h(e.form,"reset").pipe(W(o)).subscribe(()=>e.focus());let n=R("header [for=__search]");return h(n,"click").subscribe(()=>e.focus()),ls(e,{worker$:t}).pipe(w(i=>r.next(i)),_(()=>r.complete()),m(i=>$({ref:e},i)),G(1))}function li(e,{worker$:t,query$:r}){let o=new g,n=on(e.parentElement).pipe(b(Boolean)),i=e.parentElement,a=R(":scope > :first-child",e),s=R(":scope > :last-child",e);ze("search").subscribe(l=>s.setAttribute("role",l?"list":"presentation")),o.pipe(re(r),Wr(t.pipe(Ae(jt)))).subscribe(([{items:l},{value:f}])=>{switch(l.length){case 0:a.textContent=f.length?Ee("search.result.none"):Ee("search.result.placeholder");break;case 1:a.textContent=Ee("search.result.one");break;default:let u=sr(l.length);a.textContent=Ee("search.result.other",u)}});let p=o.pipe(w(()=>s.innerHTML=""),v(({items:l})=>O(I(...l.slice(0,10)),I(...l.slice(10)).pipe(Be(4),Vr(n),v(([f])=>f)))),m(Mn),pe());return p.subscribe(l=>s.appendChild(l)),p.pipe(ne(l=>{let f=fe("details",l);return typeof f=="undefined"?S:h(f,"toggle").pipe(W(o),m(()=>f))})).subscribe(l=>{l.open===!1&&l.offsetTop<=i.scrollTop&&i.scrollTo({top:l.offsetTop})}),t.pipe(b(dr),m(({data:l})=>l)).pipe(w(l=>o.next(l)),_(()=>o.complete()),m(l=>$({ref:e},l)))}function ms(e,{query$:t}){return t.pipe(m(({value:r})=>{let o=ye();return o.hash="",r=r.replace(/\s+/g,"+").replace(/&/g,"%26").replace(/=/g,"%3D"),o.search=`q=${r}`,{url:o}}))}function mi(e,t){let r=new g,o=r.pipe(Z(),ie(!0));return r.subscribe(({url:n})=>{e.setAttribute("data-clipboard-text",e.href),e.href=`${n}`}),h(e,"click").pipe(W(o)).subscribe(n=>n.preventDefault()),ms(e,t).pipe(w(n=>r.next(n)),_(()=>r.complete()),m(n=>$({ref:e},n)))}function fi(e,{worker$:t,keyboard$:r}){let o=new g,n=Se("search-query"),i=O(h(n,"keydown"),h(n,"focus")).pipe(ve(se),m(()=>n.value),K());return o.pipe(He(i),m(([{suggest:s},p])=>{let c=p.split(/([\s-]+)/);if(s!=null&&s.length&&c[c.length-1]){let l=s[s.length-1];l.startsWith(c[c.length-1])&&(c[c.length-1]=l)}else c.length=0;return c})).subscribe(s=>e.innerHTML=s.join("").replace(/\s/g," ")),r.pipe(b(({mode:s})=>s==="search")).subscribe(s=>{switch(s.type){case"ArrowRight":e.innerText.length&&n.selectionStart===n.value.length&&(n.value=e.innerText);break}}),t.pipe(b(dr),m(({data:s})=>s)).pipe(w(s=>o.next(s)),_(()=>o.complete()),m(()=>({ref:e})))}function ui(e,{index$:t,keyboard$:r}){let o=xe();try{let n=ai(o.search,t),i=Se("search-query",e),a=Se("search-result",e);h(e,"click").pipe(b(({target:p})=>p instanceof Element&&!!p.closest("a"))).subscribe(()=>Je("search",!1)),r.pipe(b(({mode:p})=>p==="search")).subscribe(p=>{let c=Ie();switch(p.type){case"Enter":if(c===i){let l=new Map;for(let f of P(":first-child [href]",a)){let u=f.firstElementChild;l.set(f,parseFloat(u.getAttribute("data-md-score")))}if(l.size){let[[f]]=[...l].sort(([,u],[,d])=>d-u);f.click()}p.claim()}break;case"Escape":case"Tab":Je("search",!1),i.blur();break;case"ArrowUp":case"ArrowDown":if(typeof c=="undefined")i.focus();else{let l=[i,...P(":not(details) > [href], summary, details[open] [href]",a)],f=Math.max(0,(Math.max(0,l.indexOf(c))+l.length+(p.type==="ArrowUp"?-1:1))%l.length);l[f].focus()}p.claim();break;default:i!==Ie()&&i.focus()}}),r.pipe(b(({mode:p})=>p==="global")).subscribe(p=>{switch(p.type){case"f":case"s":case"/":i.focus(),i.select(),p.claim();break}});let s=pi(i,{worker$:n});return O(s,li(a,{worker$:n,query$:s})).pipe(Re(...ae("search-share",e).map(p=>mi(p,{query$:s})),...ae("search-suggest",e).map(p=>fi(p,{worker$:n,keyboard$:r}))))}catch(n){return e.hidden=!0,Ye}}function di(e,{index$:t,location$:r}){return z([t,r.pipe(Q(ye()),b(o=>!!o.searchParams.get("h")))]).pipe(m(([o,n])=>ii(o.config)(n.searchParams.get("h"))),m(o=>{var a;let n=new Map,i=document.createNodeIterator(e,NodeFilter.SHOW_TEXT);for(let s=i.nextNode();s;s=i.nextNode())if((a=s.parentElement)!=null&&a.offsetHeight){let p=s.textContent,c=o(p);c.length>p.length&&n.set(s,c)}for(let[s,p]of n){let{childNodes:c}=x("span",null,p);s.replaceWith(...Array.from(c))}return{ref:e,nodes:n}}))}function fs(e,{viewport$:t,main$:r}){let o=e.closest(".md-grid"),n=o.offsetTop-o.parentElement.offsetTop;return z([r,t]).pipe(m(([{offset:i,height:a},{offset:{y:s}}])=>(a=a+Math.min(n,Math.max(0,s-i))-n,{height:a,locked:s>=i+n})),K((i,a)=>i.height===a.height&&i.locked===a.locked))}function Zr(e,o){var n=o,{header$:t}=n,r=so(n,["header$"]);let i=R(".md-sidebar__scrollwrap",e),{y:a}=Ve(i);return C(()=>{let s=new g,p=s.pipe(Z(),ie(!0)),c=s.pipe(Me(0,me));return c.pipe(re(t)).subscribe({next([{height:l},{height:f}]){i.style.height=`${l-2*a}px`,e.style.top=`${f}px`},complete(){i.style.height="",e.style.top=""}}),c.pipe(Ae()).subscribe(()=>{for(let l of P(".md-nav__link--active[href]",e)){if(!l.clientHeight)continue;let f=l.closest(".md-sidebar__scrollwrap");if(typeof f!="undefined"){let u=l.offsetTop-f.offsetTop,{height:d}=ce(f);f.scrollTo({top:u-d/2})}}}),ue(P("label[tabindex]",e)).pipe(ne(l=>h(l,"click").pipe(ve(se),m(()=>l),W(p)))).subscribe(l=>{let f=R(`[id="${l.htmlFor}"]`);R(`[aria-labelledby="${l.id}"]`).setAttribute("aria-expanded",`${f.checked}`)}),fs(e,r).pipe(w(l=>s.next(l)),_(()=>s.complete()),m(l=>$({ref:e},l)))})}function hi(e,t){if(typeof t!="undefined"){let r=`https://api.github.com/repos/${e}/${t}`;return st(je(`${r}/releases/latest`).pipe(de(()=>S),m(o=>({version:o.tag_name})),De({})),je(r).pipe(de(()=>S),m(o=>({stars:o.stargazers_count,forks:o.forks_count})),De({}))).pipe(m(([o,n])=>$($({},o),n)))}else{let r=`https://api.github.com/users/${e}`;return je(r).pipe(m(o=>({repositories:o.public_repos})),De({}))}}function bi(e,t){let r=`https://${e}/api/v4/projects/${encodeURIComponent(t)}`;return st(je(`${r}/releases/permalink/latest`).pipe(de(()=>S),m(({tag_name:o})=>({version:o})),De({})),je(r).pipe(de(()=>S),m(({star_count:o,forks_count:n})=>({stars:o,forks:n})),De({}))).pipe(m(([o,n])=>$($({},o),n)))}function vi(e){let t=e.match(/^.+github\.com\/([^/]+)\/?([^/]+)?/i);if(t){let[,r,o]=t;return hi(r,o)}if(t=e.match(/^.+?([^/]*gitlab[^/]+)\/(.+?)\/?$/i),t){let[,r,o]=t;return bi(r,o)}return S}var us;function ds(e){return us||(us=C(()=>{let t=__md_get("__source",sessionStorage);if(t)return I(t);if(ae("consent").length){let o=__md_get("__consent");if(!(o&&o.github))return S}return vi(e.href).pipe(w(o=>__md_set("__source",o,sessionStorage)))}).pipe(de(()=>S),b(t=>Object.keys(t).length>0),m(t=>({facts:t})),G(1)))}function gi(e){let t=R(":scope > :last-child",e);return C(()=>{let r=new g;return r.subscribe(({facts:o})=>{t.appendChild(_n(o)),t.classList.add("md-source__repository--active")}),ds(e).pipe(w(o=>r.next(o)),_(()=>r.complete()),m(o=>$({ref:e},o)))})}function hs(e,{viewport$:t,header$:r}){return ge(document.body).pipe(v(()=>mr(e,{header$:r,viewport$:t})),m(({offset:{y:o}})=>({hidden:o>=10})),ee("hidden"))}function yi(e,t){return C(()=>{let r=new g;return r.subscribe({next({hidden:o}){e.hidden=o},complete(){e.hidden=!1}}),(B("navigation.tabs.sticky")?I({hidden:!1}):hs(e,t)).pipe(w(o=>r.next(o)),_(()=>r.complete()),m(o=>$({ref:e},o)))})}function bs(e,{viewport$:t,header$:r}){let o=new Map,n=P(".md-nav__link",e);for(let s of n){let p=decodeURIComponent(s.hash.substring(1)),c=fe(`[id="${p}"]`);typeof c!="undefined"&&o.set(s,c)}let i=r.pipe(ee("height"),m(({height:s})=>{let p=Se("main"),c=R(":scope > :first-child",p);return s+.8*(c.offsetTop-p.offsetTop)}),pe());return ge(document.body).pipe(ee("height"),v(s=>C(()=>{let p=[];return I([...o].reduce((c,[l,f])=>{for(;p.length&&o.get(p[p.length-1]).tagName>=f.tagName;)p.pop();let u=f.offsetTop;for(;!u&&f.parentElement;)f=f.parentElement,u=f.offsetTop;let d=f.offsetParent;for(;d;d=d.offsetParent)u+=d.offsetTop;return c.set([...p=[...p,l]].reverse(),u)},new Map))}).pipe(m(p=>new Map([...p].sort(([,c],[,l])=>c-l))),He(i),v(([p,c])=>t.pipe(Fr(([l,f],{offset:{y:u},size:d})=>{let y=u+d.height>=Math.floor(s.height);for(;f.length;){let[,L]=f[0];if(L-c=u&&!y)f=[l.pop(),...f];else break}return[l,f]},[[],[...p]]),K((l,f)=>l[0]===f[0]&&l[1]===f[1])))))).pipe(m(([s,p])=>({prev:s.map(([c])=>c),next:p.map(([c])=>c)})),Q({prev:[],next:[]}),Be(2,1),m(([s,p])=>s.prev.length{let i=new g,a=i.pipe(Z(),ie(!0));if(i.subscribe(({prev:s,next:p})=>{for(let[c]of p)c.classList.remove("md-nav__link--passed"),c.classList.remove("md-nav__link--active");for(let[c,[l]]of s.entries())l.classList.add("md-nav__link--passed"),l.classList.toggle("md-nav__link--active",c===s.length-1)}),B("toc.follow")){let s=O(t.pipe(_e(1),m(()=>{})),t.pipe(_e(250),m(()=>"smooth")));i.pipe(b(({prev:p})=>p.length>0),He(o.pipe(ve(se))),re(s)).subscribe(([[{prev:p}],c])=>{let[l]=p[p.length-1];if(l.offsetHeight){let f=cr(l);if(typeof f!="undefined"){let u=l.offsetTop-f.offsetTop,{height:d}=ce(f);f.scrollTo({top:u-d/2,behavior:c})}}})}return B("navigation.tracking")&&t.pipe(W(a),ee("offset"),_e(250),Ce(1),W(n.pipe(Ce(1))),ct({delay:250}),re(i)).subscribe(([,{prev:s}])=>{let p=ye(),c=s[s.length-1];if(c&&c.length){let[l]=c,{hash:f}=new URL(l.href);p.hash!==f&&(p.hash=f,history.replaceState({},"",`${p}`))}else p.hash="",history.replaceState({},"",`${p}`)}),bs(e,{viewport$:t,header$:r}).pipe(w(s=>i.next(s)),_(()=>i.complete()),m(s=>$({ref:e},s)))})}function vs(e,{viewport$:t,main$:r,target$:o}){let n=t.pipe(m(({offset:{y:a}})=>a),Be(2,1),m(([a,s])=>a>s&&s>0),K()),i=r.pipe(m(({active:a})=>a));return z([i,n]).pipe(m(([a,s])=>!(a&&s)),K(),W(o.pipe(Ce(1))),ie(!0),ct({delay:250}),m(a=>({hidden:a})))}function Ei(e,{viewport$:t,header$:r,main$:o,target$:n}){let i=new g,a=i.pipe(Z(),ie(!0));return i.subscribe({next({hidden:s}){e.hidden=s,s?(e.setAttribute("tabindex","-1"),e.blur()):e.removeAttribute("tabindex")},complete(){e.style.top="",e.hidden=!0,e.removeAttribute("tabindex")}}),r.pipe(W(a),ee("height")).subscribe(({height:s})=>{e.style.top=`${s+16}px`}),h(e,"click").subscribe(s=>{s.preventDefault(),window.scrollTo({top:0})}),vs(e,{viewport$:t,main$:o,target$:n}).pipe(w(s=>i.next(s)),_(()=>i.complete()),m(s=>$({ref:e},s)))}function wi({document$:e,viewport$:t}){e.pipe(v(()=>P(".md-ellipsis")),ne(r=>tt(r).pipe(W(e.pipe(Ce(1))),b(o=>o),m(()=>r),Te(1))),b(r=>r.offsetWidth{let o=r.innerText,n=r.closest("a")||r;return n.title=o,B("content.tooltips")?mt(n,{viewport$:t}).pipe(W(e.pipe(Ce(1))),_(()=>n.removeAttribute("title"))):S})).subscribe(),B("content.tooltips")&&e.pipe(v(()=>P(".md-status")),ne(r=>mt(r,{viewport$:t}))).subscribe()}function Ti({document$:e,tablet$:t}){e.pipe(v(()=>P(".md-toggle--indeterminate")),w(r=>{r.indeterminate=!0,r.checked=!1}),ne(r=>h(r,"change").pipe(Dr(()=>r.classList.contains("md-toggle--indeterminate")),m(()=>r))),re(t)).subscribe(([r,o])=>{r.classList.remove("md-toggle--indeterminate"),o&&(r.checked=!1)})}function gs(){return/(iPad|iPhone|iPod)/.test(navigator.userAgent)}function Si({document$:e}){e.pipe(v(()=>P("[data-md-scrollfix]")),w(t=>t.removeAttribute("data-md-scrollfix")),b(gs),ne(t=>h(t,"touchstart").pipe(m(()=>t)))).subscribe(t=>{let r=t.scrollTop;r===0?t.scrollTop=1:r+t.offsetHeight===t.scrollHeight&&(t.scrollTop=r-1)})}function Oi({viewport$:e,tablet$:t}){z([ze("search"),t]).pipe(m(([r,o])=>r&&!o),v(r=>I(r).pipe(Ge(r?400:100))),re(e)).subscribe(([r,{offset:{y:o}}])=>{if(r)document.body.setAttribute("data-md-scrolllock",""),document.body.style.top=`-${o}px`;else{let n=-1*parseInt(document.body.style.top,10);document.body.removeAttribute("data-md-scrolllock"),document.body.style.top="",n&&window.scrollTo(0,n)}})}Object.entries||(Object.entries=function(e){let t=[];for(let r of Object.keys(e))t.push([r,e[r]]);return t});Object.values||(Object.values=function(e){let t=[];for(let r of Object.keys(e))t.push(e[r]);return t});typeof Element!="undefined"&&(Element.prototype.scrollTo||(Element.prototype.scrollTo=function(e,t){typeof e=="object"?(this.scrollLeft=e.left,this.scrollTop=e.top):(this.scrollLeft=e,this.scrollTop=t)}),Element.prototype.replaceWith||(Element.prototype.replaceWith=function(...e){let t=this.parentNode;if(t){e.length===0&&t.removeChild(this);for(let r=e.length-1;r>=0;r--){let o=e[r];typeof o=="string"?o=document.createTextNode(o):o.parentNode&&o.parentNode.removeChild(o),r?t.insertBefore(this.previousSibling,o):t.replaceChild(o,this)}}}));function ys(){return location.protocol==="file:"?Tt(`${new URL("search/search_index.js",eo.base)}`).pipe(m(()=>__index),G(1)):je(new URL("search/search_index.json",eo.base))}document.documentElement.classList.remove("no-js");document.documentElement.classList.add("js");var ot=Go(),Ut=sn(),Lt=ln(Ut),to=an(),Oe=gn(),hr=Pt("(min-width: 960px)"),Mi=Pt("(min-width: 1220px)"),_i=mn(),eo=xe(),Ai=document.forms.namedItem("search")?ys():Ye,ro=new g;Zn({alert$:ro});var oo=new g;B("navigation.instant")&&oi({location$:Ut,viewport$:Oe,progress$:oo}).subscribe(ot);var Li;((Li=eo.version)==null?void 0:Li.provider)==="mike"&&ci({document$:ot});O(Ut,Lt).pipe(Ge(125)).subscribe(()=>{Je("drawer",!1),Je("search",!1)});to.pipe(b(({mode:e})=>e==="global")).subscribe(e=>{switch(e.type){case"p":case",":let t=fe("link[rel=prev]");typeof t!="undefined"&<(t);break;case"n":case".":let r=fe("link[rel=next]");typeof r!="undefined"&<(r);break;case"Enter":let o=Ie();o instanceof HTMLLabelElement&&o.click()}});wi({viewport$:Oe,document$:ot});Ti({document$:ot,tablet$:hr});Si({document$:ot});Oi({viewport$:Oe,tablet$:hr});var rt=Kn(Se("header"),{viewport$:Oe}),Ft=ot.pipe(m(()=>Se("main")),v(e=>Gn(e,{viewport$:Oe,header$:rt})),G(1)),xs=O(...ae("consent").map(e=>En(e,{target$:Lt})),...ae("dialog").map(e=>qn(e,{alert$:ro})),...ae("header").map(e=>Yn(e,{viewport$:Oe,header$:rt,main$:Ft})),...ae("palette").map(e=>Jn(e)),...ae("progress").map(e=>Xn(e,{progress$:oo})),...ae("search").map(e=>ui(e,{index$:Ai,keyboard$:to})),...ae("source").map(e=>gi(e))),Es=C(()=>O(...ae("announce").map(e=>xn(e)),...ae("content").map(e=>zn(e,{viewport$:Oe,target$:Lt,print$:_i})),...ae("content").map(e=>B("search.highlight")?di(e,{index$:Ai,location$:Ut}):S),...ae("header-title").map(e=>Bn(e,{viewport$:Oe,header$:rt})),...ae("sidebar").map(e=>e.getAttribute("data-md-type")==="navigation"?Nr(Mi,()=>Zr(e,{viewport$:Oe,header$:rt,main$:Ft})):Nr(hr,()=>Zr(e,{viewport$:Oe,header$:rt,main$:Ft}))),...ae("tabs").map(e=>yi(e,{viewport$:Oe,header$:rt})),...ae("toc").map(e=>xi(e,{viewport$:Oe,header$:rt,main$:Ft,target$:Lt})),...ae("top").map(e=>Ei(e,{viewport$:Oe,header$:rt,main$:Ft,target$:Lt})))),Ci=ot.pipe(v(()=>Es),Re(xs),G(1));Ci.subscribe();window.document$=ot;window.location$=Ut;window.target$=Lt;window.keyboard$=to;window.viewport$=Oe;window.tablet$=hr;window.screen$=Mi;window.print$=_i;window.alert$=ro;window.progress$=oo;window.component$=Ci;})(); +//# sourceMappingURL=bundle.525ec568.min.js.map + diff --git a/assets/javascripts/bundle.525ec568.min.js.map b/assets/javascripts/bundle.525ec568.min.js.map new file mode 100644 index 0000000..ef5d8d3 --- /dev/null +++ b/assets/javascripts/bundle.525ec568.min.js.map @@ -0,0 +1,7 @@ +{ + "version": 3, + "sources": ["node_modules/focus-visible/dist/focus-visible.js", "node_modules/escape-html/index.js", "node_modules/clipboard/dist/clipboard.js", "src/templates/assets/javascripts/bundle.ts", "node_modules/tslib/tslib.es6.mjs", "node_modules/rxjs/src/internal/util/isFunction.ts", "node_modules/rxjs/src/internal/util/createErrorClass.ts", "node_modules/rxjs/src/internal/util/UnsubscriptionError.ts", "node_modules/rxjs/src/internal/util/arrRemove.ts", "node_modules/rxjs/src/internal/Subscription.ts", "node_modules/rxjs/src/internal/config.ts", "node_modules/rxjs/src/internal/scheduler/timeoutProvider.ts", "node_modules/rxjs/src/internal/util/reportUnhandledError.ts", "node_modules/rxjs/src/internal/util/noop.ts", "node_modules/rxjs/src/internal/NotificationFactories.ts", "node_modules/rxjs/src/internal/util/errorContext.ts", "node_modules/rxjs/src/internal/Subscriber.ts", "node_modules/rxjs/src/internal/symbol/observable.ts", "node_modules/rxjs/src/internal/util/identity.ts", "node_modules/rxjs/src/internal/util/pipe.ts", "node_modules/rxjs/src/internal/Observable.ts", "node_modules/rxjs/src/internal/util/lift.ts", "node_modules/rxjs/src/internal/operators/OperatorSubscriber.ts", "node_modules/rxjs/src/internal/scheduler/animationFrameProvider.ts", "node_modules/rxjs/src/internal/util/ObjectUnsubscribedError.ts", "node_modules/rxjs/src/internal/Subject.ts", "node_modules/rxjs/src/internal/BehaviorSubject.ts", "node_modules/rxjs/src/internal/scheduler/dateTimestampProvider.ts", "node_modules/rxjs/src/internal/ReplaySubject.ts", "node_modules/rxjs/src/internal/scheduler/Action.ts", "node_modules/rxjs/src/internal/scheduler/intervalProvider.ts", "node_modules/rxjs/src/internal/scheduler/AsyncAction.ts", "node_modules/rxjs/src/internal/Scheduler.ts", "node_modules/rxjs/src/internal/scheduler/AsyncScheduler.ts", "node_modules/rxjs/src/internal/scheduler/async.ts", "node_modules/rxjs/src/internal/scheduler/QueueAction.ts", "node_modules/rxjs/src/internal/scheduler/QueueScheduler.ts", "node_modules/rxjs/src/internal/scheduler/queue.ts", "node_modules/rxjs/src/internal/scheduler/AnimationFrameAction.ts", "node_modules/rxjs/src/internal/scheduler/AnimationFrameScheduler.ts", "node_modules/rxjs/src/internal/scheduler/animationFrame.ts", "node_modules/rxjs/src/internal/observable/empty.ts", "node_modules/rxjs/src/internal/util/isScheduler.ts", "node_modules/rxjs/src/internal/util/args.ts", "node_modules/rxjs/src/internal/util/isArrayLike.ts", "node_modules/rxjs/src/internal/util/isPromise.ts", "node_modules/rxjs/src/internal/util/isInteropObservable.ts", "node_modules/rxjs/src/internal/util/isAsyncIterable.ts", "node_modules/rxjs/src/internal/util/throwUnobservableError.ts", "node_modules/rxjs/src/internal/symbol/iterator.ts", "node_modules/rxjs/src/internal/util/isIterable.ts", "node_modules/rxjs/src/internal/util/isReadableStreamLike.ts", "node_modules/rxjs/src/internal/observable/innerFrom.ts", "node_modules/rxjs/src/internal/util/executeSchedule.ts", "node_modules/rxjs/src/internal/operators/observeOn.ts", "node_modules/rxjs/src/internal/operators/subscribeOn.ts", "node_modules/rxjs/src/internal/scheduled/scheduleObservable.ts", "node_modules/rxjs/src/internal/scheduled/schedulePromise.ts", "node_modules/rxjs/src/internal/scheduled/scheduleArray.ts", "node_modules/rxjs/src/internal/scheduled/scheduleIterable.ts", "node_modules/rxjs/src/internal/scheduled/scheduleAsyncIterable.ts", "node_modules/rxjs/src/internal/scheduled/scheduleReadableStreamLike.ts", "node_modules/rxjs/src/internal/scheduled/scheduled.ts", "node_modules/rxjs/src/internal/observable/from.ts", "node_modules/rxjs/src/internal/observable/of.ts", "node_modules/rxjs/src/internal/observable/throwError.ts", "node_modules/rxjs/src/internal/util/EmptyError.ts", "node_modules/rxjs/src/internal/util/isDate.ts", "node_modules/rxjs/src/internal/operators/map.ts", "node_modules/rxjs/src/internal/util/mapOneOrManyArgs.ts", "node_modules/rxjs/src/internal/util/argsArgArrayOrObject.ts", "node_modules/rxjs/src/internal/util/createObject.ts", "node_modules/rxjs/src/internal/observable/combineLatest.ts", "node_modules/rxjs/src/internal/operators/mergeInternals.ts", "node_modules/rxjs/src/internal/operators/mergeMap.ts", "node_modules/rxjs/src/internal/operators/mergeAll.ts", "node_modules/rxjs/src/internal/operators/concatAll.ts", "node_modules/rxjs/src/internal/observable/concat.ts", "node_modules/rxjs/src/internal/observable/defer.ts", "node_modules/rxjs/src/internal/observable/fromEvent.ts", "node_modules/rxjs/src/internal/observable/fromEventPattern.ts", "node_modules/rxjs/src/internal/observable/timer.ts", "node_modules/rxjs/src/internal/observable/merge.ts", "node_modules/rxjs/src/internal/observable/never.ts", "node_modules/rxjs/src/internal/util/argsOrArgArray.ts", "node_modules/rxjs/src/internal/operators/filter.ts", "node_modules/rxjs/src/internal/observable/zip.ts", "node_modules/rxjs/src/internal/operators/audit.ts", "node_modules/rxjs/src/internal/operators/auditTime.ts", "node_modules/rxjs/src/internal/operators/bufferCount.ts", "node_modules/rxjs/src/internal/operators/catchError.ts", "node_modules/rxjs/src/internal/operators/scanInternals.ts", "node_modules/rxjs/src/internal/operators/combineLatest.ts", "node_modules/rxjs/src/internal/operators/combineLatestWith.ts", "node_modules/rxjs/src/internal/operators/debounce.ts", "node_modules/rxjs/src/internal/operators/debounceTime.ts", "node_modules/rxjs/src/internal/operators/defaultIfEmpty.ts", "node_modules/rxjs/src/internal/operators/take.ts", "node_modules/rxjs/src/internal/operators/ignoreElements.ts", "node_modules/rxjs/src/internal/operators/mapTo.ts", "node_modules/rxjs/src/internal/operators/delayWhen.ts", "node_modules/rxjs/src/internal/operators/delay.ts", "node_modules/rxjs/src/internal/operators/distinctUntilChanged.ts", "node_modules/rxjs/src/internal/operators/distinctUntilKeyChanged.ts", "node_modules/rxjs/src/internal/operators/throwIfEmpty.ts", "node_modules/rxjs/src/internal/operators/endWith.ts", "node_modules/rxjs/src/internal/operators/finalize.ts", "node_modules/rxjs/src/internal/operators/first.ts", "node_modules/rxjs/src/internal/operators/takeLast.ts", "node_modules/rxjs/src/internal/operators/merge.ts", "node_modules/rxjs/src/internal/operators/mergeWith.ts", "node_modules/rxjs/src/internal/operators/repeat.ts", "node_modules/rxjs/src/internal/operators/scan.ts", "node_modules/rxjs/src/internal/operators/share.ts", "node_modules/rxjs/src/internal/operators/shareReplay.ts", "node_modules/rxjs/src/internal/operators/skip.ts", "node_modules/rxjs/src/internal/operators/skipUntil.ts", "node_modules/rxjs/src/internal/operators/startWith.ts", "node_modules/rxjs/src/internal/operators/switchMap.ts", "node_modules/rxjs/src/internal/operators/takeUntil.ts", "node_modules/rxjs/src/internal/operators/takeWhile.ts", "node_modules/rxjs/src/internal/operators/tap.ts", "node_modules/rxjs/src/internal/operators/throttle.ts", "node_modules/rxjs/src/internal/operators/throttleTime.ts", "node_modules/rxjs/src/internal/operators/withLatestFrom.ts", "node_modules/rxjs/src/internal/operators/zip.ts", "node_modules/rxjs/src/internal/operators/zipWith.ts", "src/templates/assets/javascripts/browser/document/index.ts", "src/templates/assets/javascripts/browser/element/_/index.ts", "src/templates/assets/javascripts/browser/element/focus/index.ts", "src/templates/assets/javascripts/browser/element/hover/index.ts", "src/templates/assets/javascripts/utilities/h/index.ts", "src/templates/assets/javascripts/utilities/round/index.ts", "src/templates/assets/javascripts/browser/script/index.ts", "src/templates/assets/javascripts/browser/element/size/_/index.ts", "src/templates/assets/javascripts/browser/element/size/content/index.ts", "src/templates/assets/javascripts/browser/element/offset/_/index.ts", "src/templates/assets/javascripts/browser/element/offset/content/index.ts", "src/templates/assets/javascripts/browser/element/visibility/index.ts", "src/templates/assets/javascripts/browser/toggle/index.ts", "src/templates/assets/javascripts/browser/keyboard/index.ts", "src/templates/assets/javascripts/browser/location/_/index.ts", "src/templates/assets/javascripts/browser/location/hash/index.ts", "src/templates/assets/javascripts/browser/media/index.ts", "src/templates/assets/javascripts/browser/request/index.ts", "src/templates/assets/javascripts/browser/viewport/offset/index.ts", "src/templates/assets/javascripts/browser/viewport/size/index.ts", "src/templates/assets/javascripts/browser/viewport/_/index.ts", "src/templates/assets/javascripts/browser/viewport/at/index.ts", "src/templates/assets/javascripts/browser/worker/index.ts", "src/templates/assets/javascripts/_/index.ts", "src/templates/assets/javascripts/components/_/index.ts", "src/templates/assets/javascripts/components/announce/index.ts", "src/templates/assets/javascripts/components/consent/index.ts", "src/templates/assets/javascripts/templates/tooltip/index.tsx", "src/templates/assets/javascripts/templates/annotation/index.tsx", "src/templates/assets/javascripts/templates/clipboard/index.tsx", "src/templates/assets/javascripts/templates/search/index.tsx", "src/templates/assets/javascripts/templates/source/index.tsx", "src/templates/assets/javascripts/templates/tabbed/index.tsx", "src/templates/assets/javascripts/templates/table/index.tsx", "src/templates/assets/javascripts/templates/version/index.tsx", "src/templates/assets/javascripts/components/tooltip2/index.ts", "src/templates/assets/javascripts/components/content/annotation/_/index.ts", "src/templates/assets/javascripts/components/content/annotation/list/index.ts", "src/templates/assets/javascripts/components/content/annotation/block/index.ts", "src/templates/assets/javascripts/components/content/code/_/index.ts", "src/templates/assets/javascripts/components/content/details/index.ts", "src/templates/assets/javascripts/components/content/mermaid/index.css", "src/templates/assets/javascripts/components/content/mermaid/index.ts", "src/templates/assets/javascripts/components/content/table/index.ts", "src/templates/assets/javascripts/components/content/tabs/index.ts", "src/templates/assets/javascripts/components/content/_/index.ts", "src/templates/assets/javascripts/components/dialog/index.ts", "src/templates/assets/javascripts/components/tooltip/index.ts", "src/templates/assets/javascripts/components/header/_/index.ts", "src/templates/assets/javascripts/components/header/title/index.ts", "src/templates/assets/javascripts/components/main/index.ts", "src/templates/assets/javascripts/components/palette/index.ts", "src/templates/assets/javascripts/components/progress/index.ts", "src/templates/assets/javascripts/integrations/clipboard/index.ts", "src/templates/assets/javascripts/integrations/sitemap/index.ts", "src/templates/assets/javascripts/integrations/instant/index.ts", "src/templates/assets/javascripts/integrations/search/highlighter/index.ts", "src/templates/assets/javascripts/integrations/search/worker/message/index.ts", "src/templates/assets/javascripts/integrations/search/worker/_/index.ts", "src/templates/assets/javascripts/integrations/version/findurl/index.ts", "src/templates/assets/javascripts/integrations/version/index.ts", "src/templates/assets/javascripts/components/search/query/index.ts", "src/templates/assets/javascripts/components/search/result/index.ts", "src/templates/assets/javascripts/components/search/share/index.ts", "src/templates/assets/javascripts/components/search/suggest/index.ts", "src/templates/assets/javascripts/components/search/_/index.ts", "src/templates/assets/javascripts/components/search/highlight/index.ts", "src/templates/assets/javascripts/components/sidebar/index.ts", "src/templates/assets/javascripts/components/source/facts/github/index.ts", "src/templates/assets/javascripts/components/source/facts/gitlab/index.ts", "src/templates/assets/javascripts/components/source/facts/_/index.ts", "src/templates/assets/javascripts/components/source/_/index.ts", "src/templates/assets/javascripts/components/tabs/index.ts", "src/templates/assets/javascripts/components/toc/index.ts", "src/templates/assets/javascripts/components/top/index.ts", "src/templates/assets/javascripts/patches/ellipsis/index.ts", "src/templates/assets/javascripts/patches/indeterminate/index.ts", "src/templates/assets/javascripts/patches/scrollfix/index.ts", "src/templates/assets/javascripts/patches/scrolllock/index.ts", "src/templates/assets/javascripts/polyfills/index.ts"], + "sourcesContent": ["(function (global, factory) {\n typeof exports === 'object' && typeof module !== 'undefined' ? factory() :\n typeof define === 'function' && define.amd ? define(factory) :\n (factory());\n}(this, (function () { 'use strict';\n\n /**\n * Applies the :focus-visible polyfill at the given scope.\n * A scope in this case is either the top-level Document or a Shadow Root.\n *\n * @param {(Document|ShadowRoot)} scope\n * @see https://github.com/WICG/focus-visible\n */\n function applyFocusVisiblePolyfill(scope) {\n var hadKeyboardEvent = true;\n var hadFocusVisibleRecently = false;\n var hadFocusVisibleRecentlyTimeout = null;\n\n var inputTypesAllowlist = {\n text: true,\n search: true,\n url: true,\n tel: true,\n email: true,\n password: true,\n number: true,\n date: true,\n month: true,\n week: true,\n time: true,\n datetime: true,\n 'datetime-local': true\n };\n\n /**\n * Helper function for legacy browsers and iframes which sometimes focus\n * elements like document, body, and non-interactive SVG.\n * @param {Element} el\n */\n function isValidFocusTarget(el) {\n if (\n el &&\n el !== document &&\n el.nodeName !== 'HTML' &&\n el.nodeName !== 'BODY' &&\n 'classList' in el &&\n 'contains' in el.classList\n ) {\n return true;\n }\n return false;\n }\n\n /**\n * Computes whether the given element should automatically trigger the\n * `focus-visible` class being added, i.e. whether it should always match\n * `:focus-visible` when focused.\n * @param {Element} el\n * @return {boolean}\n */\n function focusTriggersKeyboardModality(el) {\n var type = el.type;\n var tagName = el.tagName;\n\n if (tagName === 'INPUT' && inputTypesAllowlist[type] && !el.readOnly) {\n return true;\n }\n\n if (tagName === 'TEXTAREA' && !el.readOnly) {\n return true;\n }\n\n if (el.isContentEditable) {\n return true;\n }\n\n return false;\n }\n\n /**\n * Add the `focus-visible` class to the given element if it was not added by\n * the author.\n * @param {Element} el\n */\n function addFocusVisibleClass(el) {\n if (el.classList.contains('focus-visible')) {\n return;\n }\n el.classList.add('focus-visible');\n el.setAttribute('data-focus-visible-added', '');\n }\n\n /**\n * Remove the `focus-visible` class from the given element if it was not\n * originally added by the author.\n * @param {Element} el\n */\n function removeFocusVisibleClass(el) {\n if (!el.hasAttribute('data-focus-visible-added')) {\n return;\n }\n el.classList.remove('focus-visible');\n el.removeAttribute('data-focus-visible-added');\n }\n\n /**\n * If the most recent user interaction was via the keyboard;\n * and the key press did not include a meta, alt/option, or control key;\n * then the modality is keyboard. Otherwise, the modality is not keyboard.\n * Apply `focus-visible` to any current active element and keep track\n * of our keyboard modality state with `hadKeyboardEvent`.\n * @param {KeyboardEvent} e\n */\n function onKeyDown(e) {\n if (e.metaKey || e.altKey || e.ctrlKey) {\n return;\n }\n\n if (isValidFocusTarget(scope.activeElement)) {\n addFocusVisibleClass(scope.activeElement);\n }\n\n hadKeyboardEvent = true;\n }\n\n /**\n * If at any point a user clicks with a pointing device, ensure that we change\n * the modality away from keyboard.\n * This avoids the situation where a user presses a key on an already focused\n * element, and then clicks on a different element, focusing it with a\n * pointing device, while we still think we're in keyboard modality.\n * @param {Event} e\n */\n function onPointerDown(e) {\n hadKeyboardEvent = false;\n }\n\n /**\n * On `focus`, add the `focus-visible` class to the target if:\n * - the target received focus as a result of keyboard navigation, or\n * - the event target is an element that will likely require interaction\n * via the keyboard (e.g. a text box)\n * @param {Event} e\n */\n function onFocus(e) {\n // Prevent IE from focusing the document or HTML element.\n if (!isValidFocusTarget(e.target)) {\n return;\n }\n\n if (hadKeyboardEvent || focusTriggersKeyboardModality(e.target)) {\n addFocusVisibleClass(e.target);\n }\n }\n\n /**\n * On `blur`, remove the `focus-visible` class from the target.\n * @param {Event} e\n */\n function onBlur(e) {\n if (!isValidFocusTarget(e.target)) {\n return;\n }\n\n if (\n e.target.classList.contains('focus-visible') ||\n e.target.hasAttribute('data-focus-visible-added')\n ) {\n // To detect a tab/window switch, we look for a blur event followed\n // rapidly by a visibility change.\n // If we don't see a visibility change within 100ms, it's probably a\n // regular focus change.\n hadFocusVisibleRecently = true;\n window.clearTimeout(hadFocusVisibleRecentlyTimeout);\n hadFocusVisibleRecentlyTimeout = window.setTimeout(function() {\n hadFocusVisibleRecently = false;\n }, 100);\n removeFocusVisibleClass(e.target);\n }\n }\n\n /**\n * If the user changes tabs, keep track of whether or not the previously\n * focused element had .focus-visible.\n * @param {Event} e\n */\n function onVisibilityChange(e) {\n if (document.visibilityState === 'hidden') {\n // If the tab becomes active again, the browser will handle calling focus\n // on the element (Safari actually calls it twice).\n // If this tab change caused a blur on an element with focus-visible,\n // re-apply the class when the user switches back to the tab.\n if (hadFocusVisibleRecently) {\n hadKeyboardEvent = true;\n }\n addInitialPointerMoveListeners();\n }\n }\n\n /**\n * Add a group of listeners to detect usage of any pointing devices.\n * These listeners will be added when the polyfill first loads, and anytime\n * the window is blurred, so that they are active when the window regains\n * focus.\n */\n function addInitialPointerMoveListeners() {\n document.addEventListener('mousemove', onInitialPointerMove);\n document.addEventListener('mousedown', onInitialPointerMove);\n document.addEventListener('mouseup', onInitialPointerMove);\n document.addEventListener('pointermove', onInitialPointerMove);\n document.addEventListener('pointerdown', onInitialPointerMove);\n document.addEventListener('pointerup', onInitialPointerMove);\n document.addEventListener('touchmove', onInitialPointerMove);\n document.addEventListener('touchstart', onInitialPointerMove);\n document.addEventListener('touchend', onInitialPointerMove);\n }\n\n function removeInitialPointerMoveListeners() {\n document.removeEventListener('mousemove', onInitialPointerMove);\n document.removeEventListener('mousedown', onInitialPointerMove);\n document.removeEventListener('mouseup', onInitialPointerMove);\n document.removeEventListener('pointermove', onInitialPointerMove);\n document.removeEventListener('pointerdown', onInitialPointerMove);\n document.removeEventListener('pointerup', onInitialPointerMove);\n document.removeEventListener('touchmove', onInitialPointerMove);\n document.removeEventListener('touchstart', onInitialPointerMove);\n document.removeEventListener('touchend', onInitialPointerMove);\n }\n\n /**\n * When the polfyill first loads, assume the user is in keyboard modality.\n * If any event is received from a pointing device (e.g. mouse, pointer,\n * touch), turn off keyboard modality.\n * This accounts for situations where focus enters the page from the URL bar.\n * @param {Event} e\n */\n function onInitialPointerMove(e) {\n // Work around a Safari quirk that fires a mousemove on whenever the\n // window blurs, even if you're tabbing out of the page. \u00AF\\_(\u30C4)_/\u00AF\n if (e.target.nodeName && e.target.nodeName.toLowerCase() === 'html') {\n return;\n }\n\n hadKeyboardEvent = false;\n removeInitialPointerMoveListeners();\n }\n\n // For some kinds of state, we are interested in changes at the global scope\n // only. For example, global pointer input, global key presses and global\n // visibility change should affect the state at every scope:\n document.addEventListener('keydown', onKeyDown, true);\n document.addEventListener('mousedown', onPointerDown, true);\n document.addEventListener('pointerdown', onPointerDown, true);\n document.addEventListener('touchstart', onPointerDown, true);\n document.addEventListener('visibilitychange', onVisibilityChange, true);\n\n addInitialPointerMoveListeners();\n\n // For focus and blur, we specifically care about state changes in the local\n // scope. This is because focus / blur events that originate from within a\n // shadow root are not re-dispatched from the host element if it was already\n // the active element in its own scope:\n scope.addEventListener('focus', onFocus, true);\n scope.addEventListener('blur', onBlur, true);\n\n // We detect that a node is a ShadowRoot by ensuring that it is a\n // DocumentFragment and also has a host property. This check covers native\n // implementation and polyfill implementation transparently. If we only cared\n // about the native implementation, we could just check if the scope was\n // an instance of a ShadowRoot.\n if (scope.nodeType === Node.DOCUMENT_FRAGMENT_NODE && scope.host) {\n // Since a ShadowRoot is a special kind of DocumentFragment, it does not\n // have a root element to add a class to. So, we add this attribute to the\n // host element instead:\n scope.host.setAttribute('data-js-focus-visible', '');\n } else if (scope.nodeType === Node.DOCUMENT_NODE) {\n document.documentElement.classList.add('js-focus-visible');\n document.documentElement.setAttribute('data-js-focus-visible', '');\n }\n }\n\n // It is important to wrap all references to global window and document in\n // these checks to support server-side rendering use cases\n // @see https://github.com/WICG/focus-visible/issues/199\n if (typeof window !== 'undefined' && typeof document !== 'undefined') {\n // Make the polyfill helper globally available. This can be used as a signal\n // to interested libraries that wish to coordinate with the polyfill for e.g.,\n // applying the polyfill to a shadow root:\n window.applyFocusVisiblePolyfill = applyFocusVisiblePolyfill;\n\n // Notify interested libraries of the polyfill's presence, in case the\n // polyfill was loaded lazily:\n var event;\n\n try {\n event = new CustomEvent('focus-visible-polyfill-ready');\n } catch (error) {\n // IE11 does not support using CustomEvent as a constructor directly:\n event = document.createEvent('CustomEvent');\n event.initCustomEvent('focus-visible-polyfill-ready', false, false, {});\n }\n\n window.dispatchEvent(event);\n }\n\n if (typeof document !== 'undefined') {\n // Apply the polyfill to the global document, so that no JavaScript\n // coordination is required to use the polyfill in the top-level document:\n applyFocusVisiblePolyfill(document);\n }\n\n})));\n", "/*!\n * escape-html\n * Copyright(c) 2012-2013 TJ Holowaychuk\n * Copyright(c) 2015 Andreas Lubbe\n * Copyright(c) 2015 Tiancheng \"Timothy\" Gu\n * MIT Licensed\n */\n\n'use strict';\n\n/**\n * Module variables.\n * @private\n */\n\nvar matchHtmlRegExp = /[\"'&<>]/;\n\n/**\n * Module exports.\n * @public\n */\n\nmodule.exports = escapeHtml;\n\n/**\n * Escape special characters in the given string of html.\n *\n * @param {string} string The string to escape for inserting into HTML\n * @return {string}\n * @public\n */\n\nfunction escapeHtml(string) {\n var str = '' + string;\n var match = matchHtmlRegExp.exec(str);\n\n if (!match) {\n return str;\n }\n\n var escape;\n var html = '';\n var index = 0;\n var lastIndex = 0;\n\n for (index = match.index; index < str.length; index++) {\n switch (str.charCodeAt(index)) {\n case 34: // \"\n escape = '"';\n break;\n case 38: // &\n escape = '&';\n break;\n case 39: // '\n escape = ''';\n break;\n case 60: // <\n escape = '<';\n break;\n case 62: // >\n escape = '>';\n break;\n default:\n continue;\n }\n\n if (lastIndex !== index) {\n html += str.substring(lastIndex, index);\n }\n\n lastIndex = index + 1;\n html += escape;\n }\n\n return lastIndex !== index\n ? html + str.substring(lastIndex, index)\n : html;\n}\n", "/*!\n * clipboard.js v2.0.11\n * https://clipboardjs.com/\n *\n * Licensed MIT \u00A9 Zeno Rocha\n */\n(function webpackUniversalModuleDefinition(root, factory) {\n\tif(typeof exports === 'object' && typeof module === 'object')\n\t\tmodule.exports = factory();\n\telse if(typeof define === 'function' && define.amd)\n\t\tdefine([], factory);\n\telse if(typeof exports === 'object')\n\t\texports[\"ClipboardJS\"] = factory();\n\telse\n\t\troot[\"ClipboardJS\"] = factory();\n})(this, function() {\nreturn /******/ (function() { // webpackBootstrap\n/******/ \tvar __webpack_modules__ = ({\n\n/***/ 686:\n/***/ (function(__unused_webpack_module, __webpack_exports__, __webpack_require__) {\n\n\"use strict\";\n\n// EXPORTS\n__webpack_require__.d(__webpack_exports__, {\n \"default\": function() { return /* binding */ clipboard; }\n});\n\n// EXTERNAL MODULE: ./node_modules/tiny-emitter/index.js\nvar tiny_emitter = __webpack_require__(279);\nvar tiny_emitter_default = /*#__PURE__*/__webpack_require__.n(tiny_emitter);\n// EXTERNAL MODULE: ./node_modules/good-listener/src/listen.js\nvar listen = __webpack_require__(370);\nvar listen_default = /*#__PURE__*/__webpack_require__.n(listen);\n// EXTERNAL MODULE: ./node_modules/select/src/select.js\nvar src_select = __webpack_require__(817);\nvar select_default = /*#__PURE__*/__webpack_require__.n(src_select);\n;// CONCATENATED MODULE: ./src/common/command.js\n/**\n * Executes a given operation type.\n * @param {String} type\n * @return {Boolean}\n */\nfunction command(type) {\n try {\n return document.execCommand(type);\n } catch (err) {\n return false;\n }\n}\n;// CONCATENATED MODULE: ./src/actions/cut.js\n\n\n/**\n * Cut action wrapper.\n * @param {String|HTMLElement} target\n * @return {String}\n */\n\nvar ClipboardActionCut = function ClipboardActionCut(target) {\n var selectedText = select_default()(target);\n command('cut');\n return selectedText;\n};\n\n/* harmony default export */ var actions_cut = (ClipboardActionCut);\n;// CONCATENATED MODULE: ./src/common/create-fake-element.js\n/**\n * Creates a fake textarea element with a value.\n * @param {String} value\n * @return {HTMLElement}\n */\nfunction createFakeElement(value) {\n var isRTL = document.documentElement.getAttribute('dir') === 'rtl';\n var fakeElement = document.createElement('textarea'); // Prevent zooming on iOS\n\n fakeElement.style.fontSize = '12pt'; // Reset box model\n\n fakeElement.style.border = '0';\n fakeElement.style.padding = '0';\n fakeElement.style.margin = '0'; // Move element out of screen horizontally\n\n fakeElement.style.position = 'absolute';\n fakeElement.style[isRTL ? 'right' : 'left'] = '-9999px'; // Move element to the same position vertically\n\n var yPosition = window.pageYOffset || document.documentElement.scrollTop;\n fakeElement.style.top = \"\".concat(yPosition, \"px\");\n fakeElement.setAttribute('readonly', '');\n fakeElement.value = value;\n return fakeElement;\n}\n;// CONCATENATED MODULE: ./src/actions/copy.js\n\n\n\n/**\n * Create fake copy action wrapper using a fake element.\n * @param {String} target\n * @param {Object} options\n * @return {String}\n */\n\nvar fakeCopyAction = function fakeCopyAction(value, options) {\n var fakeElement = createFakeElement(value);\n options.container.appendChild(fakeElement);\n var selectedText = select_default()(fakeElement);\n command('copy');\n fakeElement.remove();\n return selectedText;\n};\n/**\n * Copy action wrapper.\n * @param {String|HTMLElement} target\n * @param {Object} options\n * @return {String}\n */\n\n\nvar ClipboardActionCopy = function ClipboardActionCopy(target) {\n var options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {\n container: document.body\n };\n var selectedText = '';\n\n if (typeof target === 'string') {\n selectedText = fakeCopyAction(target, options);\n } else if (target instanceof HTMLInputElement && !['text', 'search', 'url', 'tel', 'password'].includes(target === null || target === void 0 ? void 0 : target.type)) {\n // If input type doesn't support `setSelectionRange`. Simulate it. https://developer.mozilla.org/en-US/docs/Web/API/HTMLInputElement/setSelectionRange\n selectedText = fakeCopyAction(target.value, options);\n } else {\n selectedText = select_default()(target);\n command('copy');\n }\n\n return selectedText;\n};\n\n/* harmony default export */ var actions_copy = (ClipboardActionCopy);\n;// CONCATENATED MODULE: ./src/actions/default.js\nfunction _typeof(obj) { \"@babel/helpers - typeof\"; if (typeof Symbol === \"function\" && typeof Symbol.iterator === \"symbol\") { _typeof = function _typeof(obj) { return typeof obj; }; } else { _typeof = function _typeof(obj) { return obj && typeof Symbol === \"function\" && obj.constructor === Symbol && obj !== Symbol.prototype ? \"symbol\" : typeof obj; }; } return _typeof(obj); }\n\n\n\n/**\n * Inner function which performs selection from either `text` or `target`\n * properties and then executes copy or cut operations.\n * @param {Object} options\n */\n\nvar ClipboardActionDefault = function ClipboardActionDefault() {\n var options = arguments.length > 0 && arguments[0] !== undefined ? arguments[0] : {};\n // Defines base properties passed from constructor.\n var _options$action = options.action,\n action = _options$action === void 0 ? 'copy' : _options$action,\n container = options.container,\n target = options.target,\n text = options.text; // Sets the `action` to be performed which can be either 'copy' or 'cut'.\n\n if (action !== 'copy' && action !== 'cut') {\n throw new Error('Invalid \"action\" value, use either \"copy\" or \"cut\"');\n } // Sets the `target` property using an element that will be have its content copied.\n\n\n if (target !== undefined) {\n if (target && _typeof(target) === 'object' && target.nodeType === 1) {\n if (action === 'copy' && target.hasAttribute('disabled')) {\n throw new Error('Invalid \"target\" attribute. Please use \"readonly\" instead of \"disabled\" attribute');\n }\n\n if (action === 'cut' && (target.hasAttribute('readonly') || target.hasAttribute('disabled'))) {\n throw new Error('Invalid \"target\" attribute. You can\\'t cut text from elements with \"readonly\" or \"disabled\" attributes');\n }\n } else {\n throw new Error('Invalid \"target\" value, use a valid Element');\n }\n } // Define selection strategy based on `text` property.\n\n\n if (text) {\n return actions_copy(text, {\n container: container\n });\n } // Defines which selection strategy based on `target` property.\n\n\n if (target) {\n return action === 'cut' ? actions_cut(target) : actions_copy(target, {\n container: container\n });\n }\n};\n\n/* harmony default export */ var actions_default = (ClipboardActionDefault);\n;// CONCATENATED MODULE: ./src/clipboard.js\nfunction clipboard_typeof(obj) { \"@babel/helpers - typeof\"; if (typeof Symbol === \"function\" && typeof Symbol.iterator === \"symbol\") { clipboard_typeof = function _typeof(obj) { return typeof obj; }; } else { clipboard_typeof = function _typeof(obj) { return obj && typeof Symbol === \"function\" && obj.constructor === Symbol && obj !== Symbol.prototype ? \"symbol\" : typeof obj; }; } return clipboard_typeof(obj); }\n\nfunction _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError(\"Cannot call a class as a function\"); } }\n\nfunction _defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if (\"value\" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } }\n\nfunction _createClass(Constructor, protoProps, staticProps) { if (protoProps) _defineProperties(Constructor.prototype, protoProps); if (staticProps) _defineProperties(Constructor, staticProps); return Constructor; }\n\nfunction _inherits(subClass, superClass) { if (typeof superClass !== \"function\" && superClass !== null) { throw new TypeError(\"Super expression must either be null or a function\"); } subClass.prototype = Object.create(superClass && superClass.prototype, { constructor: { value: subClass, writable: true, configurable: true } }); if (superClass) _setPrototypeOf(subClass, superClass); }\n\nfunction _setPrototypeOf(o, p) { _setPrototypeOf = Object.setPrototypeOf || function _setPrototypeOf(o, p) { o.__proto__ = p; return o; }; return _setPrototypeOf(o, p); }\n\nfunction _createSuper(Derived) { var hasNativeReflectConstruct = _isNativeReflectConstruct(); return function _createSuperInternal() { var Super = _getPrototypeOf(Derived), result; if (hasNativeReflectConstruct) { var NewTarget = _getPrototypeOf(this).constructor; result = Reflect.construct(Super, arguments, NewTarget); } else { result = Super.apply(this, arguments); } return _possibleConstructorReturn(this, result); }; }\n\nfunction _possibleConstructorReturn(self, call) { if (call && (clipboard_typeof(call) === \"object\" || typeof call === \"function\")) { return call; } return _assertThisInitialized(self); }\n\nfunction _assertThisInitialized(self) { if (self === void 0) { throw new ReferenceError(\"this hasn't been initialised - super() hasn't been called\"); } return self; }\n\nfunction _isNativeReflectConstruct() { if (typeof Reflect === \"undefined\" || !Reflect.construct) return false; if (Reflect.construct.sham) return false; if (typeof Proxy === \"function\") return true; try { Date.prototype.toString.call(Reflect.construct(Date, [], function () {})); return true; } catch (e) { return false; } }\n\nfunction _getPrototypeOf(o) { _getPrototypeOf = Object.setPrototypeOf ? Object.getPrototypeOf : function _getPrototypeOf(o) { return o.__proto__ || Object.getPrototypeOf(o); }; return _getPrototypeOf(o); }\n\n\n\n\n\n\n/**\n * Helper function to retrieve attribute value.\n * @param {String} suffix\n * @param {Element} element\n */\n\nfunction getAttributeValue(suffix, element) {\n var attribute = \"data-clipboard-\".concat(suffix);\n\n if (!element.hasAttribute(attribute)) {\n return;\n }\n\n return element.getAttribute(attribute);\n}\n/**\n * Base class which takes one or more elements, adds event listeners to them,\n * and instantiates a new `ClipboardAction` on each click.\n */\n\n\nvar Clipboard = /*#__PURE__*/function (_Emitter) {\n _inherits(Clipboard, _Emitter);\n\n var _super = _createSuper(Clipboard);\n\n /**\n * @param {String|HTMLElement|HTMLCollection|NodeList} trigger\n * @param {Object} options\n */\n function Clipboard(trigger, options) {\n var _this;\n\n _classCallCheck(this, Clipboard);\n\n _this = _super.call(this);\n\n _this.resolveOptions(options);\n\n _this.listenClick(trigger);\n\n return _this;\n }\n /**\n * Defines if attributes would be resolved using internal setter functions\n * or custom functions that were passed in the constructor.\n * @param {Object} options\n */\n\n\n _createClass(Clipboard, [{\n key: \"resolveOptions\",\n value: function resolveOptions() {\n var options = arguments.length > 0 && arguments[0] !== undefined ? arguments[0] : {};\n this.action = typeof options.action === 'function' ? options.action : this.defaultAction;\n this.target = typeof options.target === 'function' ? options.target : this.defaultTarget;\n this.text = typeof options.text === 'function' ? options.text : this.defaultText;\n this.container = clipboard_typeof(options.container) === 'object' ? options.container : document.body;\n }\n /**\n * Adds a click event listener to the passed trigger.\n * @param {String|HTMLElement|HTMLCollection|NodeList} trigger\n */\n\n }, {\n key: \"listenClick\",\n value: function listenClick(trigger) {\n var _this2 = this;\n\n this.listener = listen_default()(trigger, 'click', function (e) {\n return _this2.onClick(e);\n });\n }\n /**\n * Defines a new `ClipboardAction` on each click event.\n * @param {Event} e\n */\n\n }, {\n key: \"onClick\",\n value: function onClick(e) {\n var trigger = e.delegateTarget || e.currentTarget;\n var action = this.action(trigger) || 'copy';\n var text = actions_default({\n action: action,\n container: this.container,\n target: this.target(trigger),\n text: this.text(trigger)\n }); // Fires an event based on the copy operation result.\n\n this.emit(text ? 'success' : 'error', {\n action: action,\n text: text,\n trigger: trigger,\n clearSelection: function clearSelection() {\n if (trigger) {\n trigger.focus();\n }\n\n window.getSelection().removeAllRanges();\n }\n });\n }\n /**\n * Default `action` lookup function.\n * @param {Element} trigger\n */\n\n }, {\n key: \"defaultAction\",\n value: function defaultAction(trigger) {\n return getAttributeValue('action', trigger);\n }\n /**\n * Default `target` lookup function.\n * @param {Element} trigger\n */\n\n }, {\n key: \"defaultTarget\",\n value: function defaultTarget(trigger) {\n var selector = getAttributeValue('target', trigger);\n\n if (selector) {\n return document.querySelector(selector);\n }\n }\n /**\n * Allow fire programmatically a copy action\n * @param {String|HTMLElement} target\n * @param {Object} options\n * @returns Text copied.\n */\n\n }, {\n key: \"defaultText\",\n\n /**\n * Default `text` lookup function.\n * @param {Element} trigger\n */\n value: function defaultText(trigger) {\n return getAttributeValue('text', trigger);\n }\n /**\n * Destroy lifecycle.\n */\n\n }, {\n key: \"destroy\",\n value: function destroy() {\n this.listener.destroy();\n }\n }], [{\n key: \"copy\",\n value: function copy(target) {\n var options = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {\n container: document.body\n };\n return actions_copy(target, options);\n }\n /**\n * Allow fire programmatically a cut action\n * @param {String|HTMLElement} target\n * @returns Text cutted.\n */\n\n }, {\n key: \"cut\",\n value: function cut(target) {\n return actions_cut(target);\n }\n /**\n * Returns the support of the given action, or all actions if no action is\n * given.\n * @param {String} [action]\n */\n\n }, {\n key: \"isSupported\",\n value: function isSupported() {\n var action = arguments.length > 0 && arguments[0] !== undefined ? arguments[0] : ['copy', 'cut'];\n var actions = typeof action === 'string' ? [action] : action;\n var support = !!document.queryCommandSupported;\n actions.forEach(function (action) {\n support = support && !!document.queryCommandSupported(action);\n });\n return support;\n }\n }]);\n\n return Clipboard;\n}((tiny_emitter_default()));\n\n/* harmony default export */ var clipboard = (Clipboard);\n\n/***/ }),\n\n/***/ 828:\n/***/ (function(module) {\n\nvar DOCUMENT_NODE_TYPE = 9;\n\n/**\n * A polyfill for Element.matches()\n */\nif (typeof Element !== 'undefined' && !Element.prototype.matches) {\n var proto = Element.prototype;\n\n proto.matches = proto.matchesSelector ||\n proto.mozMatchesSelector ||\n proto.msMatchesSelector ||\n proto.oMatchesSelector ||\n proto.webkitMatchesSelector;\n}\n\n/**\n * Finds the closest parent that matches a selector.\n *\n * @param {Element} element\n * @param {String} selector\n * @return {Function}\n */\nfunction closest (element, selector) {\n while (element && element.nodeType !== DOCUMENT_NODE_TYPE) {\n if (typeof element.matches === 'function' &&\n element.matches(selector)) {\n return element;\n }\n element = element.parentNode;\n }\n}\n\nmodule.exports = closest;\n\n\n/***/ }),\n\n/***/ 438:\n/***/ (function(module, __unused_webpack_exports, __webpack_require__) {\n\nvar closest = __webpack_require__(828);\n\n/**\n * Delegates event to a selector.\n *\n * @param {Element} element\n * @param {String} selector\n * @param {String} type\n * @param {Function} callback\n * @param {Boolean} useCapture\n * @return {Object}\n */\nfunction _delegate(element, selector, type, callback, useCapture) {\n var listenerFn = listener.apply(this, arguments);\n\n element.addEventListener(type, listenerFn, useCapture);\n\n return {\n destroy: function() {\n element.removeEventListener(type, listenerFn, useCapture);\n }\n }\n}\n\n/**\n * Delegates event to a selector.\n *\n * @param {Element|String|Array} [elements]\n * @param {String} selector\n * @param {String} type\n * @param {Function} callback\n * @param {Boolean} useCapture\n * @return {Object}\n */\nfunction delegate(elements, selector, type, callback, useCapture) {\n // Handle the regular Element usage\n if (typeof elements.addEventListener === 'function') {\n return _delegate.apply(null, arguments);\n }\n\n // Handle Element-less usage, it defaults to global delegation\n if (typeof type === 'function') {\n // Use `document` as the first parameter, then apply arguments\n // This is a short way to .unshift `arguments` without running into deoptimizations\n return _delegate.bind(null, document).apply(null, arguments);\n }\n\n // Handle Selector-based usage\n if (typeof elements === 'string') {\n elements = document.querySelectorAll(elements);\n }\n\n // Handle Array-like based usage\n return Array.prototype.map.call(elements, function (element) {\n return _delegate(element, selector, type, callback, useCapture);\n });\n}\n\n/**\n * Finds closest match and invokes callback.\n *\n * @param {Element} element\n * @param {String} selector\n * @param {String} type\n * @param {Function} callback\n * @return {Function}\n */\nfunction listener(element, selector, type, callback) {\n return function(e) {\n e.delegateTarget = closest(e.target, selector);\n\n if (e.delegateTarget) {\n callback.call(element, e);\n }\n }\n}\n\nmodule.exports = delegate;\n\n\n/***/ }),\n\n/***/ 879:\n/***/ (function(__unused_webpack_module, exports) {\n\n/**\n * Check if argument is a HTML element.\n *\n * @param {Object} value\n * @return {Boolean}\n */\nexports.node = function(value) {\n return value !== undefined\n && value instanceof HTMLElement\n && value.nodeType === 1;\n};\n\n/**\n * Check if argument is a list of HTML elements.\n *\n * @param {Object} value\n * @return {Boolean}\n */\nexports.nodeList = function(value) {\n var type = Object.prototype.toString.call(value);\n\n return value !== undefined\n && (type === '[object NodeList]' || type === '[object HTMLCollection]')\n && ('length' in value)\n && (value.length === 0 || exports.node(value[0]));\n};\n\n/**\n * Check if argument is a string.\n *\n * @param {Object} value\n * @return {Boolean}\n */\nexports.string = function(value) {\n return typeof value === 'string'\n || value instanceof String;\n};\n\n/**\n * Check if argument is a function.\n *\n * @param {Object} value\n * @return {Boolean}\n */\nexports.fn = function(value) {\n var type = Object.prototype.toString.call(value);\n\n return type === '[object Function]';\n};\n\n\n/***/ }),\n\n/***/ 370:\n/***/ (function(module, __unused_webpack_exports, __webpack_require__) {\n\nvar is = __webpack_require__(879);\nvar delegate = __webpack_require__(438);\n\n/**\n * Validates all params and calls the right\n * listener function based on its target type.\n *\n * @param {String|HTMLElement|HTMLCollection|NodeList} target\n * @param {String} type\n * @param {Function} callback\n * @return {Object}\n */\nfunction listen(target, type, callback) {\n if (!target && !type && !callback) {\n throw new Error('Missing required arguments');\n }\n\n if (!is.string(type)) {\n throw new TypeError('Second argument must be a String');\n }\n\n if (!is.fn(callback)) {\n throw new TypeError('Third argument must be a Function');\n }\n\n if (is.node(target)) {\n return listenNode(target, type, callback);\n }\n else if (is.nodeList(target)) {\n return listenNodeList(target, type, callback);\n }\n else if (is.string(target)) {\n return listenSelector(target, type, callback);\n }\n else {\n throw new TypeError('First argument must be a String, HTMLElement, HTMLCollection, or NodeList');\n }\n}\n\n/**\n * Adds an event listener to a HTML element\n * and returns a remove listener function.\n *\n * @param {HTMLElement} node\n * @param {String} type\n * @param {Function} callback\n * @return {Object}\n */\nfunction listenNode(node, type, callback) {\n node.addEventListener(type, callback);\n\n return {\n destroy: function() {\n node.removeEventListener(type, callback);\n }\n }\n}\n\n/**\n * Add an event listener to a list of HTML elements\n * and returns a remove listener function.\n *\n * @param {NodeList|HTMLCollection} nodeList\n * @param {String} type\n * @param {Function} callback\n * @return {Object}\n */\nfunction listenNodeList(nodeList, type, callback) {\n Array.prototype.forEach.call(nodeList, function(node) {\n node.addEventListener(type, callback);\n });\n\n return {\n destroy: function() {\n Array.prototype.forEach.call(nodeList, function(node) {\n node.removeEventListener(type, callback);\n });\n }\n }\n}\n\n/**\n * Add an event listener to a selector\n * and returns a remove listener function.\n *\n * @param {String} selector\n * @param {String} type\n * @param {Function} callback\n * @return {Object}\n */\nfunction listenSelector(selector, type, callback) {\n return delegate(document.body, selector, type, callback);\n}\n\nmodule.exports = listen;\n\n\n/***/ }),\n\n/***/ 817:\n/***/ (function(module) {\n\nfunction select(element) {\n var selectedText;\n\n if (element.nodeName === 'SELECT') {\n element.focus();\n\n selectedText = element.value;\n }\n else if (element.nodeName === 'INPUT' || element.nodeName === 'TEXTAREA') {\n var isReadOnly = element.hasAttribute('readonly');\n\n if (!isReadOnly) {\n element.setAttribute('readonly', '');\n }\n\n element.select();\n element.setSelectionRange(0, element.value.length);\n\n if (!isReadOnly) {\n element.removeAttribute('readonly');\n }\n\n selectedText = element.value;\n }\n else {\n if (element.hasAttribute('contenteditable')) {\n element.focus();\n }\n\n var selection = window.getSelection();\n var range = document.createRange();\n\n range.selectNodeContents(element);\n selection.removeAllRanges();\n selection.addRange(range);\n\n selectedText = selection.toString();\n }\n\n return selectedText;\n}\n\nmodule.exports = select;\n\n\n/***/ }),\n\n/***/ 279:\n/***/ (function(module) {\n\nfunction E () {\n // Keep this empty so it's easier to inherit from\n // (via https://github.com/lipsmack from https://github.com/scottcorgan/tiny-emitter/issues/3)\n}\n\nE.prototype = {\n on: function (name, callback, ctx) {\n var e = this.e || (this.e = {});\n\n (e[name] || (e[name] = [])).push({\n fn: callback,\n ctx: ctx\n });\n\n return this;\n },\n\n once: function (name, callback, ctx) {\n var self = this;\n function listener () {\n self.off(name, listener);\n callback.apply(ctx, arguments);\n };\n\n listener._ = callback\n return this.on(name, listener, ctx);\n },\n\n emit: function (name) {\n var data = [].slice.call(arguments, 1);\n var evtArr = ((this.e || (this.e = {}))[name] || []).slice();\n var i = 0;\n var len = evtArr.length;\n\n for (i; i < len; i++) {\n evtArr[i].fn.apply(evtArr[i].ctx, data);\n }\n\n return this;\n },\n\n off: function (name, callback) {\n var e = this.e || (this.e = {});\n var evts = e[name];\n var liveEvents = [];\n\n if (evts && callback) {\n for (var i = 0, len = evts.length; i < len; i++) {\n if (evts[i].fn !== callback && evts[i].fn._ !== callback)\n liveEvents.push(evts[i]);\n }\n }\n\n // Remove event from queue to prevent memory leak\n // Suggested by https://github.com/lazd\n // Ref: https://github.com/scottcorgan/tiny-emitter/commit/c6ebfaa9bc973b33d110a84a307742b7cf94c953#commitcomment-5024910\n\n (liveEvents.length)\n ? e[name] = liveEvents\n : delete e[name];\n\n return this;\n }\n};\n\nmodule.exports = E;\nmodule.exports.TinyEmitter = E;\n\n\n/***/ })\n\n/******/ \t});\n/************************************************************************/\n/******/ \t// The module cache\n/******/ \tvar __webpack_module_cache__ = {};\n/******/ \t\n/******/ \t// The require function\n/******/ \tfunction __webpack_require__(moduleId) {\n/******/ \t\t// Check if module is in cache\n/******/ \t\tif(__webpack_module_cache__[moduleId]) {\n/******/ \t\t\treturn __webpack_module_cache__[moduleId].exports;\n/******/ \t\t}\n/******/ \t\t// Create a new module (and put it into the cache)\n/******/ \t\tvar module = __webpack_module_cache__[moduleId] = {\n/******/ \t\t\t// no module.id needed\n/******/ \t\t\t// no module.loaded needed\n/******/ \t\t\texports: {}\n/******/ \t\t};\n/******/ \t\n/******/ \t\t// Execute the module function\n/******/ \t\t__webpack_modules__[moduleId](module, module.exports, __webpack_require__);\n/******/ \t\n/******/ \t\t// Return the exports of the module\n/******/ \t\treturn module.exports;\n/******/ \t}\n/******/ \t\n/************************************************************************/\n/******/ \t/* webpack/runtime/compat get default export */\n/******/ \t!function() {\n/******/ \t\t// getDefaultExport function for compatibility with non-harmony modules\n/******/ \t\t__webpack_require__.n = function(module) {\n/******/ \t\t\tvar getter = module && module.__esModule ?\n/******/ \t\t\t\tfunction() { return module['default']; } :\n/******/ \t\t\t\tfunction() { return module; };\n/******/ \t\t\t__webpack_require__.d(getter, { a: getter });\n/******/ \t\t\treturn getter;\n/******/ \t\t};\n/******/ \t}();\n/******/ \t\n/******/ \t/* webpack/runtime/define property getters */\n/******/ \t!function() {\n/******/ \t\t// define getter functions for harmony exports\n/******/ \t\t__webpack_require__.d = function(exports, definition) {\n/******/ \t\t\tfor(var key in definition) {\n/******/ \t\t\t\tif(__webpack_require__.o(definition, key) && !__webpack_require__.o(exports, key)) {\n/******/ \t\t\t\t\tObject.defineProperty(exports, key, { enumerable: true, get: definition[key] });\n/******/ \t\t\t\t}\n/******/ \t\t\t}\n/******/ \t\t};\n/******/ \t}();\n/******/ \t\n/******/ \t/* webpack/runtime/hasOwnProperty shorthand */\n/******/ \t!function() {\n/******/ \t\t__webpack_require__.o = function(obj, prop) { return Object.prototype.hasOwnProperty.call(obj, prop); }\n/******/ \t}();\n/******/ \t\n/************************************************************************/\n/******/ \t// module exports must be returned from runtime so entry inlining is disabled\n/******/ \t// startup\n/******/ \t// Load entry module and return exports\n/******/ \treturn __webpack_require__(686);\n/******/ })()\n.default;\n});", "/*\n * Copyright (c) 2016-2024 Martin Donath \n *\n * Permission is hereby granted, free of charge, to any person obtaining a copy\n * of this software and associated documentation files (the \"Software\"), to\n * deal in the Software without restriction, including without limitation the\n * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or\n * sell copies of the Software, and to permit persons to whom the Software is\n * furnished to do so, subject to the following conditions:\n *\n * The above copyright notice and this permission notice shall be included in\n * all copies or substantial portions of the Software.\n *\n * THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL THE\n * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING\n * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS\n * IN THE SOFTWARE.\n */\n\nimport \"focus-visible\"\n\nimport {\n EMPTY,\n NEVER,\n Observable,\n Subject,\n defer,\n delay,\n filter,\n map,\n merge,\n mergeWith,\n shareReplay,\n switchMap\n} from \"rxjs\"\n\nimport { configuration, feature } from \"./_\"\nimport {\n at,\n getActiveElement,\n getOptionalElement,\n requestJSON,\n setLocation,\n setToggle,\n watchDocument,\n watchKeyboard,\n watchLocation,\n watchLocationTarget,\n watchMedia,\n watchPrint,\n watchScript,\n watchViewport\n} from \"./browser\"\nimport {\n getComponentElement,\n getComponentElements,\n mountAnnounce,\n mountBackToTop,\n mountConsent,\n mountContent,\n mountDialog,\n mountHeader,\n mountHeaderTitle,\n mountPalette,\n mountProgress,\n mountSearch,\n mountSearchHiglight,\n mountSidebar,\n mountSource,\n mountTableOfContents,\n mountTabs,\n watchHeader,\n watchMain\n} from \"./components\"\nimport {\n SearchIndex,\n setupClipboardJS,\n setupInstantNavigation,\n setupVersionSelector\n} from \"./integrations\"\nimport {\n patchEllipsis,\n patchIndeterminate,\n patchScrollfix,\n patchScrolllock\n} from \"./patches\"\nimport \"./polyfills\"\n\n/* ----------------------------------------------------------------------------\n * Functions - @todo refactor\n * ------------------------------------------------------------------------- */\n\n/**\n * Fetch search index\n *\n * @returns Search index observable\n */\nfunction fetchSearchIndex(): Observable {\n if (location.protocol === \"file:\") {\n return watchScript(\n `${new URL(\"search/search_index.js\", config.base)}`\n )\n .pipe(\n // @ts-ignore - @todo fix typings\n map(() => __index),\n shareReplay(1)\n )\n } else {\n return requestJSON(\n new URL(\"search/search_index.json\", config.base)\n )\n }\n}\n\n/* ----------------------------------------------------------------------------\n * Application\n * ------------------------------------------------------------------------- */\n\n/* Yay, JavaScript is available */\ndocument.documentElement.classList.remove(\"no-js\")\ndocument.documentElement.classList.add(\"js\")\n\n/* Set up navigation observables and subjects */\nconst document$ = watchDocument()\nconst location$ = watchLocation()\nconst target$ = watchLocationTarget(location$)\nconst keyboard$ = watchKeyboard()\n\n/* Set up media observables */\nconst viewport$ = watchViewport()\nconst tablet$ = watchMedia(\"(min-width: 960px)\")\nconst screen$ = watchMedia(\"(min-width: 1220px)\")\nconst print$ = watchPrint()\n\n/* Retrieve search index, if search is enabled */\nconst config = configuration()\nconst index$ = document.forms.namedItem(\"search\")\n ? fetchSearchIndex()\n : NEVER\n\n/* Set up Clipboard.js integration */\nconst alert$ = new Subject()\nsetupClipboardJS({ alert$ })\n\n/* Set up progress indicator */\nconst progress$ = new Subject()\n\n/* Set up instant navigation, if enabled */\nif (feature(\"navigation.instant\"))\n setupInstantNavigation({ location$, viewport$, progress$ })\n .subscribe(document$)\n\n/* Set up version selector */\nif (config.version?.provider === \"mike\")\n setupVersionSelector({ document$ })\n\n/* Always close drawer and search on navigation */\nmerge(location$, target$)\n .pipe(\n delay(125)\n )\n .subscribe(() => {\n setToggle(\"drawer\", false)\n setToggle(\"search\", false)\n })\n\n/* Set up global keyboard handlers */\nkeyboard$\n .pipe(\n filter(({ mode }) => mode === \"global\")\n )\n .subscribe(key => {\n switch (key.type) {\n\n /* Go to previous page */\n case \"p\":\n case \",\":\n const prev = getOptionalElement(\"link[rel=prev]\")\n if (typeof prev !== \"undefined\")\n setLocation(prev)\n break\n\n /* Go to next page */\n case \"n\":\n case \".\":\n const next = getOptionalElement(\"link[rel=next]\")\n if (typeof next !== \"undefined\")\n setLocation(next)\n break\n\n /* Expand navigation, see https://bit.ly/3ZjG5io */\n case \"Enter\":\n const active = getActiveElement()\n if (active instanceof HTMLLabelElement)\n active.click()\n }\n })\n\n/* Set up patches */\npatchEllipsis({ viewport$, document$ })\npatchIndeterminate({ document$, tablet$ })\npatchScrollfix({ document$ })\npatchScrolllock({ viewport$, tablet$ })\n\n/* Set up header and main area observable */\nconst header$ = watchHeader(getComponentElement(\"header\"), { viewport$ })\nconst main$ = document$\n .pipe(\n map(() => getComponentElement(\"main\")),\n switchMap(el => watchMain(el, { viewport$, header$ })),\n shareReplay(1)\n )\n\n/* Set up control component observables */\nconst control$ = merge(\n\n /* Consent */\n ...getComponentElements(\"consent\")\n .map(el => mountConsent(el, { target$ })),\n\n /* Dialog */\n ...getComponentElements(\"dialog\")\n .map(el => mountDialog(el, { alert$ })),\n\n /* Header */\n ...getComponentElements(\"header\")\n .map(el => mountHeader(el, { viewport$, header$, main$ })),\n\n /* Color palette */\n ...getComponentElements(\"palette\")\n .map(el => mountPalette(el)),\n\n /* Progress bar */\n ...getComponentElements(\"progress\")\n .map(el => mountProgress(el, { progress$ })),\n\n /* Search */\n ...getComponentElements(\"search\")\n .map(el => mountSearch(el, { index$, keyboard$ })),\n\n /* Repository information */\n ...getComponentElements(\"source\")\n .map(el => mountSource(el))\n)\n\n/* Set up content component observables */\nconst content$ = defer(() => merge(\n\n /* Announcement bar */\n ...getComponentElements(\"announce\")\n .map(el => mountAnnounce(el)),\n\n /* Content */\n ...getComponentElements(\"content\")\n .map(el => mountContent(el, { viewport$, target$, print$ })),\n\n /* Search highlighting */\n ...getComponentElements(\"content\")\n .map(el => feature(\"search.highlight\")\n ? mountSearchHiglight(el, { index$, location$ })\n : EMPTY\n ),\n\n /* Header title */\n ...getComponentElements(\"header-title\")\n .map(el => mountHeaderTitle(el, { viewport$, header$ })),\n\n /* Sidebar */\n ...getComponentElements(\"sidebar\")\n .map(el => el.getAttribute(\"data-md-type\") === \"navigation\"\n ? at(screen$, () => mountSidebar(el, { viewport$, header$, main$ }))\n : at(tablet$, () => mountSidebar(el, { viewport$, header$, main$ }))\n ),\n\n /* Navigation tabs */\n ...getComponentElements(\"tabs\")\n .map(el => mountTabs(el, { viewport$, header$ })),\n\n /* Table of contents */\n ...getComponentElements(\"toc\")\n .map(el => mountTableOfContents(el, {\n viewport$, header$, main$, target$\n })),\n\n /* Back-to-top button */\n ...getComponentElements(\"top\")\n .map(el => mountBackToTop(el, { viewport$, header$, main$, target$ }))\n))\n\n/* Set up component observables */\nconst component$ = document$\n .pipe(\n switchMap(() => content$),\n mergeWith(control$),\n shareReplay(1)\n )\n\n/* Subscribe to all components */\ncomponent$.subscribe()\n\n/* ----------------------------------------------------------------------------\n * Exports\n * ------------------------------------------------------------------------- */\n\nwindow.document$ = document$ /* Document observable */\nwindow.location$ = location$ /* Location subject */\nwindow.target$ = target$ /* Location target observable */\nwindow.keyboard$ = keyboard$ /* Keyboard observable */\nwindow.viewport$ = viewport$ /* Viewport observable */\nwindow.tablet$ = tablet$ /* Media tablet observable */\nwindow.screen$ = screen$ /* Media screen observable */\nwindow.print$ = print$ /* Media print observable */\nwindow.alert$ = alert$ /* Alert subject */\nwindow.progress$ = progress$ /* Progress indicator subject */\nwindow.component$ = component$ /* Component observable */\n", "/******************************************************************************\nCopyright (c) Microsoft Corporation.\n\nPermission to use, copy, modify, and/or distribute this software for any\npurpose with or without fee is hereby granted.\n\nTHE SOFTWARE IS PROVIDED \"AS IS\" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH\nREGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY\nAND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,\nINDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM\nLOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR\nOTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR\nPERFORMANCE OF THIS SOFTWARE.\n***************************************************************************** */\n/* global Reflect, Promise, SuppressedError, Symbol, Iterator */\n\nvar extendStatics = function(d, b) {\n extendStatics = Object.setPrototypeOf ||\n ({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||\n function (d, b) { for (var p in b) if (Object.prototype.hasOwnProperty.call(b, p)) d[p] = b[p]; };\n return extendStatics(d, b);\n};\n\nexport function __extends(d, b) {\n if (typeof b !== \"function\" && b !== null)\n throw new TypeError(\"Class extends value \" + String(b) + \" is not a constructor or null\");\n extendStatics(d, b);\n function __() { this.constructor = d; }\n d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());\n}\n\nexport var __assign = function() {\n __assign = Object.assign || function __assign(t) {\n for (var s, i = 1, n = arguments.length; i < n; i++) {\n s = arguments[i];\n for (var p in s) if (Object.prototype.hasOwnProperty.call(s, p)) t[p] = s[p];\n }\n return t;\n }\n return __assign.apply(this, arguments);\n}\n\nexport function __rest(s, e) {\n var t = {};\n for (var p in s) if (Object.prototype.hasOwnProperty.call(s, p) && e.indexOf(p) < 0)\n t[p] = s[p];\n if (s != null && typeof Object.getOwnPropertySymbols === \"function\")\n for (var i = 0, p = Object.getOwnPropertySymbols(s); i < p.length; i++) {\n if (e.indexOf(p[i]) < 0 && Object.prototype.propertyIsEnumerable.call(s, p[i]))\n t[p[i]] = s[p[i]];\n }\n return t;\n}\n\nexport function __decorate(decorators, target, key, desc) {\n var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;\n if (typeof Reflect === \"object\" && typeof Reflect.decorate === \"function\") r = Reflect.decorate(decorators, target, key, desc);\n else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;\n return c > 3 && r && Object.defineProperty(target, key, r), r;\n}\n\nexport function __param(paramIndex, decorator) {\n return function (target, key) { decorator(target, key, paramIndex); }\n}\n\nexport function __esDecorate(ctor, descriptorIn, decorators, contextIn, initializers, extraInitializers) {\n function accept(f) { if (f !== void 0 && typeof f !== \"function\") throw new TypeError(\"Function expected\"); return f; }\n var kind = contextIn.kind, key = kind === \"getter\" ? \"get\" : kind === \"setter\" ? \"set\" : \"value\";\n var target = !descriptorIn && ctor ? contextIn[\"static\"] ? ctor : ctor.prototype : null;\n var descriptor = descriptorIn || (target ? Object.getOwnPropertyDescriptor(target, contextIn.name) : {});\n var _, done = false;\n for (var i = decorators.length - 1; i >= 0; i--) {\n var context = {};\n for (var p in contextIn) context[p] = p === \"access\" ? {} : contextIn[p];\n for (var p in contextIn.access) context.access[p] = contextIn.access[p];\n context.addInitializer = function (f) { if (done) throw new TypeError(\"Cannot add initializers after decoration has completed\"); extraInitializers.push(accept(f || null)); };\n var result = (0, decorators[i])(kind === \"accessor\" ? { get: descriptor.get, set: descriptor.set } : descriptor[key], context);\n if (kind === \"accessor\") {\n if (result === void 0) continue;\n if (result === null || typeof result !== \"object\") throw new TypeError(\"Object expected\");\n if (_ = accept(result.get)) descriptor.get = _;\n if (_ = accept(result.set)) descriptor.set = _;\n if (_ = accept(result.init)) initializers.unshift(_);\n }\n else if (_ = accept(result)) {\n if (kind === \"field\") initializers.unshift(_);\n else descriptor[key] = _;\n }\n }\n if (target) Object.defineProperty(target, contextIn.name, descriptor);\n done = true;\n};\n\nexport function __runInitializers(thisArg, initializers, value) {\n var useValue = arguments.length > 2;\n for (var i = 0; i < initializers.length; i++) {\n value = useValue ? initializers[i].call(thisArg, value) : initializers[i].call(thisArg);\n }\n return useValue ? value : void 0;\n};\n\nexport function __propKey(x) {\n return typeof x === \"symbol\" ? x : \"\".concat(x);\n};\n\nexport function __setFunctionName(f, name, prefix) {\n if (typeof name === \"symbol\") name = name.description ? \"[\".concat(name.description, \"]\") : \"\";\n return Object.defineProperty(f, \"name\", { configurable: true, value: prefix ? \"\".concat(prefix, \" \", name) : name });\n};\n\nexport function __metadata(metadataKey, metadataValue) {\n if (typeof Reflect === \"object\" && typeof Reflect.metadata === \"function\") return Reflect.metadata(metadataKey, metadataValue);\n}\n\nexport function __awaiter(thisArg, _arguments, P, generator) {\n function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }\n return new (P || (P = Promise))(function (resolve, reject) {\n function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }\n function rejected(value) { try { step(generator[\"throw\"](value)); } catch (e) { reject(e); } }\n function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }\n step((generator = generator.apply(thisArg, _arguments || [])).next());\n });\n}\n\nexport function __generator(thisArg, body) {\n var _ = { label: 0, sent: function() { if (t[0] & 1) throw t[1]; return t[1]; }, trys: [], ops: [] }, f, y, t, g = Object.create((typeof Iterator === \"function\" ? Iterator : Object).prototype);\n return g.next = verb(0), g[\"throw\"] = verb(1), g[\"return\"] = verb(2), typeof Symbol === \"function\" && (g[Symbol.iterator] = function() { return this; }), g;\n function verb(n) { return function (v) { return step([n, v]); }; }\n function step(op) {\n if (f) throw new TypeError(\"Generator is already executing.\");\n while (g && (g = 0, op[0] && (_ = 0)), _) try {\n if (f = 1, y && (t = op[0] & 2 ? y[\"return\"] : op[0] ? y[\"throw\"] || ((t = y[\"return\"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;\n if (y = 0, t) op = [op[0] & 2, t.value];\n switch (op[0]) {\n case 0: case 1: t = op; break;\n case 4: _.label++; return { value: op[1], done: false };\n case 5: _.label++; y = op[1]; op = [0]; continue;\n case 7: op = _.ops.pop(); _.trys.pop(); continue;\n default:\n if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) { _ = 0; continue; }\n if (op[0] === 3 && (!t || (op[1] > t[0] && op[1] < t[3]))) { _.label = op[1]; break; }\n if (op[0] === 6 && _.label < t[1]) { _.label = t[1]; t = op; break; }\n if (t && _.label < t[2]) { _.label = t[2]; _.ops.push(op); break; }\n if (t[2]) _.ops.pop();\n _.trys.pop(); continue;\n }\n op = body.call(thisArg, _);\n } catch (e) { op = [6, e]; y = 0; } finally { f = t = 0; }\n if (op[0] & 5) throw op[1]; return { value: op[0] ? op[1] : void 0, done: true };\n }\n}\n\nexport var __createBinding = Object.create ? (function(o, m, k, k2) {\n if (k2 === undefined) k2 = k;\n var desc = Object.getOwnPropertyDescriptor(m, k);\n if (!desc || (\"get\" in desc ? !m.__esModule : desc.writable || desc.configurable)) {\n desc = { enumerable: true, get: function() { return m[k]; } };\n }\n Object.defineProperty(o, k2, desc);\n}) : (function(o, m, k, k2) {\n if (k2 === undefined) k2 = k;\n o[k2] = m[k];\n});\n\nexport function __exportStar(m, o) {\n for (var p in m) if (p !== \"default\" && !Object.prototype.hasOwnProperty.call(o, p)) __createBinding(o, m, p);\n}\n\nexport function __values(o) {\n var s = typeof Symbol === \"function\" && Symbol.iterator, m = s && o[s], i = 0;\n if (m) return m.call(o);\n if (o && typeof o.length === \"number\") return {\n next: function () {\n if (o && i >= o.length) o = void 0;\n return { value: o && o[i++], done: !o };\n }\n };\n throw new TypeError(s ? \"Object is not iterable.\" : \"Symbol.iterator is not defined.\");\n}\n\nexport function __read(o, n) {\n var m = typeof Symbol === \"function\" && o[Symbol.iterator];\n if (!m) return o;\n var i = m.call(o), r, ar = [], e;\n try {\n while ((n === void 0 || n-- > 0) && !(r = i.next()).done) ar.push(r.value);\n }\n catch (error) { e = { error: error }; }\n finally {\n try {\n if (r && !r.done && (m = i[\"return\"])) m.call(i);\n }\n finally { if (e) throw e.error; }\n }\n return ar;\n}\n\n/** @deprecated */\nexport function __spread() {\n for (var ar = [], i = 0; i < arguments.length; i++)\n ar = ar.concat(__read(arguments[i]));\n return ar;\n}\n\n/** @deprecated */\nexport function __spreadArrays() {\n for (var s = 0, i = 0, il = arguments.length; i < il; i++) s += arguments[i].length;\n for (var r = Array(s), k = 0, i = 0; i < il; i++)\n for (var a = arguments[i], j = 0, jl = a.length; j < jl; j++, k++)\n r[k] = a[j];\n return r;\n}\n\nexport function __spreadArray(to, from, pack) {\n if (pack || arguments.length === 2) for (var i = 0, l = from.length, ar; i < l; i++) {\n if (ar || !(i in from)) {\n if (!ar) ar = Array.prototype.slice.call(from, 0, i);\n ar[i] = from[i];\n }\n }\n return to.concat(ar || Array.prototype.slice.call(from));\n}\n\nexport function __await(v) {\n return this instanceof __await ? (this.v = v, this) : new __await(v);\n}\n\nexport function __asyncGenerator(thisArg, _arguments, generator) {\n if (!Symbol.asyncIterator) throw new TypeError(\"Symbol.asyncIterator is not defined.\");\n var g = generator.apply(thisArg, _arguments || []), i, q = [];\n return i = Object.create((typeof AsyncIterator === \"function\" ? AsyncIterator : Object).prototype), verb(\"next\"), verb(\"throw\"), verb(\"return\", awaitReturn), i[Symbol.asyncIterator] = function () { return this; }, i;\n function awaitReturn(f) { return function (v) { return Promise.resolve(v).then(f, reject); }; }\n function verb(n, f) { if (g[n]) { i[n] = function (v) { return new Promise(function (a, b) { q.push([n, v, a, b]) > 1 || resume(n, v); }); }; if (f) i[n] = f(i[n]); } }\n function resume(n, v) { try { step(g[n](v)); } catch (e) { settle(q[0][3], e); } }\n function step(r) { r.value instanceof __await ? Promise.resolve(r.value.v).then(fulfill, reject) : settle(q[0][2], r); }\n function fulfill(value) { resume(\"next\", value); }\n function reject(value) { resume(\"throw\", value); }\n function settle(f, v) { if (f(v), q.shift(), q.length) resume(q[0][0], q[0][1]); }\n}\n\nexport function __asyncDelegator(o) {\n var i, p;\n return i = {}, verb(\"next\"), verb(\"throw\", function (e) { throw e; }), verb(\"return\"), i[Symbol.iterator] = function () { return this; }, i;\n function verb(n, f) { i[n] = o[n] ? function (v) { return (p = !p) ? { value: __await(o[n](v)), done: false } : f ? f(v) : v; } : f; }\n}\n\nexport function __asyncValues(o) {\n if (!Symbol.asyncIterator) throw new TypeError(\"Symbol.asyncIterator is not defined.\");\n var m = o[Symbol.asyncIterator], i;\n return m ? m.call(o) : (o = typeof __values === \"function\" ? __values(o) : o[Symbol.iterator](), i = {}, verb(\"next\"), verb(\"throw\"), verb(\"return\"), i[Symbol.asyncIterator] = function () { return this; }, i);\n function verb(n) { i[n] = o[n] && function (v) { return new Promise(function (resolve, reject) { v = o[n](v), settle(resolve, reject, v.done, v.value); }); }; }\n function settle(resolve, reject, d, v) { Promise.resolve(v).then(function(v) { resolve({ value: v, done: d }); }, reject); }\n}\n\nexport function __makeTemplateObject(cooked, raw) {\n if (Object.defineProperty) { Object.defineProperty(cooked, \"raw\", { value: raw }); } else { cooked.raw = raw; }\n return cooked;\n};\n\nvar __setModuleDefault = Object.create ? (function(o, v) {\n Object.defineProperty(o, \"default\", { enumerable: true, value: v });\n}) : function(o, v) {\n o[\"default\"] = v;\n};\n\nexport function __importStar(mod) {\n if (mod && mod.__esModule) return mod;\n var result = {};\n if (mod != null) for (var k in mod) if (k !== \"default\" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);\n __setModuleDefault(result, mod);\n return result;\n}\n\nexport function __importDefault(mod) {\n return (mod && mod.__esModule) ? mod : { default: mod };\n}\n\nexport function __classPrivateFieldGet(receiver, state, kind, f) {\n if (kind === \"a\" && !f) throw new TypeError(\"Private accessor was defined without a getter\");\n if (typeof state === \"function\" ? receiver !== state || !f : !state.has(receiver)) throw new TypeError(\"Cannot read private member from an object whose class did not declare it\");\n return kind === \"m\" ? f : kind === \"a\" ? f.call(receiver) : f ? f.value : state.get(receiver);\n}\n\nexport function __classPrivateFieldSet(receiver, state, value, kind, f) {\n if (kind === \"m\") throw new TypeError(\"Private method is not writable\");\n if (kind === \"a\" && !f) throw new TypeError(\"Private accessor was defined without a setter\");\n if (typeof state === \"function\" ? receiver !== state || !f : !state.has(receiver)) throw new TypeError(\"Cannot write private member to an object whose class did not declare it\");\n return (kind === \"a\" ? f.call(receiver, value) : f ? f.value = value : state.set(receiver, value)), value;\n}\n\nexport function __classPrivateFieldIn(state, receiver) {\n if (receiver === null || (typeof receiver !== \"object\" && typeof receiver !== \"function\")) throw new TypeError(\"Cannot use 'in' operator on non-object\");\n return typeof state === \"function\" ? receiver === state : state.has(receiver);\n}\n\nexport function __addDisposableResource(env, value, async) {\n if (value !== null && value !== void 0) {\n if (typeof value !== \"object\" && typeof value !== \"function\") throw new TypeError(\"Object expected.\");\n var dispose, inner;\n if (async) {\n if (!Symbol.asyncDispose) throw new TypeError(\"Symbol.asyncDispose is not defined.\");\n dispose = value[Symbol.asyncDispose];\n }\n if (dispose === void 0) {\n if (!Symbol.dispose) throw new TypeError(\"Symbol.dispose is not defined.\");\n dispose = value[Symbol.dispose];\n if (async) inner = dispose;\n }\n if (typeof dispose !== \"function\") throw new TypeError(\"Object not disposable.\");\n if (inner) dispose = function() { try { inner.call(this); } catch (e) { return Promise.reject(e); } };\n env.stack.push({ value: value, dispose: dispose, async: async });\n }\n else if (async) {\n env.stack.push({ async: true });\n }\n return value;\n}\n\nvar _SuppressedError = typeof SuppressedError === \"function\" ? SuppressedError : function (error, suppressed, message) {\n var e = new Error(message);\n return e.name = \"SuppressedError\", e.error = error, e.suppressed = suppressed, e;\n};\n\nexport function __disposeResources(env) {\n function fail(e) {\n env.error = env.hasError ? new _SuppressedError(e, env.error, \"An error was suppressed during disposal.\") : e;\n env.hasError = true;\n }\n var r, s = 0;\n function next() {\n while (r = env.stack.pop()) {\n try {\n if (!r.async && s === 1) return s = 0, env.stack.push(r), Promise.resolve().then(next);\n if (r.dispose) {\n var result = r.dispose.call(r.value);\n if (r.async) return s |= 2, Promise.resolve(result).then(next, function(e) { fail(e); return next(); });\n }\n else s |= 1;\n }\n catch (e) {\n fail(e);\n }\n }\n if (s === 1) return env.hasError ? Promise.reject(env.error) : Promise.resolve();\n if (env.hasError) throw env.error;\n }\n return next();\n}\n\nexport default {\n __extends,\n __assign,\n __rest,\n __decorate,\n __param,\n __metadata,\n __awaiter,\n __generator,\n __createBinding,\n __exportStar,\n __values,\n __read,\n __spread,\n __spreadArrays,\n __spreadArray,\n __await,\n __asyncGenerator,\n __asyncDelegator,\n __asyncValues,\n __makeTemplateObject,\n __importStar,\n __importDefault,\n __classPrivateFieldGet,\n __classPrivateFieldSet,\n __classPrivateFieldIn,\n __addDisposableResource,\n __disposeResources,\n};\n", "/**\n * Returns true if the object is a function.\n * @param value The value to check\n */\nexport function isFunction(value: any): value is (...args: any[]) => any {\n return typeof value === 'function';\n}\n", "/**\n * Used to create Error subclasses until the community moves away from ES5.\n *\n * This is because compiling from TypeScript down to ES5 has issues with subclassing Errors\n * as well as other built-in types: https://github.com/Microsoft/TypeScript/issues/12123\n *\n * @param createImpl A factory function to create the actual constructor implementation. The returned\n * function should be a named function that calls `_super` internally.\n */\nexport function createErrorClass(createImpl: (_super: any) => any): T {\n const _super = (instance: any) => {\n Error.call(instance);\n instance.stack = new Error().stack;\n };\n\n const ctorFunc = createImpl(_super);\n ctorFunc.prototype = Object.create(Error.prototype);\n ctorFunc.prototype.constructor = ctorFunc;\n return ctorFunc;\n}\n", "import { createErrorClass } from './createErrorClass';\n\nexport interface UnsubscriptionError extends Error {\n readonly errors: any[];\n}\n\nexport interface UnsubscriptionErrorCtor {\n /**\n * @deprecated Internal implementation detail. Do not construct error instances.\n * Cannot be tagged as internal: https://github.com/ReactiveX/rxjs/issues/6269\n */\n new (errors: any[]): UnsubscriptionError;\n}\n\n/**\n * An error thrown when one or more errors have occurred during the\n * `unsubscribe` of a {@link Subscription}.\n */\nexport const UnsubscriptionError: UnsubscriptionErrorCtor = createErrorClass(\n (_super) =>\n function UnsubscriptionErrorImpl(this: any, errors: (Error | string)[]) {\n _super(this);\n this.message = errors\n ? `${errors.length} errors occurred during unsubscription:\n${errors.map((err, i) => `${i + 1}) ${err.toString()}`).join('\\n ')}`\n : '';\n this.name = 'UnsubscriptionError';\n this.errors = errors;\n }\n);\n", "/**\n * Removes an item from an array, mutating it.\n * @param arr The array to remove the item from\n * @param item The item to remove\n */\nexport function arrRemove(arr: T[] | undefined | null, item: T) {\n if (arr) {\n const index = arr.indexOf(item);\n 0 <= index && arr.splice(index, 1);\n }\n}\n", "import { isFunction } from './util/isFunction';\nimport { UnsubscriptionError } from './util/UnsubscriptionError';\nimport { SubscriptionLike, TeardownLogic, Unsubscribable } from './types';\nimport { arrRemove } from './util/arrRemove';\n\n/**\n * Represents a disposable resource, such as the execution of an Observable. A\n * Subscription has one important method, `unsubscribe`, that takes no argument\n * and just disposes the resource held by the subscription.\n *\n * Additionally, subscriptions may be grouped together through the `add()`\n * method, which will attach a child Subscription to the current Subscription.\n * When a Subscription is unsubscribed, all its children (and its grandchildren)\n * will be unsubscribed as well.\n *\n * @class Subscription\n */\nexport class Subscription implements SubscriptionLike {\n /** @nocollapse */\n public static EMPTY = (() => {\n const empty = new Subscription();\n empty.closed = true;\n return empty;\n })();\n\n /**\n * A flag to indicate whether this Subscription has already been unsubscribed.\n */\n public closed = false;\n\n private _parentage: Subscription[] | Subscription | null = null;\n\n /**\n * The list of registered finalizers to execute upon unsubscription. Adding and removing from this\n * list occurs in the {@link #add} and {@link #remove} methods.\n */\n private _finalizers: Exclude[] | null = null;\n\n /**\n * @param initialTeardown A function executed first as part of the finalization\n * process that is kicked off when {@link #unsubscribe} is called.\n */\n constructor(private initialTeardown?: () => void) {}\n\n /**\n * Disposes the resources held by the subscription. May, for instance, cancel\n * an ongoing Observable execution or cancel any other type of work that\n * started when the Subscription was created.\n * @return {void}\n */\n unsubscribe(): void {\n let errors: any[] | undefined;\n\n if (!this.closed) {\n this.closed = true;\n\n // Remove this from it's parents.\n const { _parentage } = this;\n if (_parentage) {\n this._parentage = null;\n if (Array.isArray(_parentage)) {\n for (const parent of _parentage) {\n parent.remove(this);\n }\n } else {\n _parentage.remove(this);\n }\n }\n\n const { initialTeardown: initialFinalizer } = this;\n if (isFunction(initialFinalizer)) {\n try {\n initialFinalizer();\n } catch (e) {\n errors = e instanceof UnsubscriptionError ? e.errors : [e];\n }\n }\n\n const { _finalizers } = this;\n if (_finalizers) {\n this._finalizers = null;\n for (const finalizer of _finalizers) {\n try {\n execFinalizer(finalizer);\n } catch (err) {\n errors = errors ?? [];\n if (err instanceof UnsubscriptionError) {\n errors = [...errors, ...err.errors];\n } else {\n errors.push(err);\n }\n }\n }\n }\n\n if (errors) {\n throw new UnsubscriptionError(errors);\n }\n }\n }\n\n /**\n * Adds a finalizer to this subscription, so that finalization will be unsubscribed/called\n * when this subscription is unsubscribed. If this subscription is already {@link #closed},\n * because it has already been unsubscribed, then whatever finalizer is passed to it\n * will automatically be executed (unless the finalizer itself is also a closed subscription).\n *\n * Closed Subscriptions cannot be added as finalizers to any subscription. Adding a closed\n * subscription to a any subscription will result in no operation. (A noop).\n *\n * Adding a subscription to itself, or adding `null` or `undefined` will not perform any\n * operation at all. (A noop).\n *\n * `Subscription` instances that are added to this instance will automatically remove themselves\n * if they are unsubscribed. Functions and {@link Unsubscribable} objects that you wish to remove\n * will need to be removed manually with {@link #remove}\n *\n * @param teardown The finalization logic to add to this subscription.\n */\n add(teardown: TeardownLogic): void {\n // Only add the finalizer if it's not undefined\n // and don't add a subscription to itself.\n if (teardown && teardown !== this) {\n if (this.closed) {\n // If this subscription is already closed,\n // execute whatever finalizer is handed to it automatically.\n execFinalizer(teardown);\n } else {\n if (teardown instanceof Subscription) {\n // We don't add closed subscriptions, and we don't add the same subscription\n // twice. Subscription unsubscribe is idempotent.\n if (teardown.closed || teardown._hasParent(this)) {\n return;\n }\n teardown._addParent(this);\n }\n (this._finalizers = this._finalizers ?? []).push(teardown);\n }\n }\n }\n\n /**\n * Checks to see if a this subscription already has a particular parent.\n * This will signal that this subscription has already been added to the parent in question.\n * @param parent the parent to check for\n */\n private _hasParent(parent: Subscription) {\n const { _parentage } = this;\n return _parentage === parent || (Array.isArray(_parentage) && _parentage.includes(parent));\n }\n\n /**\n * Adds a parent to this subscription so it can be removed from the parent if it\n * unsubscribes on it's own.\n *\n * NOTE: THIS ASSUMES THAT {@link _hasParent} HAS ALREADY BEEN CHECKED.\n * @param parent The parent subscription to add\n */\n private _addParent(parent: Subscription) {\n const { _parentage } = this;\n this._parentage = Array.isArray(_parentage) ? (_parentage.push(parent), _parentage) : _parentage ? [_parentage, parent] : parent;\n }\n\n /**\n * Called on a child when it is removed via {@link #remove}.\n * @param parent The parent to remove\n */\n private _removeParent(parent: Subscription) {\n const { _parentage } = this;\n if (_parentage === parent) {\n this._parentage = null;\n } else if (Array.isArray(_parentage)) {\n arrRemove(_parentage, parent);\n }\n }\n\n /**\n * Removes a finalizer from this subscription that was previously added with the {@link #add} method.\n *\n * Note that `Subscription` instances, when unsubscribed, will automatically remove themselves\n * from every other `Subscription` they have been added to. This means that using the `remove` method\n * is not a common thing and should be used thoughtfully.\n *\n * If you add the same finalizer instance of a function or an unsubscribable object to a `Subscription` instance\n * more than once, you will need to call `remove` the same number of times to remove all instances.\n *\n * All finalizer instances are removed to free up memory upon unsubscription.\n *\n * @param teardown The finalizer to remove from this subscription\n */\n remove(teardown: Exclude): void {\n const { _finalizers } = this;\n _finalizers && arrRemove(_finalizers, teardown);\n\n if (teardown instanceof Subscription) {\n teardown._removeParent(this);\n }\n }\n}\n\nexport const EMPTY_SUBSCRIPTION = Subscription.EMPTY;\n\nexport function isSubscription(value: any): value is Subscription {\n return (\n value instanceof Subscription ||\n (value && 'closed' in value && isFunction(value.remove) && isFunction(value.add) && isFunction(value.unsubscribe))\n );\n}\n\nfunction execFinalizer(finalizer: Unsubscribable | (() => void)) {\n if (isFunction(finalizer)) {\n finalizer();\n } else {\n finalizer.unsubscribe();\n }\n}\n", "import { Subscriber } from './Subscriber';\nimport { ObservableNotification } from './types';\n\n/**\n * The {@link GlobalConfig} object for RxJS. It is used to configure things\n * like how to react on unhandled errors.\n */\nexport const config: GlobalConfig = {\n onUnhandledError: null,\n onStoppedNotification: null,\n Promise: undefined,\n useDeprecatedSynchronousErrorHandling: false,\n useDeprecatedNextContext: false,\n};\n\n/**\n * The global configuration object for RxJS, used to configure things\n * like how to react on unhandled errors. Accessible via {@link config}\n * object.\n */\nexport interface GlobalConfig {\n /**\n * A registration point for unhandled errors from RxJS. These are errors that\n * cannot were not handled by consuming code in the usual subscription path. For\n * example, if you have this configured, and you subscribe to an observable without\n * providing an error handler, errors from that subscription will end up here. This\n * will _always_ be called asynchronously on another job in the runtime. This is because\n * we do not want errors thrown in this user-configured handler to interfere with the\n * behavior of the library.\n */\n onUnhandledError: ((err: any) => void) | null;\n\n /**\n * A registration point for notifications that cannot be sent to subscribers because they\n * have completed, errored or have been explicitly unsubscribed. By default, next, complete\n * and error notifications sent to stopped subscribers are noops. However, sometimes callers\n * might want a different behavior. For example, with sources that attempt to report errors\n * to stopped subscribers, a caller can configure RxJS to throw an unhandled error instead.\n * This will _always_ be called asynchronously on another job in the runtime. This is because\n * we do not want errors thrown in this user-configured handler to interfere with the\n * behavior of the library.\n */\n onStoppedNotification: ((notification: ObservableNotification, subscriber: Subscriber) => void) | null;\n\n /**\n * The promise constructor used by default for {@link Observable#toPromise toPromise} and {@link Observable#forEach forEach}\n * methods.\n *\n * @deprecated As of version 8, RxJS will no longer support this sort of injection of a\n * Promise constructor. If you need a Promise implementation other than native promises,\n * please polyfill/patch Promise as you see appropriate. Will be removed in v8.\n */\n Promise?: PromiseConstructorLike;\n\n /**\n * If true, turns on synchronous error rethrowing, which is a deprecated behavior\n * in v6 and higher. This behavior enables bad patterns like wrapping a subscribe\n * call in a try/catch block. It also enables producer interference, a nasty bug\n * where a multicast can be broken for all observers by a downstream consumer with\n * an unhandled error. DO NOT USE THIS FLAG UNLESS IT'S NEEDED TO BUY TIME\n * FOR MIGRATION REASONS.\n *\n * @deprecated As of version 8, RxJS will no longer support synchronous throwing\n * of unhandled errors. All errors will be thrown on a separate call stack to prevent bad\n * behaviors described above. Will be removed in v8.\n */\n useDeprecatedSynchronousErrorHandling: boolean;\n\n /**\n * If true, enables an as-of-yet undocumented feature from v5: The ability to access\n * `unsubscribe()` via `this` context in `next` functions created in observers passed\n * to `subscribe`.\n *\n * This is being removed because the performance was severely problematic, and it could also cause\n * issues when types other than POJOs are passed to subscribe as subscribers, as they will likely have\n * their `this` context overwritten.\n *\n * @deprecated As of version 8, RxJS will no longer support altering the\n * context of next functions provided as part of an observer to Subscribe. Instead,\n * you will have access to a subscription or a signal or token that will allow you to do things like\n * unsubscribe and test closed status. Will be removed in v8.\n */\n useDeprecatedNextContext: boolean;\n}\n", "import type { TimerHandle } from './timerHandle';\ntype SetTimeoutFunction = (handler: () => void, timeout?: number, ...args: any[]) => TimerHandle;\ntype ClearTimeoutFunction = (handle: TimerHandle) => void;\n\ninterface TimeoutProvider {\n setTimeout: SetTimeoutFunction;\n clearTimeout: ClearTimeoutFunction;\n delegate:\n | {\n setTimeout: SetTimeoutFunction;\n clearTimeout: ClearTimeoutFunction;\n }\n | undefined;\n}\n\nexport const timeoutProvider: TimeoutProvider = {\n // When accessing the delegate, use the variable rather than `this` so that\n // the functions can be called without being bound to the provider.\n setTimeout(handler: () => void, timeout?: number, ...args) {\n const { delegate } = timeoutProvider;\n if (delegate?.setTimeout) {\n return delegate.setTimeout(handler, timeout, ...args);\n }\n return setTimeout(handler, timeout, ...args);\n },\n clearTimeout(handle) {\n const { delegate } = timeoutProvider;\n return (delegate?.clearTimeout || clearTimeout)(handle as any);\n },\n delegate: undefined,\n};\n", "import { config } from '../config';\nimport { timeoutProvider } from '../scheduler/timeoutProvider';\n\n/**\n * Handles an error on another job either with the user-configured {@link onUnhandledError},\n * or by throwing it on that new job so it can be picked up by `window.onerror`, `process.on('error')`, etc.\n *\n * This should be called whenever there is an error that is out-of-band with the subscription\n * or when an error hits a terminal boundary of the subscription and no error handler was provided.\n *\n * @param err the error to report\n */\nexport function reportUnhandledError(err: any) {\n timeoutProvider.setTimeout(() => {\n const { onUnhandledError } = config;\n if (onUnhandledError) {\n // Execute the user-configured error handler.\n onUnhandledError(err);\n } else {\n // Throw so it is picked up by the runtime's uncaught error mechanism.\n throw err;\n }\n });\n}\n", "/* tslint:disable:no-empty */\nexport function noop() { }\n", "import { CompleteNotification, NextNotification, ErrorNotification } from './types';\n\n/**\n * A completion object optimized for memory use and created to be the\n * same \"shape\" as other notifications in v8.\n * @internal\n */\nexport const COMPLETE_NOTIFICATION = (() => createNotification('C', undefined, undefined) as CompleteNotification)();\n\n/**\n * Internal use only. Creates an optimized error notification that is the same \"shape\"\n * as other notifications.\n * @internal\n */\nexport function errorNotification(error: any): ErrorNotification {\n return createNotification('E', undefined, error) as any;\n}\n\n/**\n * Internal use only. Creates an optimized next notification that is the same \"shape\"\n * as other notifications.\n * @internal\n */\nexport function nextNotification(value: T) {\n return createNotification('N', value, undefined) as NextNotification;\n}\n\n/**\n * Ensures that all notifications created internally have the same \"shape\" in v8.\n *\n * TODO: This is only exported to support a crazy legacy test in `groupBy`.\n * @internal\n */\nexport function createNotification(kind: 'N' | 'E' | 'C', value: any, error: any) {\n return {\n kind,\n value,\n error,\n };\n}\n", "import { config } from '../config';\n\nlet context: { errorThrown: boolean; error: any } | null = null;\n\n/**\n * Handles dealing with errors for super-gross mode. Creates a context, in which\n * any synchronously thrown errors will be passed to {@link captureError}. Which\n * will record the error such that it will be rethrown after the call back is complete.\n * TODO: Remove in v8\n * @param cb An immediately executed function.\n */\nexport function errorContext(cb: () => void) {\n if (config.useDeprecatedSynchronousErrorHandling) {\n const isRoot = !context;\n if (isRoot) {\n context = { errorThrown: false, error: null };\n }\n cb();\n if (isRoot) {\n const { errorThrown, error } = context!;\n context = null;\n if (errorThrown) {\n throw error;\n }\n }\n } else {\n // This is the general non-deprecated path for everyone that\n // isn't crazy enough to use super-gross mode (useDeprecatedSynchronousErrorHandling)\n cb();\n }\n}\n\n/**\n * Captures errors only in super-gross mode.\n * @param err the error to capture\n */\nexport function captureError(err: any) {\n if (config.useDeprecatedSynchronousErrorHandling && context) {\n context.errorThrown = true;\n context.error = err;\n }\n}\n", "import { isFunction } from './util/isFunction';\nimport { Observer, ObservableNotification } from './types';\nimport { isSubscription, Subscription } from './Subscription';\nimport { config } from './config';\nimport { reportUnhandledError } from './util/reportUnhandledError';\nimport { noop } from './util/noop';\nimport { nextNotification, errorNotification, COMPLETE_NOTIFICATION } from './NotificationFactories';\nimport { timeoutProvider } from './scheduler/timeoutProvider';\nimport { captureError } from './util/errorContext';\n\n/**\n * Implements the {@link Observer} interface and extends the\n * {@link Subscription} class. While the {@link Observer} is the public API for\n * consuming the values of an {@link Observable}, all Observers get converted to\n * a Subscriber, in order to provide Subscription-like capabilities such as\n * `unsubscribe`. Subscriber is a common type in RxJS, and crucial for\n * implementing operators, but it is rarely used as a public API.\n *\n * @class Subscriber\n */\nexport class Subscriber extends Subscription implements Observer {\n /**\n * A static factory for a Subscriber, given a (potentially partial) definition\n * of an Observer.\n * @param next The `next` callback of an Observer.\n * @param error The `error` callback of an\n * Observer.\n * @param complete The `complete` callback of an\n * Observer.\n * @return A Subscriber wrapping the (partially defined)\n * Observer represented by the given arguments.\n * @nocollapse\n * @deprecated Do not use. Will be removed in v8. There is no replacement for this\n * method, and there is no reason to be creating instances of `Subscriber` directly.\n * If you have a specific use case, please file an issue.\n */\n static create(next?: (x?: T) => void, error?: (e?: any) => void, complete?: () => void): Subscriber {\n return new SafeSubscriber(next, error, complete);\n }\n\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n protected isStopped: boolean = false;\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n protected destination: Subscriber | Observer; // this `any` is the escape hatch to erase extra type param (e.g. R)\n\n /**\n * @deprecated Internal implementation detail, do not use directly. Will be made internal in v8.\n * There is no reason to directly create an instance of Subscriber. This type is exported for typings reasons.\n */\n constructor(destination?: Subscriber | Observer) {\n super();\n if (destination) {\n this.destination = destination;\n // Automatically chain subscriptions together here.\n // if destination is a Subscription, then it is a Subscriber.\n if (isSubscription(destination)) {\n destination.add(this);\n }\n } else {\n this.destination = EMPTY_OBSERVER;\n }\n }\n\n /**\n * The {@link Observer} callback to receive notifications of type `next` from\n * the Observable, with a value. The Observable may call this method 0 or more\n * times.\n * @param {T} [value] The `next` value.\n * @return {void}\n */\n next(value?: T): void {\n if (this.isStopped) {\n handleStoppedNotification(nextNotification(value), this);\n } else {\n this._next(value!);\n }\n }\n\n /**\n * The {@link Observer} callback to receive notifications of type `error` from\n * the Observable, with an attached `Error`. Notifies the Observer that\n * the Observable has experienced an error condition.\n * @param {any} [err] The `error` exception.\n * @return {void}\n */\n error(err?: any): void {\n if (this.isStopped) {\n handleStoppedNotification(errorNotification(err), this);\n } else {\n this.isStopped = true;\n this._error(err);\n }\n }\n\n /**\n * The {@link Observer} callback to receive a valueless notification of type\n * `complete` from the Observable. Notifies the Observer that the Observable\n * has finished sending push-based notifications.\n * @return {void}\n */\n complete(): void {\n if (this.isStopped) {\n handleStoppedNotification(COMPLETE_NOTIFICATION, this);\n } else {\n this.isStopped = true;\n this._complete();\n }\n }\n\n unsubscribe(): void {\n if (!this.closed) {\n this.isStopped = true;\n super.unsubscribe();\n this.destination = null!;\n }\n }\n\n protected _next(value: T): void {\n this.destination.next(value);\n }\n\n protected _error(err: any): void {\n try {\n this.destination.error(err);\n } finally {\n this.unsubscribe();\n }\n }\n\n protected _complete(): void {\n try {\n this.destination.complete();\n } finally {\n this.unsubscribe();\n }\n }\n}\n\n/**\n * This bind is captured here because we want to be able to have\n * compatibility with monoid libraries that tend to use a method named\n * `bind`. In particular, a library called Monio requires this.\n */\nconst _bind = Function.prototype.bind;\n\nfunction bind any>(fn: Fn, thisArg: any): Fn {\n return _bind.call(fn, thisArg);\n}\n\n/**\n * Internal optimization only, DO NOT EXPOSE.\n * @internal\n */\nclass ConsumerObserver implements Observer {\n constructor(private partialObserver: Partial>) {}\n\n next(value: T): void {\n const { partialObserver } = this;\n if (partialObserver.next) {\n try {\n partialObserver.next(value);\n } catch (error) {\n handleUnhandledError(error);\n }\n }\n }\n\n error(err: any): void {\n const { partialObserver } = this;\n if (partialObserver.error) {\n try {\n partialObserver.error(err);\n } catch (error) {\n handleUnhandledError(error);\n }\n } else {\n handleUnhandledError(err);\n }\n }\n\n complete(): void {\n const { partialObserver } = this;\n if (partialObserver.complete) {\n try {\n partialObserver.complete();\n } catch (error) {\n handleUnhandledError(error);\n }\n }\n }\n}\n\nexport class SafeSubscriber extends Subscriber {\n constructor(\n observerOrNext?: Partial> | ((value: T) => void) | null,\n error?: ((e?: any) => void) | null,\n complete?: (() => void) | null\n ) {\n super();\n\n let partialObserver: Partial>;\n if (isFunction(observerOrNext) || !observerOrNext) {\n // The first argument is a function, not an observer. The next\n // two arguments *could* be observers, or they could be empty.\n partialObserver = {\n next: (observerOrNext ?? undefined) as (((value: T) => void) | undefined),\n error: error ?? undefined,\n complete: complete ?? undefined,\n };\n } else {\n // The first argument is a partial observer.\n let context: any;\n if (this && config.useDeprecatedNextContext) {\n // This is a deprecated path that made `this.unsubscribe()` available in\n // next handler functions passed to subscribe. This only exists behind a flag\n // now, as it is *very* slow.\n context = Object.create(observerOrNext);\n context.unsubscribe = () => this.unsubscribe();\n partialObserver = {\n next: observerOrNext.next && bind(observerOrNext.next, context),\n error: observerOrNext.error && bind(observerOrNext.error, context),\n complete: observerOrNext.complete && bind(observerOrNext.complete, context),\n };\n } else {\n // The \"normal\" path. Just use the partial observer directly.\n partialObserver = observerOrNext;\n }\n }\n\n // Wrap the partial observer to ensure it's a full observer, and\n // make sure proper error handling is accounted for.\n this.destination = new ConsumerObserver(partialObserver);\n }\n}\n\nfunction handleUnhandledError(error: any) {\n if (config.useDeprecatedSynchronousErrorHandling) {\n captureError(error);\n } else {\n // Ideal path, we report this as an unhandled error,\n // which is thrown on a new call stack.\n reportUnhandledError(error);\n }\n}\n\n/**\n * An error handler used when no error handler was supplied\n * to the SafeSubscriber -- meaning no error handler was supplied\n * do the `subscribe` call on our observable.\n * @param err The error to handle\n */\nfunction defaultErrorHandler(err: any) {\n throw err;\n}\n\n/**\n * A handler for notifications that cannot be sent to a stopped subscriber.\n * @param notification The notification being sent\n * @param subscriber The stopped subscriber\n */\nfunction handleStoppedNotification(notification: ObservableNotification, subscriber: Subscriber) {\n const { onStoppedNotification } = config;\n onStoppedNotification && timeoutProvider.setTimeout(() => onStoppedNotification(notification, subscriber));\n}\n\n/**\n * The observer used as a stub for subscriptions where the user did not\n * pass any arguments to `subscribe`. Comes with the default error handling\n * behavior.\n */\nexport const EMPTY_OBSERVER: Readonly> & { closed: true } = {\n closed: true,\n next: noop,\n error: defaultErrorHandler,\n complete: noop,\n};\n", "/**\n * Symbol.observable or a string \"@@observable\". Used for interop\n *\n * @deprecated We will no longer be exporting this symbol in upcoming versions of RxJS.\n * Instead polyfill and use Symbol.observable directly *or* use https://www.npmjs.com/package/symbol-observable\n */\nexport const observable: string | symbol = (() => (typeof Symbol === 'function' && Symbol.observable) || '@@observable')();\n", "/**\n * This function takes one parameter and just returns it. Simply put,\n * this is like `(x: T): T => x`.\n *\n * ## Examples\n *\n * This is useful in some cases when using things like `mergeMap`\n *\n * ```ts\n * import { interval, take, map, range, mergeMap, identity } from 'rxjs';\n *\n * const source$ = interval(1000).pipe(take(5));\n *\n * const result$ = source$.pipe(\n * map(i => range(i)),\n * mergeMap(identity) // same as mergeMap(x => x)\n * );\n *\n * result$.subscribe({\n * next: console.log\n * });\n * ```\n *\n * Or when you want to selectively apply an operator\n *\n * ```ts\n * import { interval, take, identity } from 'rxjs';\n *\n * const shouldLimit = () => Math.random() < 0.5;\n *\n * const source$ = interval(1000);\n *\n * const result$ = source$.pipe(shouldLimit() ? take(5) : identity);\n *\n * result$.subscribe({\n * next: console.log\n * });\n * ```\n *\n * @param x Any value that is returned by this function\n * @returns The value passed as the first parameter to this function\n */\nexport function identity(x: T): T {\n return x;\n}\n", "import { identity } from './identity';\nimport { UnaryFunction } from '../types';\n\nexport function pipe(): typeof identity;\nexport function pipe(fn1: UnaryFunction): UnaryFunction;\nexport function pipe(fn1: UnaryFunction, fn2: UnaryFunction): UnaryFunction;\nexport function pipe(fn1: UnaryFunction, fn2: UnaryFunction, fn3: UnaryFunction): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction\n): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction,\n fn5: UnaryFunction\n): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction,\n fn5: UnaryFunction,\n fn6: UnaryFunction\n): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction,\n fn5: UnaryFunction,\n fn6: UnaryFunction,\n fn7: UnaryFunction\n): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction,\n fn5: UnaryFunction,\n fn6: UnaryFunction,\n fn7: UnaryFunction,\n fn8: UnaryFunction\n): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction,\n fn5: UnaryFunction,\n fn6: UnaryFunction,\n fn7: UnaryFunction,\n fn8: UnaryFunction,\n fn9: UnaryFunction\n): UnaryFunction;\nexport function pipe(\n fn1: UnaryFunction,\n fn2: UnaryFunction,\n fn3: UnaryFunction,\n fn4: UnaryFunction,\n fn5: UnaryFunction,\n fn6: UnaryFunction,\n fn7: UnaryFunction,\n fn8: UnaryFunction,\n fn9: UnaryFunction,\n ...fns: UnaryFunction[]\n): UnaryFunction;\n\n/**\n * pipe() can be called on one or more functions, each of which can take one argument (\"UnaryFunction\")\n * and uses it to return a value.\n * It returns a function that takes one argument, passes it to the first UnaryFunction, and then\n * passes the result to the next one, passes that result to the next one, and so on. \n */\nexport function pipe(...fns: Array>): UnaryFunction {\n return pipeFromArray(fns);\n}\n\n/** @internal */\nexport function pipeFromArray(fns: Array>): UnaryFunction {\n if (fns.length === 0) {\n return identity as UnaryFunction;\n }\n\n if (fns.length === 1) {\n return fns[0];\n }\n\n return function piped(input: T): R {\n return fns.reduce((prev: any, fn: UnaryFunction) => fn(prev), input as any);\n };\n}\n", "import { Operator } from './Operator';\nimport { SafeSubscriber, Subscriber } from './Subscriber';\nimport { isSubscription, Subscription } from './Subscription';\nimport { TeardownLogic, OperatorFunction, Subscribable, Observer } from './types';\nimport { observable as Symbol_observable } from './symbol/observable';\nimport { pipeFromArray } from './util/pipe';\nimport { config } from './config';\nimport { isFunction } from './util/isFunction';\nimport { errorContext } from './util/errorContext';\n\n/**\n * A representation of any set of values over any amount of time. This is the most basic building block\n * of RxJS.\n *\n * @class Observable\n */\nexport class Observable implements Subscribable {\n /**\n * @deprecated Internal implementation detail, do not use directly. Will be made internal in v8.\n */\n source: Observable | undefined;\n\n /**\n * @deprecated Internal implementation detail, do not use directly. Will be made internal in v8.\n */\n operator: Operator | undefined;\n\n /**\n * @constructor\n * @param {Function} subscribe the function that is called when the Observable is\n * initially subscribed to. This function is given a Subscriber, to which new values\n * can be `next`ed, or an `error` method can be called to raise an error, or\n * `complete` can be called to notify of a successful completion.\n */\n constructor(subscribe?: (this: Observable, subscriber: Subscriber) => TeardownLogic) {\n if (subscribe) {\n this._subscribe = subscribe;\n }\n }\n\n // HACK: Since TypeScript inherits static properties too, we have to\n // fight against TypeScript here so Subject can have a different static create signature\n /**\n * Creates a new Observable by calling the Observable constructor\n * @owner Observable\n * @method create\n * @param {Function} subscribe? the subscriber function to be passed to the Observable constructor\n * @return {Observable} a new observable\n * @nocollapse\n * @deprecated Use `new Observable()` instead. Will be removed in v8.\n */\n static create: (...args: any[]) => any = (subscribe?: (subscriber: Subscriber) => TeardownLogic) => {\n return new Observable(subscribe);\n };\n\n /**\n * Creates a new Observable, with this Observable instance as the source, and the passed\n * operator defined as the new observable's operator.\n * @method lift\n * @param operator the operator defining the operation to take on the observable\n * @return a new observable with the Operator applied\n * @deprecated Internal implementation detail, do not use directly. Will be made internal in v8.\n * If you have implemented an operator using `lift`, it is recommended that you create an\n * operator by simply returning `new Observable()` directly. See \"Creating new operators from\n * scratch\" section here: https://rxjs.dev/guide/operators\n */\n lift(operator?: Operator): Observable {\n const observable = new Observable();\n observable.source = this;\n observable.operator = operator;\n return observable;\n }\n\n subscribe(observerOrNext?: Partial> | ((value: T) => void)): Subscription;\n /** @deprecated Instead of passing separate callback arguments, use an observer argument. Signatures taking separate callback arguments will be removed in v8. Details: https://rxjs.dev/deprecations/subscribe-arguments */\n subscribe(next?: ((value: T) => void) | null, error?: ((error: any) => void) | null, complete?: (() => void) | null): Subscription;\n /**\n * Invokes an execution of an Observable and registers Observer handlers for notifications it will emit.\n *\n * Use it when you have all these Observables, but still nothing is happening.\n *\n * `subscribe` is not a regular operator, but a method that calls Observable's internal `subscribe` function. It\n * might be for example a function that you passed to Observable's constructor, but most of the time it is\n * a library implementation, which defines what will be emitted by an Observable, and when it be will emitted. This means\n * that calling `subscribe` is actually the moment when Observable starts its work, not when it is created, as it is often\n * the thought.\n *\n * Apart from starting the execution of an Observable, this method allows you to listen for values\n * that an Observable emits, as well as for when it completes or errors. You can achieve this in two\n * of the following ways.\n *\n * The first way is creating an object that implements {@link Observer} interface. It should have methods\n * defined by that interface, but note that it should be just a regular JavaScript object, which you can create\n * yourself in any way you want (ES6 class, classic function constructor, object literal etc.). In particular, do\n * not attempt to use any RxJS implementation details to create Observers - you don't need them. Remember also\n * that your object does not have to implement all methods. If you find yourself creating a method that doesn't\n * do anything, you can simply omit it. Note however, if the `error` method is not provided and an error happens,\n * it will be thrown asynchronously. Errors thrown asynchronously cannot be caught using `try`/`catch`. Instead,\n * use the {@link onUnhandledError} configuration option or use a runtime handler (like `window.onerror` or\n * `process.on('error)`) to be notified of unhandled errors. Because of this, it's recommended that you provide\n * an `error` method to avoid missing thrown errors.\n *\n * The second way is to give up on Observer object altogether and simply provide callback functions in place of its methods.\n * This means you can provide three functions as arguments to `subscribe`, where the first function is equivalent\n * of a `next` method, the second of an `error` method and the third of a `complete` method. Just as in case of an Observer,\n * if you do not need to listen for something, you can omit a function by passing `undefined` or `null`,\n * since `subscribe` recognizes these functions by where they were placed in function call. When it comes\n * to the `error` function, as with an Observer, if not provided, errors emitted by an Observable will be thrown asynchronously.\n *\n * You can, however, subscribe with no parameters at all. This may be the case where you're not interested in terminal events\n * and you also handled emissions internally by using operators (e.g. using `tap`).\n *\n * Whichever style of calling `subscribe` you use, in both cases it returns a Subscription object.\n * This object allows you to call `unsubscribe` on it, which in turn will stop the work that an Observable does and will clean\n * up all resources that an Observable used. Note that cancelling a subscription will not call `complete` callback\n * provided to `subscribe` function, which is reserved for a regular completion signal that comes from an Observable.\n *\n * Remember that callbacks provided to `subscribe` are not guaranteed to be called asynchronously.\n * It is an Observable itself that decides when these functions will be called. For example {@link of}\n * by default emits all its values synchronously. Always check documentation for how given Observable\n * will behave when subscribed and if its default behavior can be modified with a `scheduler`.\n *\n * #### Examples\n *\n * Subscribe with an {@link guide/observer Observer}\n *\n * ```ts\n * import { of } from 'rxjs';\n *\n * const sumObserver = {\n * sum: 0,\n * next(value) {\n * console.log('Adding: ' + value);\n * this.sum = this.sum + value;\n * },\n * error() {\n * // We actually could just remove this method,\n * // since we do not really care about errors right now.\n * },\n * complete() {\n * console.log('Sum equals: ' + this.sum);\n * }\n * };\n *\n * of(1, 2, 3) // Synchronously emits 1, 2, 3 and then completes.\n * .subscribe(sumObserver);\n *\n * // Logs:\n * // 'Adding: 1'\n * // 'Adding: 2'\n * // 'Adding: 3'\n * // 'Sum equals: 6'\n * ```\n *\n * Subscribe with functions ({@link deprecations/subscribe-arguments deprecated})\n *\n * ```ts\n * import { of } from 'rxjs'\n *\n * let sum = 0;\n *\n * of(1, 2, 3).subscribe(\n * value => {\n * console.log('Adding: ' + value);\n * sum = sum + value;\n * },\n * undefined,\n * () => console.log('Sum equals: ' + sum)\n * );\n *\n * // Logs:\n * // 'Adding: 1'\n * // 'Adding: 2'\n * // 'Adding: 3'\n * // 'Sum equals: 6'\n * ```\n *\n * Cancel a subscription\n *\n * ```ts\n * import { interval } from 'rxjs';\n *\n * const subscription = interval(1000).subscribe({\n * next(num) {\n * console.log(num)\n * },\n * complete() {\n * // Will not be called, even when cancelling subscription.\n * console.log('completed!');\n * }\n * });\n *\n * setTimeout(() => {\n * subscription.unsubscribe();\n * console.log('unsubscribed!');\n * }, 2500);\n *\n * // Logs:\n * // 0 after 1s\n * // 1 after 2s\n * // 'unsubscribed!' after 2.5s\n * ```\n *\n * @param {Observer|Function} observerOrNext (optional) Either an observer with methods to be called,\n * or the first of three possible handlers, which is the handler for each value emitted from the subscribed\n * Observable.\n * @param {Function} error (optional) A handler for a terminal event resulting from an error. If no error handler is provided,\n * the error will be thrown asynchronously as unhandled.\n * @param {Function} complete (optional) A handler for a terminal event resulting from successful completion.\n * @return {Subscription} a subscription reference to the registered handlers\n * @method subscribe\n */\n subscribe(\n observerOrNext?: Partial> | ((value: T) => void) | null,\n error?: ((error: any) => void) | null,\n complete?: (() => void) | null\n ): Subscription {\n const subscriber = isSubscriber(observerOrNext) ? observerOrNext : new SafeSubscriber(observerOrNext, error, complete);\n\n errorContext(() => {\n const { operator, source } = this;\n subscriber.add(\n operator\n ? // We're dealing with a subscription in the\n // operator chain to one of our lifted operators.\n operator.call(subscriber, source)\n : source\n ? // If `source` has a value, but `operator` does not, something that\n // had intimate knowledge of our API, like our `Subject`, must have\n // set it. We're going to just call `_subscribe` directly.\n this._subscribe(subscriber)\n : // In all other cases, we're likely wrapping a user-provided initializer\n // function, so we need to catch errors and handle them appropriately.\n this._trySubscribe(subscriber)\n );\n });\n\n return subscriber;\n }\n\n /** @internal */\n protected _trySubscribe(sink: Subscriber): TeardownLogic {\n try {\n return this._subscribe(sink);\n } catch (err) {\n // We don't need to return anything in this case,\n // because it's just going to try to `add()` to a subscription\n // above.\n sink.error(err);\n }\n }\n\n /**\n * Used as a NON-CANCELLABLE means of subscribing to an observable, for use with\n * APIs that expect promises, like `async/await`. You cannot unsubscribe from this.\n *\n * **WARNING**: Only use this with observables you *know* will complete. If the source\n * observable does not complete, you will end up with a promise that is hung up, and\n * potentially all of the state of an async function hanging out in memory. To avoid\n * this situation, look into adding something like {@link timeout}, {@link take},\n * {@link takeWhile}, or {@link takeUntil} amongst others.\n *\n * #### Example\n *\n * ```ts\n * import { interval, take } from 'rxjs';\n *\n * const source$ = interval(1000).pipe(take(4));\n *\n * async function getTotal() {\n * let total = 0;\n *\n * await source$.forEach(value => {\n * total += value;\n * console.log('observable -> ' + value);\n * });\n *\n * return total;\n * }\n *\n * getTotal().then(\n * total => console.log('Total: ' + total)\n * );\n *\n * // Expected:\n * // 'observable -> 0'\n * // 'observable -> 1'\n * // 'observable -> 2'\n * // 'observable -> 3'\n * // 'Total: 6'\n * ```\n *\n * @param next a handler for each value emitted by the observable\n * @return a promise that either resolves on observable completion or\n * rejects with the handled error\n */\n forEach(next: (value: T) => void): Promise;\n\n /**\n * @param next a handler for each value emitted by the observable\n * @param promiseCtor a constructor function used to instantiate the Promise\n * @return a promise that either resolves on observable completion or\n * rejects with the handled error\n * @deprecated Passing a Promise constructor will no longer be available\n * in upcoming versions of RxJS. This is because it adds weight to the library, for very\n * little benefit. If you need this functionality, it is recommended that you either\n * polyfill Promise, or you create an adapter to convert the returned native promise\n * to whatever promise implementation you wanted. Will be removed in v8.\n */\n forEach(next: (value: T) => void, promiseCtor: PromiseConstructorLike): Promise;\n\n forEach(next: (value: T) => void, promiseCtor?: PromiseConstructorLike): Promise {\n promiseCtor = getPromiseCtor(promiseCtor);\n\n return new promiseCtor((resolve, reject) => {\n const subscriber = new SafeSubscriber({\n next: (value) => {\n try {\n next(value);\n } catch (err) {\n reject(err);\n subscriber.unsubscribe();\n }\n },\n error: reject,\n complete: resolve,\n });\n this.subscribe(subscriber);\n }) as Promise;\n }\n\n /** @internal */\n protected _subscribe(subscriber: Subscriber): TeardownLogic {\n return this.source?.subscribe(subscriber);\n }\n\n /**\n * An interop point defined by the es7-observable spec https://github.com/zenparsing/es-observable\n * @method Symbol.observable\n * @return {Observable} this instance of the observable\n */\n [Symbol_observable]() {\n return this;\n }\n\n /* tslint:disable:max-line-length */\n pipe(): Observable;\n pipe(op1: OperatorFunction): Observable;\n pipe(op1: OperatorFunction, op2: OperatorFunction): Observable;\n pipe(op1: OperatorFunction, op2: OperatorFunction, op3: OperatorFunction): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction\n ): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction,\n op5: OperatorFunction\n ): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction,\n op5: OperatorFunction,\n op6: OperatorFunction\n ): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction,\n op5: OperatorFunction,\n op6: OperatorFunction,\n op7: OperatorFunction\n ): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction,\n op5: OperatorFunction,\n op6: OperatorFunction,\n op7: OperatorFunction,\n op8: OperatorFunction\n ): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction,\n op5: OperatorFunction,\n op6: OperatorFunction,\n op7: OperatorFunction,\n op8: OperatorFunction,\n op9: OperatorFunction\n ): Observable;\n pipe(\n op1: OperatorFunction,\n op2: OperatorFunction,\n op3: OperatorFunction,\n op4: OperatorFunction,\n op5: OperatorFunction,\n op6: OperatorFunction,\n op7: OperatorFunction,\n op8: OperatorFunction,\n op9: OperatorFunction,\n ...operations: OperatorFunction[]\n ): Observable;\n /* tslint:enable:max-line-length */\n\n /**\n * Used to stitch together functional operators into a chain.\n * @method pipe\n * @return {Observable} the Observable result of all of the operators having\n * been called in the order they were passed in.\n *\n * ## Example\n *\n * ```ts\n * import { interval, filter, map, scan } from 'rxjs';\n *\n * interval(1000)\n * .pipe(\n * filter(x => x % 2 === 0),\n * map(x => x + x),\n * scan((acc, x) => acc + x)\n * )\n * .subscribe(x => console.log(x));\n * ```\n */\n pipe(...operations: OperatorFunction[]): Observable {\n return pipeFromArray(operations)(this);\n }\n\n /* tslint:disable:max-line-length */\n /** @deprecated Replaced with {@link firstValueFrom} and {@link lastValueFrom}. Will be removed in v8. Details: https://rxjs.dev/deprecations/to-promise */\n toPromise(): Promise;\n /** @deprecated Replaced with {@link firstValueFrom} and {@link lastValueFrom}. Will be removed in v8. Details: https://rxjs.dev/deprecations/to-promise */\n toPromise(PromiseCtor: typeof Promise): Promise;\n /** @deprecated Replaced with {@link firstValueFrom} and {@link lastValueFrom}. Will be removed in v8. Details: https://rxjs.dev/deprecations/to-promise */\n toPromise(PromiseCtor: PromiseConstructorLike): Promise;\n /* tslint:enable:max-line-length */\n\n /**\n * Subscribe to this Observable and get a Promise resolving on\n * `complete` with the last emission (if any).\n *\n * **WARNING**: Only use this with observables you *know* will complete. If the source\n * observable does not complete, you will end up with a promise that is hung up, and\n * potentially all of the state of an async function hanging out in memory. To avoid\n * this situation, look into adding something like {@link timeout}, {@link take},\n * {@link takeWhile}, or {@link takeUntil} amongst others.\n *\n * @method toPromise\n * @param [promiseCtor] a constructor function used to instantiate\n * the Promise\n * @return A Promise that resolves with the last value emit, or\n * rejects on an error. If there were no emissions, Promise\n * resolves with undefined.\n * @deprecated Replaced with {@link firstValueFrom} and {@link lastValueFrom}. Will be removed in v8. Details: https://rxjs.dev/deprecations/to-promise\n */\n toPromise(promiseCtor?: PromiseConstructorLike): Promise {\n promiseCtor = getPromiseCtor(promiseCtor);\n\n return new promiseCtor((resolve, reject) => {\n let value: T | undefined;\n this.subscribe(\n (x: T) => (value = x),\n (err: any) => reject(err),\n () => resolve(value)\n );\n }) as Promise;\n }\n}\n\n/**\n * Decides between a passed promise constructor from consuming code,\n * A default configured promise constructor, and the native promise\n * constructor and returns it. If nothing can be found, it will throw\n * an error.\n * @param promiseCtor The optional promise constructor to passed by consuming code\n */\nfunction getPromiseCtor(promiseCtor: PromiseConstructorLike | undefined) {\n return promiseCtor ?? config.Promise ?? Promise;\n}\n\nfunction isObserver(value: any): value is Observer {\n return value && isFunction(value.next) && isFunction(value.error) && isFunction(value.complete);\n}\n\nfunction isSubscriber(value: any): value is Subscriber {\n return (value && value instanceof Subscriber) || (isObserver(value) && isSubscription(value));\n}\n", "import { Observable } from '../Observable';\nimport { Subscriber } from '../Subscriber';\nimport { OperatorFunction } from '../types';\nimport { isFunction } from './isFunction';\n\n/**\n * Used to determine if an object is an Observable with a lift function.\n */\nexport function hasLift(source: any): source is { lift: InstanceType['lift'] } {\n return isFunction(source?.lift);\n}\n\n/**\n * Creates an `OperatorFunction`. Used to define operators throughout the library in a concise way.\n * @param init The logic to connect the liftedSource to the subscriber at the moment of subscription.\n */\nexport function operate(\n init: (liftedSource: Observable, subscriber: Subscriber) => (() => void) | void\n): OperatorFunction {\n return (source: Observable) => {\n if (hasLift(source)) {\n return source.lift(function (this: Subscriber, liftedSource: Observable) {\n try {\n return init(liftedSource, this);\n } catch (err) {\n this.error(err);\n }\n });\n }\n throw new TypeError('Unable to lift unknown Observable type');\n };\n}\n", "import { Subscriber } from '../Subscriber';\n\n/**\n * Creates an instance of an `OperatorSubscriber`.\n * @param destination The downstream subscriber.\n * @param onNext Handles next values, only called if this subscriber is not stopped or closed. Any\n * error that occurs in this function is caught and sent to the `error` method of this subscriber.\n * @param onError Handles errors from the subscription, any errors that occur in this handler are caught\n * and send to the `destination` error handler.\n * @param onComplete Handles completion notification from the subscription. Any errors that occur in\n * this handler are sent to the `destination` error handler.\n * @param onFinalize Additional teardown logic here. This will only be called on teardown if the\n * subscriber itself is not already closed. This is called after all other teardown logic is executed.\n */\nexport function createOperatorSubscriber(\n destination: Subscriber,\n onNext?: (value: T) => void,\n onComplete?: () => void,\n onError?: (err: any) => void,\n onFinalize?: () => void\n): Subscriber {\n return new OperatorSubscriber(destination, onNext, onComplete, onError, onFinalize);\n}\n\n/**\n * A generic helper for allowing operators to be created with a Subscriber and\n * use closures to capture necessary state from the operator function itself.\n */\nexport class OperatorSubscriber extends Subscriber {\n /**\n * Creates an instance of an `OperatorSubscriber`.\n * @param destination The downstream subscriber.\n * @param onNext Handles next values, only called if this subscriber is not stopped or closed. Any\n * error that occurs in this function is caught and sent to the `error` method of this subscriber.\n * @param onError Handles errors from the subscription, any errors that occur in this handler are caught\n * and send to the `destination` error handler.\n * @param onComplete Handles completion notification from the subscription. Any errors that occur in\n * this handler are sent to the `destination` error handler.\n * @param onFinalize Additional finalization logic here. This will only be called on finalization if the\n * subscriber itself is not already closed. This is called after all other finalization logic is executed.\n * @param shouldUnsubscribe An optional check to see if an unsubscribe call should truly unsubscribe.\n * NOTE: This currently **ONLY** exists to support the strange behavior of {@link groupBy}, where unsubscription\n * to the resulting observable does not actually disconnect from the source if there are active subscriptions\n * to any grouped observable. (DO NOT EXPOSE OR USE EXTERNALLY!!!)\n */\n constructor(\n destination: Subscriber,\n onNext?: (value: T) => void,\n onComplete?: () => void,\n onError?: (err: any) => void,\n private onFinalize?: () => void,\n private shouldUnsubscribe?: () => boolean\n ) {\n // It's important - for performance reasons - that all of this class's\n // members are initialized and that they are always initialized in the same\n // order. This will ensure that all OperatorSubscriber instances have the\n // same hidden class in V8. This, in turn, will help keep the number of\n // hidden classes involved in property accesses within the base class as\n // low as possible. If the number of hidden classes involved exceeds four,\n // the property accesses will become megamorphic and performance penalties\n // will be incurred - i.e. inline caches won't be used.\n //\n // The reasons for ensuring all instances have the same hidden class are\n // further discussed in this blog post from Benedikt Meurer:\n // https://benediktmeurer.de/2018/03/23/impact-of-polymorphism-on-component-based-frameworks-like-react/\n super(destination);\n this._next = onNext\n ? function (this: OperatorSubscriber, value: T) {\n try {\n onNext(value);\n } catch (err) {\n destination.error(err);\n }\n }\n : super._next;\n this._error = onError\n ? function (this: OperatorSubscriber, err: any) {\n try {\n onError(err);\n } catch (err) {\n // Send any errors that occur down stream.\n destination.error(err);\n } finally {\n // Ensure finalization.\n this.unsubscribe();\n }\n }\n : super._error;\n this._complete = onComplete\n ? function (this: OperatorSubscriber) {\n try {\n onComplete();\n } catch (err) {\n // Send any errors that occur down stream.\n destination.error(err);\n } finally {\n // Ensure finalization.\n this.unsubscribe();\n }\n }\n : super._complete;\n }\n\n unsubscribe() {\n if (!this.shouldUnsubscribe || this.shouldUnsubscribe()) {\n const { closed } = this;\n super.unsubscribe();\n // Execute additional teardown if we have any and we didn't already do so.\n !closed && this.onFinalize?.();\n }\n }\n}\n", "import { Subscription } from '../Subscription';\n\ninterface AnimationFrameProvider {\n schedule(callback: FrameRequestCallback): Subscription;\n requestAnimationFrame: typeof requestAnimationFrame;\n cancelAnimationFrame: typeof cancelAnimationFrame;\n delegate:\n | {\n requestAnimationFrame: typeof requestAnimationFrame;\n cancelAnimationFrame: typeof cancelAnimationFrame;\n }\n | undefined;\n}\n\nexport const animationFrameProvider: AnimationFrameProvider = {\n // When accessing the delegate, use the variable rather than `this` so that\n // the functions can be called without being bound to the provider.\n schedule(callback) {\n let request = requestAnimationFrame;\n let cancel: typeof cancelAnimationFrame | undefined = cancelAnimationFrame;\n const { delegate } = animationFrameProvider;\n if (delegate) {\n request = delegate.requestAnimationFrame;\n cancel = delegate.cancelAnimationFrame;\n }\n const handle = request((timestamp) => {\n // Clear the cancel function. The request has been fulfilled, so\n // attempting to cancel the request upon unsubscription would be\n // pointless.\n cancel = undefined;\n callback(timestamp);\n });\n return new Subscription(() => cancel?.(handle));\n },\n requestAnimationFrame(...args) {\n const { delegate } = animationFrameProvider;\n return (delegate?.requestAnimationFrame || requestAnimationFrame)(...args);\n },\n cancelAnimationFrame(...args) {\n const { delegate } = animationFrameProvider;\n return (delegate?.cancelAnimationFrame || cancelAnimationFrame)(...args);\n },\n delegate: undefined,\n};\n", "import { createErrorClass } from './createErrorClass';\n\nexport interface ObjectUnsubscribedError extends Error {}\n\nexport interface ObjectUnsubscribedErrorCtor {\n /**\n * @deprecated Internal implementation detail. Do not construct error instances.\n * Cannot be tagged as internal: https://github.com/ReactiveX/rxjs/issues/6269\n */\n new (): ObjectUnsubscribedError;\n}\n\n/**\n * An error thrown when an action is invalid because the object has been\n * unsubscribed.\n *\n * @see {@link Subject}\n * @see {@link BehaviorSubject}\n *\n * @class ObjectUnsubscribedError\n */\nexport const ObjectUnsubscribedError: ObjectUnsubscribedErrorCtor = createErrorClass(\n (_super) =>\n function ObjectUnsubscribedErrorImpl(this: any) {\n _super(this);\n this.name = 'ObjectUnsubscribedError';\n this.message = 'object unsubscribed';\n }\n);\n", "import { Operator } from './Operator';\nimport { Observable } from './Observable';\nimport { Subscriber } from './Subscriber';\nimport { Subscription, EMPTY_SUBSCRIPTION } from './Subscription';\nimport { Observer, SubscriptionLike, TeardownLogic } from './types';\nimport { ObjectUnsubscribedError } from './util/ObjectUnsubscribedError';\nimport { arrRemove } from './util/arrRemove';\nimport { errorContext } from './util/errorContext';\n\n/**\n * A Subject is a special type of Observable that allows values to be\n * multicasted to many Observers. Subjects are like EventEmitters.\n *\n * Every Subject is an Observable and an Observer. You can subscribe to a\n * Subject, and you can call next to feed values as well as error and complete.\n */\nexport class Subject extends Observable implements SubscriptionLike {\n closed = false;\n\n private currentObservers: Observer[] | null = null;\n\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n observers: Observer[] = [];\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n isStopped = false;\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n hasError = false;\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n thrownError: any = null;\n\n /**\n * Creates a \"subject\" by basically gluing an observer to an observable.\n *\n * @nocollapse\n * @deprecated Recommended you do not use. Will be removed at some point in the future. Plans for replacement still under discussion.\n */\n static create: (...args: any[]) => any = (destination: Observer, source: Observable): AnonymousSubject => {\n return new AnonymousSubject(destination, source);\n };\n\n constructor() {\n // NOTE: This must be here to obscure Observable's constructor.\n super();\n }\n\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n lift(operator: Operator): Observable {\n const subject = new AnonymousSubject(this, this);\n subject.operator = operator as any;\n return subject as any;\n }\n\n /** @internal */\n protected _throwIfClosed() {\n if (this.closed) {\n throw new ObjectUnsubscribedError();\n }\n }\n\n next(value: T) {\n errorContext(() => {\n this._throwIfClosed();\n if (!this.isStopped) {\n if (!this.currentObservers) {\n this.currentObservers = Array.from(this.observers);\n }\n for (const observer of this.currentObservers) {\n observer.next(value);\n }\n }\n });\n }\n\n error(err: any) {\n errorContext(() => {\n this._throwIfClosed();\n if (!this.isStopped) {\n this.hasError = this.isStopped = true;\n this.thrownError = err;\n const { observers } = this;\n while (observers.length) {\n observers.shift()!.error(err);\n }\n }\n });\n }\n\n complete() {\n errorContext(() => {\n this._throwIfClosed();\n if (!this.isStopped) {\n this.isStopped = true;\n const { observers } = this;\n while (observers.length) {\n observers.shift()!.complete();\n }\n }\n });\n }\n\n unsubscribe() {\n this.isStopped = this.closed = true;\n this.observers = this.currentObservers = null!;\n }\n\n get observed() {\n return this.observers?.length > 0;\n }\n\n /** @internal */\n protected _trySubscribe(subscriber: Subscriber): TeardownLogic {\n this._throwIfClosed();\n return super._trySubscribe(subscriber);\n }\n\n /** @internal */\n protected _subscribe(subscriber: Subscriber): Subscription {\n this._throwIfClosed();\n this._checkFinalizedStatuses(subscriber);\n return this._innerSubscribe(subscriber);\n }\n\n /** @internal */\n protected _innerSubscribe(subscriber: Subscriber) {\n const { hasError, isStopped, observers } = this;\n if (hasError || isStopped) {\n return EMPTY_SUBSCRIPTION;\n }\n this.currentObservers = null;\n observers.push(subscriber);\n return new Subscription(() => {\n this.currentObservers = null;\n arrRemove(observers, subscriber);\n });\n }\n\n /** @internal */\n protected _checkFinalizedStatuses(subscriber: Subscriber) {\n const { hasError, thrownError, isStopped } = this;\n if (hasError) {\n subscriber.error(thrownError);\n } else if (isStopped) {\n subscriber.complete();\n }\n }\n\n /**\n * Creates a new Observable with this Subject as the source. You can do this\n * to create custom Observer-side logic of the Subject and conceal it from\n * code that uses the Observable.\n * @return {Observable} Observable that the Subject casts to\n */\n asObservable(): Observable {\n const observable: any = new Observable();\n observable.source = this;\n return observable;\n }\n}\n\n/**\n * @class AnonymousSubject\n */\nexport class AnonymousSubject extends Subject {\n constructor(\n /** @deprecated Internal implementation detail, do not use directly. Will be made internal in v8. */\n public destination?: Observer,\n source?: Observable\n ) {\n super();\n this.source = source;\n }\n\n next(value: T) {\n this.destination?.next?.(value);\n }\n\n error(err: any) {\n this.destination?.error?.(err);\n }\n\n complete() {\n this.destination?.complete?.();\n }\n\n /** @internal */\n protected _subscribe(subscriber: Subscriber): Subscription {\n return this.source?.subscribe(subscriber) ?? EMPTY_SUBSCRIPTION;\n }\n}\n", "import { Subject } from './Subject';\nimport { Subscriber } from './Subscriber';\nimport { Subscription } from './Subscription';\n\n/**\n * A variant of Subject that requires an initial value and emits its current\n * value whenever it is subscribed to.\n *\n * @class BehaviorSubject\n */\nexport class BehaviorSubject extends Subject {\n constructor(private _value: T) {\n super();\n }\n\n get value(): T {\n return this.getValue();\n }\n\n /** @internal */\n protected _subscribe(subscriber: Subscriber): Subscription {\n const subscription = super._subscribe(subscriber);\n !subscription.closed && subscriber.next(this._value);\n return subscription;\n }\n\n getValue(): T {\n const { hasError, thrownError, _value } = this;\n if (hasError) {\n throw thrownError;\n }\n this._throwIfClosed();\n return _value;\n }\n\n next(value: T): void {\n super.next((this._value = value));\n }\n}\n", "import { TimestampProvider } from '../types';\n\ninterface DateTimestampProvider extends TimestampProvider {\n delegate: TimestampProvider | undefined;\n}\n\nexport const dateTimestampProvider: DateTimestampProvider = {\n now() {\n // Use the variable rather than `this` so that the function can be called\n // without being bound to the provider.\n return (dateTimestampProvider.delegate || Date).now();\n },\n delegate: undefined,\n};\n", "import { Subject } from './Subject';\nimport { TimestampProvider } from './types';\nimport { Subscriber } from './Subscriber';\nimport { Subscription } from './Subscription';\nimport { dateTimestampProvider } from './scheduler/dateTimestampProvider';\n\n/**\n * A variant of {@link Subject} that \"replays\" old values to new subscribers by emitting them when they first subscribe.\n *\n * `ReplaySubject` has an internal buffer that will store a specified number of values that it has observed. Like `Subject`,\n * `ReplaySubject` \"observes\" values by having them passed to its `next` method. When it observes a value, it will store that\n * value for a time determined by the configuration of the `ReplaySubject`, as passed to its constructor.\n *\n * When a new subscriber subscribes to the `ReplaySubject` instance, it will synchronously emit all values in its buffer in\n * a First-In-First-Out (FIFO) manner. The `ReplaySubject` will also complete, if it has observed completion; and it will\n * error if it has observed an error.\n *\n * There are two main configuration items to be concerned with:\n *\n * 1. `bufferSize` - This will determine how many items are stored in the buffer, defaults to infinite.\n * 2. `windowTime` - The amount of time to hold a value in the buffer before removing it from the buffer.\n *\n * Both configurations may exist simultaneously. So if you would like to buffer a maximum of 3 values, as long as the values\n * are less than 2 seconds old, you could do so with a `new ReplaySubject(3, 2000)`.\n *\n * ### Differences with BehaviorSubject\n *\n * `BehaviorSubject` is similar to `new ReplaySubject(1)`, with a couple of exceptions:\n *\n * 1. `BehaviorSubject` comes \"primed\" with a single value upon construction.\n * 2. `ReplaySubject` will replay values, even after observing an error, where `BehaviorSubject` will not.\n *\n * @see {@link Subject}\n * @see {@link BehaviorSubject}\n * @see {@link shareReplay}\n */\nexport class ReplaySubject extends Subject {\n private _buffer: (T | number)[] = [];\n private _infiniteTimeWindow = true;\n\n /**\n * @param bufferSize The size of the buffer to replay on subscription\n * @param windowTime The amount of time the buffered items will stay buffered\n * @param timestampProvider An object with a `now()` method that provides the current timestamp. This is used to\n * calculate the amount of time something has been buffered.\n */\n constructor(\n private _bufferSize = Infinity,\n private _windowTime = Infinity,\n private _timestampProvider: TimestampProvider = dateTimestampProvider\n ) {\n super();\n this._infiniteTimeWindow = _windowTime === Infinity;\n this._bufferSize = Math.max(1, _bufferSize);\n this._windowTime = Math.max(1, _windowTime);\n }\n\n next(value: T): void {\n const { isStopped, _buffer, _infiniteTimeWindow, _timestampProvider, _windowTime } = this;\n if (!isStopped) {\n _buffer.push(value);\n !_infiniteTimeWindow && _buffer.push(_timestampProvider.now() + _windowTime);\n }\n this._trimBuffer();\n super.next(value);\n }\n\n /** @internal */\n protected _subscribe(subscriber: Subscriber): Subscription {\n this._throwIfClosed();\n this._trimBuffer();\n\n const subscription = this._innerSubscribe(subscriber);\n\n const { _infiniteTimeWindow, _buffer } = this;\n // We use a copy here, so reentrant code does not mutate our array while we're\n // emitting it to a new subscriber.\n const copy = _buffer.slice();\n for (let i = 0; i < copy.length && !subscriber.closed; i += _infiniteTimeWindow ? 1 : 2) {\n subscriber.next(copy[i] as T);\n }\n\n this._checkFinalizedStatuses(subscriber);\n\n return subscription;\n }\n\n private _trimBuffer() {\n const { _bufferSize, _timestampProvider, _buffer, _infiniteTimeWindow } = this;\n // If we don't have an infinite buffer size, and we're over the length,\n // use splice to truncate the old buffer values off. Note that we have to\n // double the size for instances where we're not using an infinite time window\n // because we're storing the values and the timestamps in the same array.\n const adjustedBufferSize = (_infiniteTimeWindow ? 1 : 2) * _bufferSize;\n _bufferSize < Infinity && adjustedBufferSize < _buffer.length && _buffer.splice(0, _buffer.length - adjustedBufferSize);\n\n // Now, if we're not in an infinite time window, remove all values where the time is\n // older than what is allowed.\n if (!_infiniteTimeWindow) {\n const now = _timestampProvider.now();\n let last = 0;\n // Search the array for the first timestamp that isn't expired and\n // truncate the buffer up to that point.\n for (let i = 1; i < _buffer.length && (_buffer[i] as number) <= now; i += 2) {\n last = i;\n }\n last && _buffer.splice(0, last + 1);\n }\n }\n}\n", "import { Scheduler } from '../Scheduler';\nimport { Subscription } from '../Subscription';\nimport { SchedulerAction } from '../types';\n\n/**\n * A unit of work to be executed in a `scheduler`. An action is typically\n * created from within a {@link SchedulerLike} and an RxJS user does not need to concern\n * themselves about creating and manipulating an Action.\n *\n * ```ts\n * class Action extends Subscription {\n * new (scheduler: Scheduler, work: (state?: T) => void);\n * schedule(state?: T, delay: number = 0): Subscription;\n * }\n * ```\n *\n * @class Action\n */\nexport class Action extends Subscription {\n constructor(scheduler: Scheduler, work: (this: SchedulerAction, state?: T) => void) {\n super();\n }\n /**\n * Schedules this action on its parent {@link SchedulerLike} for execution. May be passed\n * some context object, `state`. May happen at some point in the future,\n * according to the `delay` parameter, if specified.\n * @param {T} [state] Some contextual data that the `work` function uses when\n * called by the Scheduler.\n * @param {number} [delay] Time to wait before executing the work, where the\n * time unit is implicit and defined by the Scheduler.\n * @return {void}\n */\n public schedule(state?: T, delay: number = 0): Subscription {\n return this;\n }\n}\n", "import type { TimerHandle } from './timerHandle';\ntype SetIntervalFunction = (handler: () => void, timeout?: number, ...args: any[]) => TimerHandle;\ntype ClearIntervalFunction = (handle: TimerHandle) => void;\n\ninterface IntervalProvider {\n setInterval: SetIntervalFunction;\n clearInterval: ClearIntervalFunction;\n delegate:\n | {\n setInterval: SetIntervalFunction;\n clearInterval: ClearIntervalFunction;\n }\n | undefined;\n}\n\nexport const intervalProvider: IntervalProvider = {\n // When accessing the delegate, use the variable rather than `this` so that\n // the functions can be called without being bound to the provider.\n setInterval(handler: () => void, timeout?: number, ...args) {\n const { delegate } = intervalProvider;\n if (delegate?.setInterval) {\n return delegate.setInterval(handler, timeout, ...args);\n }\n return setInterval(handler, timeout, ...args);\n },\n clearInterval(handle) {\n const { delegate } = intervalProvider;\n return (delegate?.clearInterval || clearInterval)(handle as any);\n },\n delegate: undefined,\n};\n", "import { Action } from './Action';\nimport { SchedulerAction } from '../types';\nimport { Subscription } from '../Subscription';\nimport { AsyncScheduler } from './AsyncScheduler';\nimport { intervalProvider } from './intervalProvider';\nimport { arrRemove } from '../util/arrRemove';\nimport { TimerHandle } from './timerHandle';\n\nexport class AsyncAction extends Action {\n public id: TimerHandle | undefined;\n public state?: T;\n // @ts-ignore: Property has no initializer and is not definitely assigned\n public delay: number;\n protected pending: boolean = false;\n\n constructor(protected scheduler: AsyncScheduler, protected work: (this: SchedulerAction, state?: T) => void) {\n super(scheduler, work);\n }\n\n public schedule(state?: T, delay: number = 0): Subscription {\n if (this.closed) {\n return this;\n }\n\n // Always replace the current state with the new state.\n this.state = state;\n\n const id = this.id;\n const scheduler = this.scheduler;\n\n //\n // Important implementation note:\n //\n // Actions only execute once by default, unless rescheduled from within the\n // scheduled callback. This allows us to implement single and repeat\n // actions via the same code path, without adding API surface area, as well\n // as mimic traditional recursion but across asynchronous boundaries.\n //\n // However, JS runtimes and timers distinguish between intervals achieved by\n // serial `setTimeout` calls vs. a single `setInterval` call. An interval of\n // serial `setTimeout` calls can be individually delayed, which delays\n // scheduling the next `setTimeout`, and so on. `setInterval` attempts to\n // guarantee the interval callback will be invoked more precisely to the\n // interval period, regardless of load.\n //\n // Therefore, we use `setInterval` to schedule single and repeat actions.\n // If the action reschedules itself with the same delay, the interval is not\n // canceled. If the action doesn't reschedule, or reschedules with a\n // different delay, the interval will be canceled after scheduled callback\n // execution.\n //\n if (id != null) {\n this.id = this.recycleAsyncId(scheduler, id, delay);\n }\n\n // Set the pending flag indicating that this action has been scheduled, or\n // has recursively rescheduled itself.\n this.pending = true;\n\n this.delay = delay;\n // If this action has already an async Id, don't request a new one.\n this.id = this.id ?? this.requestAsyncId(scheduler, this.id, delay);\n\n return this;\n }\n\n protected requestAsyncId(scheduler: AsyncScheduler, _id?: TimerHandle, delay: number = 0): TimerHandle {\n return intervalProvider.setInterval(scheduler.flush.bind(scheduler, this), delay);\n }\n\n protected recycleAsyncId(_scheduler: AsyncScheduler, id?: TimerHandle, delay: number | null = 0): TimerHandle | undefined {\n // If this action is rescheduled with the same delay time, don't clear the interval id.\n if (delay != null && this.delay === delay && this.pending === false) {\n return id;\n }\n // Otherwise, if the action's delay time is different from the current delay,\n // or the action has been rescheduled before it's executed, clear the interval id\n if (id != null) {\n intervalProvider.clearInterval(id);\n }\n\n return undefined;\n }\n\n /**\n * Immediately executes this action and the `work` it contains.\n * @return {any}\n */\n public execute(state: T, delay: number): any {\n if (this.closed) {\n return new Error('executing a cancelled action');\n }\n\n this.pending = false;\n const error = this._execute(state, delay);\n if (error) {\n return error;\n } else if (this.pending === false && this.id != null) {\n // Dequeue if the action didn't reschedule itself. Don't call\n // unsubscribe(), because the action could reschedule later.\n // For example:\n // ```\n // scheduler.schedule(function doWork(counter) {\n // /* ... I'm a busy worker bee ... */\n // var originalAction = this;\n // /* wait 100ms before rescheduling the action */\n // setTimeout(function () {\n // originalAction.schedule(counter + 1);\n // }, 100);\n // }, 1000);\n // ```\n this.id = this.recycleAsyncId(this.scheduler, this.id, null);\n }\n }\n\n protected _execute(state: T, _delay: number): any {\n let errored: boolean = false;\n let errorValue: any;\n try {\n this.work(state);\n } catch (e) {\n errored = true;\n // HACK: Since code elsewhere is relying on the \"truthiness\" of the\n // return here, we can't have it return \"\" or 0 or false.\n // TODO: Clean this up when we refactor schedulers mid-version-8 or so.\n errorValue = e ? e : new Error('Scheduled action threw falsy error');\n }\n if (errored) {\n this.unsubscribe();\n return errorValue;\n }\n }\n\n unsubscribe() {\n if (!this.closed) {\n const { id, scheduler } = this;\n const { actions } = scheduler;\n\n this.work = this.state = this.scheduler = null!;\n this.pending = false;\n\n arrRemove(actions, this);\n if (id != null) {\n this.id = this.recycleAsyncId(scheduler, id, null);\n }\n\n this.delay = null!;\n super.unsubscribe();\n }\n }\n}\n", "import { Action } from './scheduler/Action';\nimport { Subscription } from './Subscription';\nimport { SchedulerLike, SchedulerAction } from './types';\nimport { dateTimestampProvider } from './scheduler/dateTimestampProvider';\n\n/**\n * An execution context and a data structure to order tasks and schedule their\n * execution. Provides a notion of (potentially virtual) time, through the\n * `now()` getter method.\n *\n * Each unit of work in a Scheduler is called an `Action`.\n *\n * ```ts\n * class Scheduler {\n * now(): number;\n * schedule(work, delay?, state?): Subscription;\n * }\n * ```\n *\n * @class Scheduler\n * @deprecated Scheduler is an internal implementation detail of RxJS, and\n * should not be used directly. Rather, create your own class and implement\n * {@link SchedulerLike}. Will be made internal in v8.\n */\nexport class Scheduler implements SchedulerLike {\n public static now: () => number = dateTimestampProvider.now;\n\n constructor(private schedulerActionCtor: typeof Action, now: () => number = Scheduler.now) {\n this.now = now;\n }\n\n /**\n * A getter method that returns a number representing the current time\n * (at the time this function was called) according to the scheduler's own\n * internal clock.\n * @return {number} A number that represents the current time. May or may not\n * have a relation to wall-clock time. May or may not refer to a time unit\n * (e.g. milliseconds).\n */\n public now: () => number;\n\n /**\n * Schedules a function, `work`, for execution. May happen at some point in\n * the future, according to the `delay` parameter, if specified. May be passed\n * some context object, `state`, which will be passed to the `work` function.\n *\n * The given arguments will be processed an stored as an Action object in a\n * queue of actions.\n *\n * @param {function(state: ?T): ?Subscription} work A function representing a\n * task, or some unit of work to be executed by the Scheduler.\n * @param {number} [delay] Time to wait before executing the work, where the\n * time unit is implicit and defined by the Scheduler itself.\n * @param {T} [state] Some contextual data that the `work` function uses when\n * called by the Scheduler.\n * @return {Subscription} A subscription in order to be able to unsubscribe\n * the scheduled work.\n */\n public schedule(work: (this: SchedulerAction, state?: T) => void, delay: number = 0, state?: T): Subscription {\n return new this.schedulerActionCtor(this, work).schedule(state, delay);\n }\n}\n", "import { Scheduler } from '../Scheduler';\nimport { Action } from './Action';\nimport { AsyncAction } from './AsyncAction';\nimport { TimerHandle } from './timerHandle';\n\nexport class AsyncScheduler extends Scheduler {\n public actions: Array> = [];\n /**\n * A flag to indicate whether the Scheduler is currently executing a batch of\n * queued actions.\n * @type {boolean}\n * @internal\n */\n public _active: boolean = false;\n /**\n * An internal ID used to track the latest asynchronous task such as those\n * coming from `setTimeout`, `setInterval`, `requestAnimationFrame`, and\n * others.\n * @type {any}\n * @internal\n */\n public _scheduled: TimerHandle | undefined;\n\n constructor(SchedulerAction: typeof Action, now: () => number = Scheduler.now) {\n super(SchedulerAction, now);\n }\n\n public flush(action: AsyncAction): void {\n const { actions } = this;\n\n if (this._active) {\n actions.push(action);\n return;\n }\n\n let error: any;\n this._active = true;\n\n do {\n if ((error = action.execute(action.state, action.delay))) {\n break;\n }\n } while ((action = actions.shift()!)); // exhaust the scheduler queue\n\n this._active = false;\n\n if (error) {\n while ((action = actions.shift()!)) {\n action.unsubscribe();\n }\n throw error;\n }\n }\n}\n", "import { AsyncAction } from './AsyncAction';\nimport { AsyncScheduler } from './AsyncScheduler';\n\n/**\n *\n * Async Scheduler\n *\n * Schedule task as if you used setTimeout(task, duration)\n *\n * `async` scheduler schedules tasks asynchronously, by putting them on the JavaScript\n * event loop queue. It is best used to delay tasks in time or to schedule tasks repeating\n * in intervals.\n *\n * If you just want to \"defer\" task, that is to perform it right after currently\n * executing synchronous code ends (commonly achieved by `setTimeout(deferredTask, 0)`),\n * better choice will be the {@link asapScheduler} scheduler.\n *\n * ## Examples\n * Use async scheduler to delay task\n * ```ts\n * import { asyncScheduler } from 'rxjs';\n *\n * const task = () => console.log('it works!');\n *\n * asyncScheduler.schedule(task, 2000);\n *\n * // After 2 seconds logs:\n * // \"it works!\"\n * ```\n *\n * Use async scheduler to repeat task in intervals\n * ```ts\n * import { asyncScheduler } from 'rxjs';\n *\n * function task(state) {\n * console.log(state);\n * this.schedule(state + 1, 1000); // `this` references currently executing Action,\n * // which we reschedule with new state and delay\n * }\n *\n * asyncScheduler.schedule(task, 3000, 0);\n *\n * // Logs:\n * // 0 after 3s\n * // 1 after 4s\n * // 2 after 5s\n * // 3 after 6s\n * ```\n */\n\nexport const asyncScheduler = new AsyncScheduler(AsyncAction);\n\n/**\n * @deprecated Renamed to {@link asyncScheduler}. Will be removed in v8.\n */\nexport const async = asyncScheduler;\n", "import { AsyncAction } from './AsyncAction';\nimport { Subscription } from '../Subscription';\nimport { QueueScheduler } from './QueueScheduler';\nimport { SchedulerAction } from '../types';\nimport { TimerHandle } from './timerHandle';\n\nexport class QueueAction extends AsyncAction {\n constructor(protected scheduler: QueueScheduler, protected work: (this: SchedulerAction, state?: T) => void) {\n super(scheduler, work);\n }\n\n public schedule(state?: T, delay: number = 0): Subscription {\n if (delay > 0) {\n return super.schedule(state, delay);\n }\n this.delay = delay;\n this.state = state;\n this.scheduler.flush(this);\n return this;\n }\n\n public execute(state: T, delay: number): any {\n return delay > 0 || this.closed ? super.execute(state, delay) : this._execute(state, delay);\n }\n\n protected requestAsyncId(scheduler: QueueScheduler, id?: TimerHandle, delay: number = 0): TimerHandle {\n // If delay exists and is greater than 0, or if the delay is null (the\n // action wasn't rescheduled) but was originally scheduled as an async\n // action, then recycle as an async action.\n\n if ((delay != null && delay > 0) || (delay == null && this.delay > 0)) {\n return super.requestAsyncId(scheduler, id, delay);\n }\n\n // Otherwise flush the scheduler starting with this action.\n scheduler.flush(this);\n\n // HACK: In the past, this was returning `void`. However, `void` isn't a valid\n // `TimerHandle`, and generally the return value here isn't really used. So the\n // compromise is to return `0` which is both \"falsy\" and a valid `TimerHandle`,\n // as opposed to refactoring every other instanceo of `requestAsyncId`.\n return 0;\n }\n}\n", "import { AsyncScheduler } from './AsyncScheduler';\n\nexport class QueueScheduler extends AsyncScheduler {\n}\n", "import { QueueAction } from './QueueAction';\nimport { QueueScheduler } from './QueueScheduler';\n\n/**\n *\n * Queue Scheduler\n *\n * Put every next task on a queue, instead of executing it immediately\n *\n * `queue` scheduler, when used with delay, behaves the same as {@link asyncScheduler} scheduler.\n *\n * When used without delay, it schedules given task synchronously - executes it right when\n * it is scheduled. However when called recursively, that is when inside the scheduled task,\n * another task is scheduled with queue scheduler, instead of executing immediately as well,\n * that task will be put on a queue and wait for current one to finish.\n *\n * This means that when you execute task with `queue` scheduler, you are sure it will end\n * before any other task scheduled with that scheduler will start.\n *\n * ## Examples\n * Schedule recursively first, then do something\n * ```ts\n * import { queueScheduler } from 'rxjs';\n *\n * queueScheduler.schedule(() => {\n * queueScheduler.schedule(() => console.log('second')); // will not happen now, but will be put on a queue\n *\n * console.log('first');\n * });\n *\n * // Logs:\n * // \"first\"\n * // \"second\"\n * ```\n *\n * Reschedule itself recursively\n * ```ts\n * import { queueScheduler } from 'rxjs';\n *\n * queueScheduler.schedule(function(state) {\n * if (state !== 0) {\n * console.log('before', state);\n * this.schedule(state - 1); // `this` references currently executing Action,\n * // which we reschedule with new state\n * console.log('after', state);\n * }\n * }, 0, 3);\n *\n * // In scheduler that runs recursively, you would expect:\n * // \"before\", 3\n * // \"before\", 2\n * // \"before\", 1\n * // \"after\", 1\n * // \"after\", 2\n * // \"after\", 3\n *\n * // But with queue it logs:\n * // \"before\", 3\n * // \"after\", 3\n * // \"before\", 2\n * // \"after\", 2\n * // \"before\", 1\n * // \"after\", 1\n * ```\n */\n\nexport const queueScheduler = new QueueScheduler(QueueAction);\n\n/**\n * @deprecated Renamed to {@link queueScheduler}. Will be removed in v8.\n */\nexport const queue = queueScheduler;\n", "import { AsyncAction } from './AsyncAction';\nimport { AnimationFrameScheduler } from './AnimationFrameScheduler';\nimport { SchedulerAction } from '../types';\nimport { animationFrameProvider } from './animationFrameProvider';\nimport { TimerHandle } from './timerHandle';\n\nexport class AnimationFrameAction extends AsyncAction {\n constructor(protected scheduler: AnimationFrameScheduler, protected work: (this: SchedulerAction, state?: T) => void) {\n super(scheduler, work);\n }\n\n protected requestAsyncId(scheduler: AnimationFrameScheduler, id?: TimerHandle, delay: number = 0): TimerHandle {\n // If delay is greater than 0, request as an async action.\n if (delay !== null && delay > 0) {\n return super.requestAsyncId(scheduler, id, delay);\n }\n // Push the action to the end of the scheduler queue.\n scheduler.actions.push(this);\n // If an animation frame has already been requested, don't request another\n // one. If an animation frame hasn't been requested yet, request one. Return\n // the current animation frame request id.\n return scheduler._scheduled || (scheduler._scheduled = animationFrameProvider.requestAnimationFrame(() => scheduler.flush(undefined)));\n }\n\n protected recycleAsyncId(scheduler: AnimationFrameScheduler, id?: TimerHandle, delay: number = 0): TimerHandle | undefined {\n // If delay exists and is greater than 0, or if the delay is null (the\n // action wasn't rescheduled) but was originally scheduled as an async\n // action, then recycle as an async action.\n if (delay != null ? delay > 0 : this.delay > 0) {\n return super.recycleAsyncId(scheduler, id, delay);\n }\n // If the scheduler queue has no remaining actions with the same async id,\n // cancel the requested animation frame and set the scheduled flag to\n // undefined so the next AnimationFrameAction will request its own.\n const { actions } = scheduler;\n if (id != null && actions[actions.length - 1]?.id !== id) {\n animationFrameProvider.cancelAnimationFrame(id as number);\n scheduler._scheduled = undefined;\n }\n // Return undefined so the action knows to request a new async id if it's rescheduled.\n return undefined;\n }\n}\n", "import { AsyncAction } from './AsyncAction';\nimport { AsyncScheduler } from './AsyncScheduler';\n\nexport class AnimationFrameScheduler extends AsyncScheduler {\n public flush(action?: AsyncAction): void {\n this._active = true;\n // The async id that effects a call to flush is stored in _scheduled.\n // Before executing an action, it's necessary to check the action's async\n // id to determine whether it's supposed to be executed in the current\n // flush.\n // Previous implementations of this method used a count to determine this,\n // but that was unsound, as actions that are unsubscribed - i.e. cancelled -\n // are removed from the actions array and that can shift actions that are\n // scheduled to be executed in a subsequent flush into positions at which\n // they are executed within the current flush.\n const flushId = this._scheduled;\n this._scheduled = undefined;\n\n const { actions } = this;\n let error: any;\n action = action || actions.shift()!;\n\n do {\n if ((error = action.execute(action.state, action.delay))) {\n break;\n }\n } while ((action = actions[0]) && action.id === flushId && actions.shift());\n\n this._active = false;\n\n if (error) {\n while ((action = actions[0]) && action.id === flushId && actions.shift()) {\n action.unsubscribe();\n }\n throw error;\n }\n }\n}\n", "import { AnimationFrameAction } from './AnimationFrameAction';\nimport { AnimationFrameScheduler } from './AnimationFrameScheduler';\n\n/**\n *\n * Animation Frame Scheduler\n *\n * Perform task when `window.requestAnimationFrame` would fire\n *\n * When `animationFrame` scheduler is used with delay, it will fall back to {@link asyncScheduler} scheduler\n * behaviour.\n *\n * Without delay, `animationFrame` scheduler can be used to create smooth browser animations.\n * It makes sure scheduled task will happen just before next browser content repaint,\n * thus performing animations as efficiently as possible.\n *\n * ## Example\n * Schedule div height animation\n * ```ts\n * // html:
\n * import { animationFrameScheduler } from 'rxjs';\n *\n * const div = document.querySelector('div');\n *\n * animationFrameScheduler.schedule(function(height) {\n * div.style.height = height + \"px\";\n *\n * this.schedule(height + 1); // `this` references currently executing Action,\n * // which we reschedule with new state\n * }, 0, 0);\n *\n * // You will see a div element growing in height\n * ```\n */\n\nexport const animationFrameScheduler = new AnimationFrameScheduler(AnimationFrameAction);\n\n/**\n * @deprecated Renamed to {@link animationFrameScheduler}. Will be removed in v8.\n */\nexport const animationFrame = animationFrameScheduler;\n", "import { Observable } from '../Observable';\nimport { SchedulerLike } from '../types';\n\n/**\n * A simple Observable that emits no items to the Observer and immediately\n * emits a complete notification.\n *\n * Just emits 'complete', and nothing else.\n *\n * ![](empty.png)\n *\n * A simple Observable that only emits the complete notification. It can be used\n * for composing with other Observables, such as in a {@link mergeMap}.\n *\n * ## Examples\n *\n * Log complete notification\n *\n * ```ts\n * import { EMPTY } from 'rxjs';\n *\n * EMPTY.subscribe({\n * next: () => console.log('Next'),\n * complete: () => console.log('Complete!')\n * });\n *\n * // Outputs\n * // Complete!\n * ```\n *\n * Emit the number 7, then complete\n *\n * ```ts\n * import { EMPTY, startWith } from 'rxjs';\n *\n * const result = EMPTY.pipe(startWith(7));\n * result.subscribe(x => console.log(x));\n *\n * // Outputs\n * // 7\n * ```\n *\n * Map and flatten only odd numbers to the sequence `'a'`, `'b'`, `'c'`\n *\n * ```ts\n * import { interval, mergeMap, of, EMPTY } from 'rxjs';\n *\n * const interval$ = interval(1000);\n * const result = interval$.pipe(\n * mergeMap(x => x % 2 === 1 ? of('a', 'b', 'c') : EMPTY),\n * );\n * result.subscribe(x => console.log(x));\n *\n * // Results in the following to the console:\n * // x is equal to the count on the interval, e.g. (0, 1, 2, 3, ...)\n * // x will occur every 1000ms\n * // if x % 2 is equal to 1, print a, b, c (each on its own)\n * // if x % 2 is not equal to 1, nothing will be output\n * ```\n *\n * @see {@link Observable}\n * @see {@link NEVER}\n * @see {@link of}\n * @see {@link throwError}\n */\nexport const EMPTY = new Observable((subscriber) => subscriber.complete());\n\n/**\n * @param scheduler A {@link SchedulerLike} to use for scheduling\n * the emission of the complete notification.\n * @deprecated Replaced with the {@link EMPTY} constant or {@link scheduled} (e.g. `scheduled([], scheduler)`). Will be removed in v8.\n */\nexport function empty(scheduler?: SchedulerLike) {\n return scheduler ? emptyScheduled(scheduler) : EMPTY;\n}\n\nfunction emptyScheduled(scheduler: SchedulerLike) {\n return new Observable((subscriber) => scheduler.schedule(() => subscriber.complete()));\n}\n", "import { SchedulerLike } from '../types';\nimport { isFunction } from './isFunction';\n\nexport function isScheduler(value: any): value is SchedulerLike {\n return value && isFunction(value.schedule);\n}\n", "import { SchedulerLike } from '../types';\nimport { isFunction } from './isFunction';\nimport { isScheduler } from './isScheduler';\n\nfunction last(arr: T[]): T | undefined {\n return arr[arr.length - 1];\n}\n\nexport function popResultSelector(args: any[]): ((...args: unknown[]) => unknown) | undefined {\n return isFunction(last(args)) ? args.pop() : undefined;\n}\n\nexport function popScheduler(args: any[]): SchedulerLike | undefined {\n return isScheduler(last(args)) ? args.pop() : undefined;\n}\n\nexport function popNumber(args: any[], defaultValue: number): number {\n return typeof last(args) === 'number' ? args.pop()! : defaultValue;\n}\n", "export const isArrayLike = ((x: any): x is ArrayLike => x && typeof x.length === 'number' && typeof x !== 'function');", "import { isFunction } from \"./isFunction\";\n\n/**\n * Tests to see if the object is \"thennable\".\n * @param value the object to test\n */\nexport function isPromise(value: any): value is PromiseLike {\n return isFunction(value?.then);\n}\n", "import { InteropObservable } from '../types';\nimport { observable as Symbol_observable } from '../symbol/observable';\nimport { isFunction } from './isFunction';\n\n/** Identifies an input as being Observable (but not necessary an Rx Observable) */\nexport function isInteropObservable(input: any): input is InteropObservable {\n return isFunction(input[Symbol_observable]);\n}\n", "import { isFunction } from './isFunction';\n\nexport function isAsyncIterable(obj: any): obj is AsyncIterable {\n return Symbol.asyncIterator && isFunction(obj?.[Symbol.asyncIterator]);\n}\n", "/**\n * Creates the TypeError to throw if an invalid object is passed to `from` or `scheduled`.\n * @param input The object that was passed.\n */\nexport function createInvalidObservableTypeError(input: any) {\n // TODO: We should create error codes that can be looked up, so this can be less verbose.\n return new TypeError(\n `You provided ${\n input !== null && typeof input === 'object' ? 'an invalid object' : `'${input}'`\n } where a stream was expected. You can provide an Observable, Promise, ReadableStream, Array, AsyncIterable, or Iterable.`\n );\n}\n", "export function getSymbolIterator(): symbol {\n if (typeof Symbol !== 'function' || !Symbol.iterator) {\n return '@@iterator' as any;\n }\n\n return Symbol.iterator;\n}\n\nexport const iterator = getSymbolIterator();\n", "import { iterator as Symbol_iterator } from '../symbol/iterator';\nimport { isFunction } from './isFunction';\n\n/** Identifies an input as being an Iterable */\nexport function isIterable(input: any): input is Iterable {\n return isFunction(input?.[Symbol_iterator]);\n}\n", "import { ReadableStreamLike } from '../types';\nimport { isFunction } from './isFunction';\n\nexport async function* readableStreamLikeToAsyncGenerator(readableStream: ReadableStreamLike): AsyncGenerator {\n const reader = readableStream.getReader();\n try {\n while (true) {\n const { value, done } = await reader.read();\n if (done) {\n return;\n }\n yield value!;\n }\n } finally {\n reader.releaseLock();\n }\n}\n\nexport function isReadableStreamLike(obj: any): obj is ReadableStreamLike {\n // We don't want to use instanceof checks because they would return\n // false for instances from another Realm, like an +

+---> + +

The coordinate system +The coordinate system

+

Code style

+

Contributions of code should follow standardised or community-agreed styles and be provided in (or added to) a structure suitable for packaging and uploading to package libraries. For Python this includes pip and/or conda, for R this would be CRAN, for Matlab this would be a toolbox on the MATLAB File Exchange, etc.

+ +
+
+
    +
  1. +

    ISO. 2019. ISO 80000-2. Part 2: Mathematics. ↩

    +
    2. +
+
+ + + + + + + + + + + + + + + + + + + + + + + + + + + +
+
+
+ + + + + + + + + + \ No newline at end of file diff --git a/developing/index.html b/developing/index.html new file mode 100644 index 0000000..3830bc1 --- /dev/null +++ b/developing/index.html @@ -0,0 +1,604 @@ + + + + + + + + + + + + + + + + + + + + + + + + + Developing echoSMs - echoSMs + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ +
+ + + + + + +
+ + +
+ +
+ + + + + + +
+
+ + + +
+
+
+ + + + + + + +
+
+
+ + + + +
+
+ + + + + + + + + + + + + + + +

Developing echoSMs

+
+

These notes are a work in progress.

+
+

This page contains notes and instructions on developing and adding new models to echoSMs.

+

Obtaining the source code

+

The echoSMs source code is kept on github under an ICES account. Clone the repository with this URL:

+
https://github.com/ices-tools-dev/echoSMs.git
+
+

Generating packages for PyPI

+

EchoSMs is a pure Python package. The build configuration is done via a pyproject.toml file and hatchling is used to produce packages.

+

A github action in the echoSMS repository will generate a Python wheel and source package and upload these to PyPI. This action is triggered whenever a tagged commit occurs to the repository. The tag is used as the new version number. EchoSMs version numbers follow the semantic versioning convention.

+

Every commit to the echoSMs repository will generate a development package being uploaded to TestPyPI. This is used to always check that a commit does not prevent production of a package and is where a package containing the latest commit can be obtained.

+

Documentation

+

The echoSMs documentation is produced using mkdocs and mkdocstrings. The documentation pages are hosted by github and are regenerated after every commit to the repository using a github action.

+

Documentation edits can be tested locally by running:

+
mkdocs serve
+
+

in the top level of the echoSMs repository. The documentation is then available at http://127.0.0.1:8000.

+

Tests

+

EchoSMs uses the pytest testing framework. After installing pytest, run the tests using

+
pytest -v
+
+

in the top level of the echoSMs repository.

+

Adding a new scattering model

+

TBD.

+ + + + + + + + + + + + + +
+
+ + + +
+ + + +
+ + + +
+
+
+
+ + + + + + + + + + \ No newline at end of file diff --git a/index.html b/index.html new file mode 100644 index 0000000..5661c96 --- /dev/null +++ b/index.html @@ -0,0 +1,917 @@ + + + + + + + + + + + + + + + + + + + + + + + echoSMs + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ +
+ + + + + + +
+ + +
+ +
+ + + + + + +
+
+ + + +
+
+
+ + + + + + + +
+
+
+ + + + +
+
+ + + + + + + + + + + + + + + +

echoSMs

+
+

Making acoustic scattering models available to fisheries and plankton scientists via the world wide web.

+
+

Background

+

This project is an international collaboration that is, in part, a component of a U.S. NOAA-Fisheries active acoustic strategic initiative, AA-SI.

+

Quantitative interpretation of acoustic echograms requires software expertise to develop advanced analytical methods for echo classification using mathematical models that predict acoustic backscatter (e.g., target strength, TS re 1 m² [dB]). These models and predictions can be used to inform echo classification by validating empirical measurements and generating training data for machine learning (ML), artificial intelligence (AI), and other advanced analytical methods, such as inverse methods. Application of these models to fish and plankton requires anatomical and morphological data that are easily accessible and available to the models.

+

The goal of this project is to make acoustic scattering models available to fisheries and plankton acoustic scientists via the world wide web. By providing the models in an open-access and open-source software language (e.g, Python, R) and providing morphological and anatomical data in open data formats (e.g., HDF5, relational databases), the proper and appropriate use of these models can extend to the entire fisheries and plankton acoustics’ community.

+

Contributing to echoSMs

+

We welcome all contributions to echoSMs, be it code, test cases, bug reports, discussion of models, etc. Please use the github facilities for this (i.e., issues, pull requests, and discussions). We are also happy to accept directly code that we can add to echoSMs on your behalf.

+

An objective of echoSMs is to provide scattering models in a form that is easy to access, use, and compare to other models. To help with that, we specify model parameter units, angle conventions, and required model outputs that code contributions should support. We also suggest coding conventions that should be followed.

+

Scattering Models

+

The initial set of acoustic scattering models will be those used in Jech et al. (2015). Acoustic model development will follow 3–4 phases:

+
    +
  1. Exact solutions and canonical shapes (see table below),
    2. +
  2. Approximate analytical models applied to canonical shapes (see table below),
    3. +
  3. Approximate analytical models applied to complex shapes approximating biological targets, such as fish and zooplankton,
    4. +
  4. Numerical models applied to canonical shapes and biological targets (this phase will depend on time and funding).
    5. +
+

Exact Solutions

+

The exact solutions, shapes, and supported boundary conditions will be:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelShapeDescriptionReferences
Modal Series solution (MSS)SphereFluid filled1,2
Fixed rigid2,3
Pressure release2
Gas filled2
Weakly scattering2
Spherical fluid shell with fluid interior2
Fixed rigid spherical shell2
Spherical fluid shell with pressure release interior2
Spherical fluid shell with gas interior2
Spherical fluid shell with weakly scattering interior2
Prolate spheroid modal series solutionProlate spheroidRigid fixed2,4,5
Pressure release2,4,5
Gas filled2,4,5
Infinite cylinder?3
Infinite plane?
+
    +
  1. Anderson, V. C. 1950. Sound scattering from a fluid sphere. JASA. 22(4): 426-431]
    2. +
  2. Jech et al. 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. JASA. 138: 3742-3764.
    3. +
  3. Faran, J. J. 1951. Sound scattering by solid cylinders and spheres. JASA. 23(4): 405-418.
    4. +
  4. Skudrzyk. 1971. The Foundations of Acoustics (Springer, NY), pp. 455-465.
    5. +
  5. Furusawa. 1988. Prolate spheroidal models for predicting general trends of fish target strength, J. Acoust. Soc. Jpn. (E) 9, 13–14.
    6. +
+

Approximate Analytical Models and Shapes

+

The approximate analytical models,shapes, and supported boundary conditions will be:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelShapeDescriptionReferences
Modal series-based deformed cylinder modelFinite cylinderFixed rigid1,2,3
Pressure release1,2,3
Gas filled1,2,3
Weakly scattering1,2,3
Prolate spheroidFixed rigid1,2,3
Pressure release1,2,3
Gas filled1,2,3
Weakly scattering1,2,3
Kirchhoff approximation (KA)SphereFixed rigid3,4,5
Prolate spheroidFixed rigid3,4,5
Finite cylinderFixed rigid3,4,5
Kirchhoff ray mode (KRM)SphereGas filled3,6,7,8
Weakly scattering3,6,7,8
Spherical shellgas filled3,6,7,8
Weakly scattering3,6,7,8
Prolate spheroidgas filled3,6,7,8
Weakly scattering3,6,7,8
Finite cylindergas filled3,6,7,8
Weakly scattering3,6,7,8
Distorted wave Born approximation (DWBA)SphereWeakly scattering3,9,10,11
Prolate spheroidWeakly scattering3,9,10,11
Finite cylinderWeakly scattering3,9,10,11
Phase-tracking distorted wave Born approximation (PT-DWBA)SphereWeakly scattering3,12
Spherical shellWeakly scattering3,12
Prolate spheroidWeakly scattering3,12
Finite cylinderWeakly scattering3,12
Stochastic distorted wave Born approximation (SDWBA)SphereWeakly scattering13,14,15
Prolate spheroidWeakly scattering13,14,15
Finite cylinderWeakly scattering13,14,15
+
    +
  1. Stanton. 1988. Sound scattering by cylinders of finite length. I. Fluid cylinders. JASA. 83, 55–63.
    2. +
  2. Stanton. 1989. Sound scattering by cylinders of finite length. III. Deformed cylinders. JASA. 86: 691-705.
    3. +
  3. Jech et al. 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. JASA. 138: 3742-3764.
    4. +
  4. Foote. 1985. Rather-high-frequency sound scattering by swimbladdered fish. JASA. 78: 688-700.
    5. +
  5. Foote and Francis. 2002. Comparing Kirchhoff approximation and boundary-element models for computing gadoid target strengths. JASA. 111: 1644-1654.
    6. +
  6. Clay and Horne. 1994. Acoustic models of fish: The Atlantic cod (Gadus morhua). JASA. 96: 1661-1668.
    7. +
  7. Clay. 1991. Low-resolution acoustic scattering models: Fluid-filled cylinders and fish with swim bladders. JASA. 89: 2168-2179.
    8. +
  8. Clay. 1992. Composite ray-mode approximations for backscattered sound from gas-filled cylinders and swimbladders. JASA. 92: 2173-2180.
    9. +
  9. Chu et al. 1993. Further analysis of target strength measurements of Antarctic krill at 38 and 120 kHz: Comparison with deformed cylinder model and inference of orientation distribution. JASA. 93: 2985-2988.
    10. +
  10. Stanton et al. 1993. Average echoes from randomly oriented random-length finite cylinders: Zooplankton models. JASA. 94: 3463-3472.]
    11. +
  11. Stanton et al. 1998. Sound scattering by several zooplankton groups II: Scattering models. JASA. 103: 236-253.
    12. +
  12. Jones et al. 2009. Use of the distorted wave Born approximation to predict scattering by inhomogeneous objects: Application to squid. JASA. 125: 73-88.
    13. +
  13. Demer and Conti. 2003. Reconciling theoretical versus empirical target strengths of krill: Effects of phase variability on the distorted wave Born approximation. ICES J. Mar. Sci. 60: 429-434.
    14. +
  14. Demer and Conti. 2004. Erratum: Reconciling theoretical versus empirical target strengths of krill; effects of phase variability on the distorted-wave, Born approximation. ICES J. Mar. Sci. 61: 157-158.
    15. +
  15. TBC
    16. +
+ + + + + + + + + + + + + +
+
+ + + +
+ + + +
+ + + +
+
+
+
+ + + + + + + + + + \ No newline at end of file diff --git a/objects.inv b/objects.inv new file mode 100644 index 0000000..b4d9c96 Binary files /dev/null and b/objects.inv differ diff --git a/other_software/index.html b/other_software/index.html new file mode 100644 index 0000000..0a6fc9e --- /dev/null +++ b/other_software/index.html @@ -0,0 +1,518 @@ + + + + + + + + + + + + + + + + + + + + + + + Other software - echoSMs + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ +
+ + + + + + +
+ + +
+ +
+ + + + + + +
+
+ + + +
+
+
+ + + + + + + +
+
+
+ + + + +
+
+ + + + + + + + + + + + + + + +

Other software

+

Other software that provides source code for acoustic scattering models of relevance to fisheries and plankton acoustics includes:

+
    +
  • acousticTS: R code for calculating scattering using the DCM, DWBA, SDWBA, SDWBA_curved, KRM, MSS model, as well as that of calibration spheres.
    • +
  • Coupled BEM acoustic: Julia code that calculates the TS of three-dimensional shapes with an included object (e.g., a swimbladder).
    • +
  • scatmod: Open source acoustic scattering models for fisheries acoustics. Python and R code for fluid spheres.
    • +
  • FishAcoustics: Contains a Python module that implements the phase-tracking DWBA model.
    • +
  • KRM Model: A web page that uses the KRM model to estimate the TS of predefined or user-supplied shapes over a range of input parameters.
    • +
  • KRMr: KRM model for fish in R.
    • +
  • Liquid spheroid: Julia and C++ code to calculate the scattering by fluid prolate and oblate spheroids.
    • +
  • SDWBA Model: A web page that uses the SDWBA model to estimate the TS of predefined shapes over a range of input parameters.
    • +
  • SDWBA_TS: Matlab code that implements the SDWBA model for Antarctic krill.
    • +
  • sphereTS: Python code to calculate the TS of calibration spheres.
    • +
  • Standard sphere target strength calculator: A web page that calculates the TS of calibration spheres.
    • +
  • tetrascatt: R and C++ code that implements the DWBA model on arbitrary geometries.
    • +
  • ZooScatR: R code that implements the DWBA model.
    • +
+ + + + + + + + + + + + + +
+
+ + + +
+ + + +
+ + + +
+
+
+
+ + + + + + + + + + \ No newline at end of file diff --git a/resources/coordinate_system_dark.svg b/resources/coordinate_system_dark.svg new file mode 100644 index 0000000..30a4741 --- /dev/null +++ b/resources/coordinate_system_dark.svg @@ -0,0 +1,19 @@ + +PyVista Export + +Creator: GL2PS 1.4.2, (C) 1999-2020 C. Geuzaine +For: VTK +CreationDate: Sun Oct 6 10:30:05 2024 + + + + + +φ +θ +ψ +x +y +z + + \ No newline at end of file diff --git a/resources/coordinate_system_light.svg b/resources/coordinate_system_light.svg new file mode 100644 index 0000000..9d1ccb5 --- /dev/null +++ b/resources/coordinate_system_light.svg @@ -0,0 +1,19 @@ + +PyVista Export + +Creator: GL2PS 1.4.2, (C) 1999-2020 C. Geuzaine +For: VTK +CreationDate: Sun Oct 6 10:30:05 2024 + + + + + +φ +θ +ψ +x +y +z + + \ No newline at end of file diff --git a/resources/echoSMs_logo.png b/resources/echoSMs_logo.png new file mode 100644 index 0000000..edc0ecd Binary files /dev/null and b/resources/echoSMs_logo.png differ diff --git a/resources/herring.stl b/resources/herring.stl new file mode 100644 index 0000000..48083c7 Binary files /dev/null and b/resources/herring.stl differ diff --git a/search/search_index.json b/search/search_index.json new file mode 100644 index 0000000..ddff9db --- /dev/null +++ b/search/search_index.json @@ -0,0 +1 @@ +{"config":{"lang":["en"],"separator":"[\\s\\-]+","pipeline":["stopWordFilter"]},"docs":[{"location":"","title":"echoSMs","text":"

Making acoustic scattering models available to fisheries and plankton scientists via the world wide web.

"},{"location":"#background","title":"Background","text":"

This project is an international collaboration that is, in part, a component of a U.S. NOAA-Fisheries active acoustic strategic initiative, AA-SI.

Quantitative interpretation of acoustic echograms requires software expertise to develop advanced analytical methods for echo classification using mathematical models that predict acoustic backscatter (e.g., target strength, TS re 1 m\u00b2 [dB]). These models and predictions can be used to inform echo classification by validating empirical measurements and generating training data for machine learning (ML), artificial intelligence (AI), and other advanced analytical methods, such as inverse methods. Application of these models to fish and plankton requires anatomical and morphological data that are easily accessible and available to the models.

The goal of this project is to make acoustic scattering models available to fisheries and plankton acoustic scientists via the world wide web. By providing the models in an open-access and open-source software language (e.g, Python, R) and providing morphological and anatomical data in open data formats (e.g., HDF5, relational databases), the proper and appropriate use of these models can extend to the entire fisheries and plankton acoustics\u2019 community.

"},{"location":"#contributing-to-echosms","title":"Contributing to echoSMs","text":"

We welcome all contributions to echoSMs, be it code, test cases, bug reports, discussion of models, etc. Please use the github facilities for this (i.e., issues, pull requests, and discussions). We are also happy to accept directly code that we can add to echoSMs on your behalf.

An objective of echoSMs is to provide scattering models in a form that is easy to access, use, and compare to other models. To help with that, we specify model parameter units, angle conventions, and required model outputs that code contributions should support. We also suggest coding conventions that should be followed.

"},{"location":"#scattering-models","title":"Scattering Models","text":"

The initial set of acoustic scattering models will be those used in Jech et al. (2015). Acoustic model development will follow 3\u20134 phases:

  1. Exact solutions and canonical shapes (see table below),
  2. Approximate analytical models applied to canonical shapes (see table below),
  3. Approximate analytical models applied to complex shapes approximating biological targets, such as fish and zooplankton,
  4. Numerical models applied to canonical shapes and biological targets (this phase will depend on time and funding).
"},{"location":"#exact-solutions","title":"Exact Solutions","text":"

The exact solutions, shapes, and supported boundary conditions will be:

Model Shape Description References Modal Series solution (MSS) Sphere Fluid filled 1,2 Fixed rigid 2,3 Pressure release 2 Gas filled 2 Weakly scattering 2 Spherical fluid shell with fluid interior 2 Fixed rigid spherical shell 2 Spherical fluid shell with pressure release interior 2 Spherical fluid shell with gas interior 2 Spherical fluid shell with weakly scattering interior 2 Prolate spheroid modal series solution Prolate spheroid Rigid fixed 2,4,5 Pressure release 2,4,5 Gas filled 2,4,5 Infinite cylinder? 3 Infinite plane?
  1. Anderson, V. C. 1950. Sound scattering from a fluid sphere. JASA. 22(4): 426-431]
  2. Jech et al. 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. JASA. 138: 3742-3764.
  3. Faran, J. J. 1951. Sound scattering by solid cylinders and spheres. JASA. 23(4): 405-418.
  4. Skudrzyk. 1971. The Foundations of Acoustics (Springer, NY), pp. 455-465.
  5. Furusawa. 1988. Prolate spheroidal models for predicting general trends of fish target strength, J. Acoust. Soc. Jpn. (E) 9, 13\u201314.
"},{"location":"#approximate-analytical-models-and-shapes","title":"Approximate Analytical Models and Shapes","text":"

The approximate analytical models,shapes, and supported boundary conditions will be:

Model Shape Description References Modal series-based deformed cylinder model Finite cylinder Fixed rigid 1,2,3 Pressure release 1,2,3 Gas filled 1,2,3 Weakly scattering 1,2,3 Prolate spheroid Fixed rigid 1,2,3 Pressure release 1,2,3 Gas filled 1,2,3 Weakly scattering 1,2,3 Kirchhoff approximation (KA) Sphere Fixed rigid 3,4,5 Prolate spheroid Fixed rigid 3,4,5 Finite cylinder Fixed rigid 3,4,5 Kirchhoff ray mode (KRM) Sphere Gas filled 3,6,7,8 Weakly scattering 3,6,7,8 Spherical shell gas filled 3,6,7,8 Weakly scattering 3,6,7,8 Prolate spheroid gas filled 3,6,7,8 Weakly scattering 3,6,7,8 Finite cylinder gas filled 3,6,7,8 Weakly scattering 3,6,7,8 Distorted wave Born approximation (DWBA) Sphere Weakly scattering 3,9,10,11 Prolate spheroid Weakly scattering 3,9,10,11 Finite cylinder Weakly scattering 3,9,10,11 Phase-tracking distorted wave Born approximation (PT-DWBA) Sphere Weakly scattering 3,12 Spherical shell Weakly scattering 3,12 Prolate spheroid Weakly scattering 3,12 Finite cylinder Weakly scattering 3,12 Stochastic distorted wave Born approximation (SDWBA) Sphere Weakly scattering 13,14,15 Prolate spheroid Weakly scattering 13,14,15 Finite cylinder Weakly scattering 13,14,15
  1. Stanton. 1988. Sound scattering by cylinders of finite length. I. Fluid cylinders. JASA. 83, 55\u201363.
  2. Stanton. 1989. Sound scattering by cylinders of finite length. III. Deformed cylinders. JASA. 86: 691-705.
  3. Jech et al. 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. JASA. 138: 3742-3764.
  4. Foote. 1985. Rather-high-frequency sound scattering by swimbladdered fish. JASA. 78: 688-700.
  5. Foote and Francis. 2002. Comparing Kirchhoff approximation and boundary-element models for computing gadoid target strengths. JASA. 111: 1644-1654.
  6. Clay and Horne. 1994. Acoustic models of fish: The Atlantic cod (Gadus morhua). JASA. 96: 1661-1668.
  7. Clay. 1991. Low-resolution acoustic scattering models: Fluid-filled cylinders and fish with swim bladders. JASA. 89: 2168-2179.
  8. Clay. 1992. Composite ray-mode approximations for backscattered sound from gas-filled cylinders and swimbladders. JASA. 92: 2173-2180.
  9. Chu et al. 1993. Further analysis of target strength measurements of Antarctic krill at 38 and 120 kHz: Comparison with deformed cylinder model and inference of orientation distribution. JASA. 93: 2985-2988.
  10. Stanton et al. 1993. Average echoes from randomly oriented random-length finite cylinders: Zooplankton models. JASA. 94: 3463-3472.]
  11. Stanton et al. 1998. Sound scattering by several zooplankton groups II: Scattering models. JASA. 103: 236-253.
  12. Jones et al. 2009. Use of the distorted wave Born approximation to predict scattering by inhomogeneous objects: Application to squid. JASA. 125: 73-88.
  13. Demer and Conti. 2003. Reconciling theoretical versus empirical target strengths of krill: Effects of phase variability on the distorted wave Born approximation. ICES J. Mar. Sci. 60: 429-434.
  14. Demer and Conti. 2004. Erratum: Reconciling theoretical versus empirical target strengths of krill; effects of phase variability on the distorted-wave, Born approximation. ICES J. Mar. Sci. 61: 157-158.
  15. TBC
"},{"location":"api_reference/","title":"API reference","text":"

This is the API reference for the echoSMs package.

Each type of model is contained in a separate Python class (with name ending in Model), but with common calling signatures across all model classes, as defined in ScatterModelBase. There are also classes to provide ready access to the benchmark models and reference model definitions. There are also utility functions.

"},{"location":"api_reference/#scattermodelbase","title":"ScatterModelBase","text":"

Bases: ABC

Base class for a class that provides a scattering model.

All scattering models should inherit from this class, have a name that ends with 'Model', and provide initialisation and calculate_ts_single() functions.

Initialise.

Attributes:

Name Type Description long_name str

The long name of the model.

short_name str

A short version of the model's long name, typically an acronym.

analytical_type str

Whether the model implements an exact or an approximate model.

boundary_types list[str]

The types of boundary conditions that the model provides, e.g., 'fixed rigid', 'pressure release', 'fluid filled'

shapes list[str]

The target shapes that the model can represent.

max_ka float

An approximate maximum ka value that will result in accurate target strength results. Note that ka is often not the only parameter that determines the accuracy of the model (e.g., aspect ratio and incident angle can also affect the accuracy).

no_expand_parameters list[str]

The model parameters that are not expanded into Pandas DataFrame columns or Xarray DataArray coordinates. They will instead end up as a dict in the DataFrame or DataArray attrs attribute.

Source code in src/echosms/scattermodelbase.py 
@abc.abstractmethod\ndef __init__(self):\n    \"\"\"Initialise.\n\n    Attributes\n    ----------\n    long_name : str\n        The long name of the model.\n    short_name : str\n        A short version of the model's long name, typically an acronym.\n    analytical_type : str\n        Whether the model implements an ``exact`` or an ``approximate`` model.\n    boundary_types : list[str]\n        The types of boundary conditions that the model provides, e.g., 'fixed rigid',\n        'pressure release', 'fluid filled'\n    shapes : list[str]\n        The target shapes that the model can represent.\n    max_ka : float\n        An approximate maximum ka value that will result in accurate target strength results.\n        Note that ka is often not the only parameter that determines the accuracy of the\n        model (e.g., aspect ratio and incident angle can also affect the accuracy).\n    no_expand_parameters : list[str]\n        The model parameters that are not expanded into Pandas DataFrame columns or\n        Xarray DataArray coordinates. They will instead end up as a dict in the DataFrame or\n        DataArray `attrs` attribute.\n    \"\"\"\n    self.long_name = ''\n    self.short_name = ''\n    self.analytical_type = ''\n    self.boundary_types = []\n    self.shapes = []\n    self.max_ka = np.nan\n    self.no_expand_parameters = []\n
"},{"location":"api_reference/#echosms.ScatterModelBase.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.ScatterModelBase.calculate_ts_single","title":"calculate_ts_single() abstractmethod","text":"

Calculate the TS for one parameter set.

Source code in src/echosms/scattermodelbase.py 
@abc.abstractmethod\ndef calculate_ts_single(self):\n    \"\"\"Calculate the TS for one parameter set.\"\"\"\n
"},{"location":"api_reference/#echosms.ScatterModelBase.validate_parameters","title":"validate_parameters(p) abstractmethod","text":"

Validate the model parameters.

Parameters:

Name Type Description Default p dict

Dict containing the model parameters.

required

Raises:

Type Description ValueError

If any of the model parameters are invalid.

KeyError

If any required model parameters are not present.

Source code in src/echosms/scattermodelbase.py 
@abc.abstractmethod\ndef validate_parameters(self, p):\n    \"\"\"Validate the model parameters.\n\n    Parameters\n    ----------\n    p : dict\n        Dict containing the model parameters.\n\n    Raises\n    ------\n    ValueError\n        If any of the model parameters are invalid.\n    KeyError\n        If any required model parameters are not present.\n    \"\"\"\n
"},{"location":"api_reference/#dcmmodel","title":"DCMModel","text":"

Bases: ScatterModelBase

Modal series deformed cylinder model (DCM).

This class contains methods to calculate acoustic scatter from finite straight cylinders with various boundary conditions.

Source code in src/echosms/dcmmodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'deformed cylinder model'\n    self.short_name = 'dcm'\n    self.analytical_type = 'approximate analytical'\n    self.boundary_types = ['fixed rigid', 'pressure release', 'fluid filled']\n    self.shapes = ['finite cylinder']\n    self.max_ka = 20  # [1]\n
"},{"location":"api_reference/#echosms.DCMModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.DCMModel.calculate_ts_single","title":"calculate_ts_single(medium_c, medium_rho, a, b, theta, f, boundary_type, target_c=None, target_rho=None, validate_parameters=True, **kwargs)","text":"

Calculate the scatter from a finite cylinder using the modal series deformed cylinder model.

Parameters:

Name Type Description Default medium_c float

Sound speed in the fluid medium surrounding the target [m/s].

required medium_rho float

Density of the fluid medium surrounding the target [kg/m\u00b3].

required a float

Radius of the cylinderical target [m].

required b float

Length of the cylinderical target [m].

required theta float

Pitch angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required f float

Frequency to calculate the scattering at [Hz].

required boundary_type str

The model type. Supported model types are given in the boundary_types class attribute.

required target_c float

Sound speed in the fluid inside the sphere [m/s]. Only required for boundary_type of fluid filled.

None target_rho float

Density of the fluid inside the sphere [kg/m\u00b3]. Only required for boundary_type of fluid filled.

None validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) of the target [dB].

Notes

The class implements the code in Section B.1 of Jech et al. (2015).

References

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. Journal of the Acoustical Society of America 138, 3742\u20133764. https://doi.org/10.1121/1.4937607

Source code in src/echosms/dcmmodel.py 
def calculate_ts_single(self, medium_c, medium_rho, a, b, theta, f, boundary_type,\n                        target_c=None, target_rho=None, validate_parameters=True,\n                        **kwargs):\n    \"\"\"\n    Calculate the scatter from a finite cylinder using the modal series deformed cylinder model.\n\n    Parameters\n    ----------\n    medium_c : float\n        Sound speed in the fluid medium surrounding the target [m/s].\n    medium_rho : float\n        Density of the fluid medium surrounding the target [kg/m\u00b3].\n    a : float\n        Radius of the cylinderical target [m].\n    b : float\n        Length of the cylinderical target [m].\n    theta : float\n        Pitch angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    f : float\n        Frequency to calculate the scattering at [Hz].\n    boundary_type : str\n        The model type. Supported model types are given in the `boundary_types` class attribute.\n    target_c : float, optional\n        Sound speed in the fluid inside the sphere [m/s].\n        Only required for `boundary_type` of ``fluid filled``.\n    target_rho : float, optional\n        Density of the fluid inside the sphere [kg/m\u00b3].\n        Only required for `boundary_type` of ``fluid filled``.\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) of the target [dB].\n\n    Notes\n    -----\n    The class implements the code in Section B.1 of Jech et al. (2015).\n\n    References\n    ----------\n    Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B.,\n    Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015.\n    Comparisons among ten models of acoustic backscattering used in aquatic ecosystem\n    research. Journal of the Acoustical Society of America 138, 3742\u20133764.\n    <https://doi.org/10.1121/1.4937607>\n    \"\"\"\n    if validate_parameters:\n        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'b': b, 'f': f,\n             'boundary_type': boundary_type, 'target_c': target_c, 'target_rho': target_rho,\n             'theta': theta}\n        self.validate_parameters(p)\n\n    if theta == 0.0:\n        return nan\n\n    theta_rad = theta*pi/180.\n    kL = wavenumber(medium_c, f)*b\n    K = wavenumber(medium_c, f) * sin(theta_rad)\n    Ka = K*a\n\n    m = range(30)  # TODO this needs to vary with f\n\n    match boundary_type:\n        case 'fixed rigid':\n            series = map(lambda m: (-1)**m * Neumann(m)*(jvp(m, Ka) / h1vp(m, Ka)), m)\n        case 'pressure release':\n            series = map(lambda m: (-1)**m * Neumann(m)*(jv(m, Ka) / hankel1(m, Ka)), m)\n        case 'fluid filled':\n            g = target_rho/medium_rho\n            h = target_c/medium_c\n            gh = g*h\n            Kda = K/h*a\n\n            def Cm(m):\n                numer = (jvp(m, Kda)*yv(m, Ka)) / (jv(m, Kda)*jvp(m, Ka))\\\n                    - gh*(yvp(m, Ka)/jvp(m, Ka))\n                denom = (jvp(m, Kda)*jv(m, Ka)) / (jv(m, Kda)*jvp(m, Ka)) - gh\n                return numer/denom\n\n            series = map(lambda m: 1j**(2*m) * Neumann(m) / (1 + 1j*Cm(m)), m)\n        case _:\n            raise ValueError(f'The {self.long_name} model does not support '\n                             f'a model type of \"{boundary_type}\".')\n\n    fbs = 1j*b/pi * (sin(kL*cos(theta_rad)) / (kL*cos(theta_rad))) * sum(series)\n    return 20*log10(abs(fbs))  # ts\n
"},{"location":"api_reference/#echosms.DCMModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/dcmmodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n    super()._present_and_in(p, ['boundary_type'], self.boundary_types)\n    super()._present_and_positive(p, ['medium_rho', 'medium_c', 'a', 'b', 'f'])\n\n    for bt in np.atleast_1d(p['boundary_type']):\n        if bt == 'fluid filled':\n            super()._present_and_positive(p, ['target_c', 'target_rho'])\n
"},{"location":"api_reference/#dwba-models","title":"DWBA models","text":"

There are several models that use the distorted-wave Born approximation, documented below:

"},{"location":"api_reference/#dwba","title":"DWBA","text":"

Bases: ScatterModelBase

Distorted-wave Born approximation scattering model.

Note

The DWBA model is not yet functional.

Source code in src/echosms/dwbamodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'distorted-wave Born approximation'\n    self.short_name = 'dwba'\n    self.analytical_type = 'approximate'\n    self.boundary_types = ['weakly scattering']\n    self.shapes = ['any']\n    self.max_ka = 20\n
"},{"location":"api_reference/#echosms.DWBAModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.DWBAModel.calculate_ts_single","title":"calculate_ts_single(theta, phi, f, target_rho, target_c, validate_parameters=True)","text":"

Distorted-wave Born approximation scattering model.

Implements the distorted-wave Born approximation model for calculating the acoustic backscatter from weakly scattering bodies.

Parameters:

Name Type Description Default theta float

Pitch angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required phi float

Roll angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required f float

Frequency to run the model at [Hz]

required target_rho iterable[float]

Densities of each material. Must have at least the same number of entries as unique integers in volume [kg/m\u00b3].

required target_c iterable[float]

Sound speed of each material. Must have at least the same number of entries as unique integers in volume [m/s].

required validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) [dB] of the target.

Notes

This class implements the method presented in Chu et al. (1993).

References

Chu, D., Foote, K. G., & Stanton, T. K. (1993). Further analysis of target strength measurements of Antarctic krill at 38 and 120 kHz: Comparison with deformed cylinder model and inference or orientation distribution. The Journal of the Acoustical Society of America, 93(5), 2985\u20132988. https://doi.org/10.1121/1.405818

Source code in src/echosms/dwbamodel.py 
def calculate_ts_single(self, theta, phi, f, target_rho, target_c, validate_parameters=True):\n    \"\"\"Distorted-wave Born approximation scattering model.\n\n    Implements the distorted-wave Born approximation\n    model for calculating the acoustic backscatter from weakly scattering bodies.\n\n    Parameters\n    ----------\n    theta : float\n        Pitch angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    phi : float\n        Roll angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    f : float\n        Frequency to run the model at [Hz]\n    target_rho : iterable[float]\n        Densities of each material. Must have at least the same number of entries as unique\n        integers in `volume` [kg/m\u00b3].\n    target_c : iterable[float]\n        Sound speed of each material. Must have at least the same number of entries as unique\n        integers in `volume` [m/s].\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) [dB] of the target.\n\n    Notes\n    -----\n    This class implements the method presented in Chu et al. (1993).\n\n    References\n    ----------\n    Chu, D., Foote, K. G., & Stanton, T. K. (1993). Further analysis of target strength\n    measurements of Antarctic krill at 38 and 120 kHz: Comparison with deformed cylinder\n    model and inference or orientation distribution. The Journal of the Acoustical Society\n    of America, 93(5), 2985\u20132988. <https://doi.org/10.1121/1.405818>\n\n    \"\"\"\n    if validate_parameters:\n        p = {'theta': theta, 'phi': phi, 'f': f, 'target_rho': f, 'target_c': target_c}\n        self.validate_parameters(p)\n    return None\n
"},{"location":"api_reference/#echosms.DWBAModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/dwbamodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n
"},{"location":"api_reference/#pt-dwba","title":"PT-DWBA","text":"

Bases: ScatterModelBase

Phase-tracking distorted-wave Born approximation scattering model.

Source code in src/echosms/ptdwbamodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'phase-tracking distorted-wave Born approximation'\n    self.short_name = 'pt-dwba'\n    self.analytical_type = 'approximate'\n    self.boundary_types = ['weakly scattering']\n    self.shapes = ['unrestricted voxel-based']\n    self.max_ka = 20\n    self.no_expand_parameters = ['volume', 'voxel_size', 'rho', 'c']\n
"},{"location":"api_reference/#echosms.PTDWBAModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.PTDWBAModel.calculate_ts_single","title":"calculate_ts_single(volume, theta, phi, f, voxel_size, rho, c, validate_parameters=True, **kwargs)","text":"

Phase-tracking distorted-wave Born approximation scattering model.

Implements the phase-tracking distorted-wave Born approximation model for calculating the acoustic backscatter from weakly scattering bodies.

Parameters:

Name Type Description Default volume Numpy ndarray[int]

The object to be modelled as a 3D volume of voxels. Array contents should be 0 for the surrounding medium, then increasing by 1 for each additional material type (i.e., 1, 2, 3, etc). volume should be arranged as per the echoSMs coordinate system, where

  • axis 0 (rows) is the x-axis
  • axis 1 (columns) is the y-axis
  • axis 2: (slices) is the z-axis

Increasing axes indices correspond to increasing x, y, and z values.

required theta float

Pitch angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required phi float

Roll angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required f float

Frequency to run the model at [Hz]

required voxel_size iterable[float]

The size of the voxels in volume [m], ordered (x, y, z). This code assumes that the voxels are cubes so y and z are currently irrelevant.

required rho iterable[float]

Densities of each material. Must have at least the same number of entries as unique integers in volume [kg/m\u00b3].

required c iterable[float]

Sound speed of each material. Must have at least the same number of entries as unique integers in volume [m/s].

required validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) [dB] of the target.

Notes

This class implements the method presented in Jones et. al. (2009). The code is based closely on the Matlab code in Jones (2006).

References

Jones, B. A. (2006). Acoustic scattering of broadband echolocation signals from prey of Blainville's beaked whales: Modeling and analysis. Master of Science, Massachusetts Institute of Technology. https://doi.org/10.1575/1912/1283

Jones, B. A., Lavery, A. C., & Stanton, T. K. (2009). Use of the distorted wave Born approximation to predict scattering by inhomogeneous objects: Application to squid. The Journal of the Acoustical Society of America, 125(1), 73-88. https://doi.org/10.1121/1.3021298

Source code in src/echosms/ptdwbamodel.py 
def calculate_ts_single(self, volume, theta, phi, f, voxel_size, rho, c,\n                        validate_parameters=True, **kwargs):\n    \"\"\"Phase-tracking distorted-wave Born approximation scattering model.\n\n    Implements the phase-tracking distorted-wave Born approximation\n    model for calculating the acoustic backscatter from weakly scattering bodies.\n\n    Parameters\n    ----------\n    volume : Numpy ndarray[int]\n        The object to be modelled as a 3D volume of voxels. Array contents should be 0\n        for the surrounding medium, then increasing by 1 for each additional material\n        type (i.e., 1, 2, 3, etc). `volume` should be arranged as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems), where\n\n        - axis 0 (rows) is the _x_-axis\n        - axis 1 (columns) is the _y_-axis\n        - axis 2: (slices) is the _z_-axis\n\n        Increasing axes indices correspond to increasing _x_, _y_, and _z_ values.\n\n    theta : float\n        Pitch angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n\n    phi : float\n        Roll angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n\n    f : float\n        Frequency to run the model at [Hz]\n\n    voxel_size : iterable[float]\n        The size of the voxels in `volume` [m], ordered (_x_, _y_, _z_).\n        This code assumes that the voxels are cubes so _y_ and _z_ are currently irrelevant.\n\n    rho : iterable[float]\n        Densities of each material. Must have at least the same number of entries as unique\n        integers in `volume` [kg/m\u00b3].\n\n    c : iterable[float]\n        Sound speed of each material. Must have at least the same number of entries as unique\n        integers in `volume` [m/s].\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) [dB] of the target.\n\n    Notes\n    -----\n    This class implements the method presented in Jones et. al. (2009). The code is\n    based closely on the Matlab code in Jones (2006).\n\n    References\n    ----------\n    Jones, B. A. (2006). Acoustic scattering of broadband echolocation signals\n    from prey of Blainville's beaked whales: Modeling and analysis. Master of Science,\n    Massachusetts Institute of Technology. <https://doi.org/10.1575/1912/1283>\n\n    Jones, B. A., Lavery, A. C., & Stanton, T. K. (2009). Use of the distorted\n    wave Born approximation to predict scattering by inhomogeneous objects:\n    Application to squid. The Journal of the Acoustical Society of America,\n    125(1), 73-88. <https://doi.org/10.1121/1.3021298>\n    \"\"\"\n    if validate_parameters:\n        p = {'volume': volume, 'theta': theta, 'phi': phi, 'f': f,\n             'voxel_size': voxel_size, 'rho': rho, 'c': c}\n        self.validate_parameters(p)\n\n    # Make sure things are numpy arrays\n    rho = np.atleast_1d(rho)\n    c = np.atleast_1d(c)\n    voxel_size = np.array(voxel_size)\n\n    # volume of the voxels [m^3]\n    dv = voxel_size.prod()\n\n    # input parameter checks\n    if not len(volume.shape) == 3:\n        raise TypeError('The volume input variable must be 3-dimensional.')\n\n    if not voxel_size.shape[0] == 3:\n        raise TypeError('The voxel_size input variable must contain 3 items.')\n\n    if not np.any(voxel_size > 0):\n        raise ValueError('All voxel_size values must be positive.')\n\n    if f < 0.0:\n        raise ValueError('The f input variable must contain only positive values.')\n\n    if (theta < -0.0) or (theta > 180.0):\n        raise ValueError('The theta (pitch) angle must be between -180.0 and +180.0')\n\n    if (phi < -180.0) or (phi > 180.0):\n        raise ValueError('The phi (roll) angle must be between -180.0 and +180.0')\n\n    if volume.min() != 0:\n        raise ValueError('The volume input variable must contain zeros.')\n\n    categories = np.unique(volume)\n    if not len(categories == (volume.max() + 1)):\n        raise ValueError('The integers in volume must include all values in the series '\n                         '(0, 1, 2, ..., n), where n is the largest integer in volume.')\n\n    if not len(rho) >= len(categories):\n        raise ValueError('The target_rho variable must contain at least as many values as '\n                         'unique integers in the volume variable.')\n\n    if not len(c) >= len(categories):\n        raise ValueError('The target_c variable must contain at least as many values '\n                         'as unique integers in the volume variable.')\n\n    # density and sound speed ratios for all object materials\n    g = rho[1:] / rho[0]\n    h = c[1:] / c[0]\n\n    # Do the pitch and roll rotations\n\n    # Convert echoSMs rotation angles (which are intrinsic) into extrinsic as\n    # that is what ndimage.rotate() below uses.\n    if phi == 0.0:  # short circuit the coordinate transformation if we can\n        pitch = theta-90\n        roll = 0.0\n    else:\n        rot = R.from_euler('ZYX', (0, theta-90, -phi), degrees=True)\n        # for backscatter we don't care about yaw\n        _, pitch, roll = rot.as_euler('zyz', degrees=True)\n\n    v = ndimage.rotate(volume, pitch, axes=(0, 2), order=0)\n    v = ndimage.rotate(v, roll, axes=(1, 2), order=0)\n\n    categories = np.unique(v)  # or just take the max?\n\n    # wavenumbers in the various media\n    k = 2.0*np.pi * f / c\n\n    # DWBA coefficients\n    # amplitudes in media 1,2,...,n\n    Cb = 1.0/(g * h**2) + 1.0/g - 2.0  # gamma_kappa - gamma_rho\n    Ca = k[0]**2 * Cb / (4.0*np.pi)  # summation coefficient\n\n    # Differential phase for each voxel.\n    dph = np.zeros(v.shape)\n    masks = []\n    for i, category in enumerate(categories):\n        masks.append(np.isin(v, category))\n        dph[masks[i]] = k[i] * voxel_size[0]\n    masks.pop(0)  # don't need to keep the category[0] mask\n\n    # cumulative summation of phase along the z-direction\n    phase = dph.cumsum(axis=2) - dph/2.0\n    dA = np.zeros(phase.shape, dtype=np.complex128)\n\n    # differential phases for each voxel\n    for i, m in enumerate(masks):\n        dA[m] = Ca[i] * np.exp(2.0*1j*phase[m]) * dv\n\n    # Convert to TS\n    return 20.0 * np.log10(np.abs(dA.sum()))\n
"},{"location":"api_reference/#echosms.PTDWBAModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/ptdwbamodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n
"},{"location":"api_reference/#sdwba","title":"SDWBA","text":"

Bases: ScatterModelBase

Stochastic distorted-wave Born approximation scattering model.

Note

The SDWBA model is not yet functional.

Source code in src/echosms/sdwbamodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = \"stochastic distorted-wave Born approximation\"\n    self.short_name = \"sdwba\"\n    self.analytical_type = \"approximate\"\n    self.boundary_types = [\"weakly scattering\"]\n    self.shapes = [\"any\"]\n    self.max_ka = 20\n
"},{"location":"api_reference/#echosms.SDWBAModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.SDWBAModel.calculate_ts_single","title":"calculate_ts_single(theta, phi, f, target_rho, target_c, validate_parameters=True)","text":"

Stochastic distorted-wave Born approximation scattering model.

Implements the stochastic distorted-wave Born approximation model for calculating the acoustic backscatter from weakly scattering bodies.

Parameters:

Name Type Description Default theta float

Pitch angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required phi float

Roll angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required f float

Frequency to run the model at [Hz]

required target_rho iterable[float]

Densities of each material. Must have at least the same number of entries as unique integers in volume [kg/m\u00b3].

required target_c iterable[float]

Sound speed of each material. Must have at least the same number of entries as unique integers in volume [m/s].

required validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) [dB] of the target.

Notes

This class implements the method presented in Demer & Conti (2003), Demer & Conti (2004), and Conti & Demer (2006).

References

Demer, D. A., & Conti, S. G. (2003). Reconciling theoretical versus empirical target strengths of krill: Effects of phase variability on the distorted-wave Born approximation. ICES Journal of Marine Science, 60, 429-434. https://doi.org/10.1016/S1054-3139(03)00002-X

Demer, D. A., & Conti, S. G. (2004). Reconciling theoretical versus empirical target strengths of krill: Effects of phase variability on the distorted-wave Born approximation. ICES Journal of Marine Science, 61(1), 157-158. https://doi.org/10.1016/j.icesjms.2003.12.003

Conti, S. G., & Demer, D. A. (2006). Improved parameterization of the SDWBA for estimating krill target strength. ICES Journal of Marine Science, 63(5), 928-935. https://doi.org/10.1016/j.icesjms.2006.02.007

Source code in src/echosms/sdwbamodel.py 
def calculate_ts_single(self, theta, phi, f, target_rho, target_c, validate_parameters=True):\n    \"\"\"Stochastic distorted-wave Born approximation scattering model.\n\n    Implements the stochastic distorted-wave Born approximation\n    model for calculating the acoustic backscatter from weakly scattering bodies.\n\n    Parameters\n    ----------\n    theta : float\n        Pitch angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    phi : float\n        Roll angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    f : float\n        Frequency to run the model at [Hz]\n    target_rho : iterable[float]\n        Densities of each material. Must have at least the same number of entries as unique\n        integers in `volume` [kg/m\u00b3].\n    target_c : iterable[float]\n        Sound speed of each material. Must have at least the same number of entries as unique\n        integers in `volume` [m/s].\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) [dB] of the target.\n\n    Notes\n    -----\n    This class implements the method presented in Demer & Conti (2003), Demer & Conti (2004),\n    and Conti & Demer (2006).\n\n    References\n    ----------\n    Demer, D. A., & Conti, S. G. (2003). Reconciling theoretical versus empirical target\n    strengths of krill: Effects of phase variability on the distorted-wave Born approximation.\n    ICES Journal of Marine Science, 60, 429-434.\n    <https://doi.org/10.1016/S1054-3139(03)00002-X>\n\n    Demer, D. A., & Conti, S. G. (2004). Reconciling theoretical versus empirical\n    target strengths of krill: Effects of phase variability on the distorted-wave Born\n    approximation. ICES Journal of Marine Science, 61(1), 157-158.\n    <https://doi.org/10.1016/j.icesjms.2003.12.003>\n\n    Conti, S. G., & Demer, D. A. (2006). Improved parameterization of the SDWBA for estimating\n    krill target strength. ICES Journal of Marine Science, 63(5), 928-935.\n    <https://doi.org/10.1016/j.icesjms.2006.02.007>\n    \"\"\"\n    if validate_parameters:\n        p = {'theta': theta, 'phi': phi, 'f': f, 'target_rho': f, 'target_c': target_c}\n        self.validate_parameters(p)\n\n    return None\n
"},{"location":"api_reference/#echosms.SDWBAModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/sdwbamodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n
"},{"location":"api_reference/#esmodel","title":"ESModel","text":"

Bases: ScatterModelBase

Elastic sphere (ES) scattering model.

This class calculates acoustic backscatter from elastic spheres.

Source code in src/echosms/esmodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'elastic sphere'\n    self.short_name = 'es'\n    self.analytical_type = 'exact'\n    self.boundary_types = ['elastic']\n    self.shapes = ['sphere']\n    self.max_ka = 20  # [1]\n
"},{"location":"api_reference/#echosms.ESModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.ESModel.calculate_ts_single","title":"calculate_ts_single(medium_c, medium_rho, a, f, target_longitudinal_c, target_transverse_c, target_rho, validate_parameters=True, **kwargs)","text":"

Calculate the backscatter from an elastic sphere for one set of parameters.

Parameters:

Name Type Description Default medium_c float

Sound speed in the fluid medium surrounding the sphere [m/s].

required medium_rho float

Density of the fluid medium surrounding the sphere [kg/m\u00b3].

required a float

Radius of the sphere [m].

required f float

Frequency to calculate the scattering at [Hz].

required target_longitudinal_c float

Longitudinal sound speed in the material inside the sphere [m/s].

required target_transverse_c float

Transverse sound speed in the material inside the sphere [m/s].

required target_rho float

Density of the material inside the sphere [kg/m\u00b3].

required validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) of the sphere [dB].

Notes

The class implements the code in MacLennan (1981).

References

MacLennan, D. N. (1981). The Theory of Solid Spheres as Sonar Calibration Targets. Scottish Fisheries Research Report Number 22. Department of Agriculture and Fisheries for Scotland.

Source code in src/echosms/esmodel.py 
def calculate_ts_single(self, medium_c, medium_rho, a, f,\n                        target_longitudinal_c, target_transverse_c, target_rho,\n                        validate_parameters=True,\n                        **kwargs) -> float:\n    \"\"\"\n    Calculate the backscatter from an elastic sphere for one set of parameters.\n\n    Parameters\n    ----------\n    medium_c : float\n        Sound speed in the fluid medium surrounding the sphere [m/s].\n    medium_rho : float\n        Density of the fluid medium surrounding the sphere [kg/m\u00b3].\n    a : float\n        Radius of the sphere [m].\n    f : float\n        Frequency to calculate the scattering at [Hz].\n    target_longitudinal_c : float\n        Longitudinal sound speed in the material inside the sphere [m/s].\n    target_transverse_c : float\n        Transverse sound speed in the material inside the sphere [m/s].\n    target_rho : float\n        Density of the material inside the sphere [kg/m\u00b3].\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) of the sphere [dB].\n\n    Notes\n    -----\n    The class implements the code in MacLennan (1981).\n\n    References\n    ----------\n    MacLennan, D. N. (1981). The Theory of Solid Spheres as Sonar Calibration Targets.\n    Scottish Fisheries Research Report Number 22. Department of Agriculture and Fisheries\n    for Scotland.\n    \"\"\"\n    if validate_parameters:\n        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'f': f,\n             'target_longitudinal_c': target_longitudinal_c,\n             'target_transverse_c': target_transverse_c,\n             'target_rho': target_rho}\n        self.validate_parameters(p)\n\n    q = wavenumber(medium_c, f)*a\n    q1 = q*medium_c/target_longitudinal_c\n    q2 = q*medium_c/target_transverse_c\n    alpha = 2. * (target_rho/medium_rho) * (target_transverse_c/medium_c)**2\n    beta = (target_rho/medium_rho) * (target_longitudinal_c/medium_c)**2 - alpha\n\n    # Use n instead of l (ell) because l looks like 1.\n    def S(n):\n        A2 = (n**2 + n-2) * spherical_jn(n, q2) + q2**2 * spherical_jnpp(n, q2)\n        A1 = 2*n*(n+1) * (q1*spherical_jn(n, q1, True) - spherical_jn(n, q1))\n        B2 = A2*q1**2 * (beta*spherical_jn(n, q1) - alpha*spherical_jnpp(n, q1))\\\n            - A1*alpha * (spherical_jn(n, q2) - q2*spherical_jn(n, q2, True))\n        B1 = q * (A2*q1*spherical_jn(n, q1, True) - A1*spherical_jn(n, q2))\n        eta_n = atan(-(B2*spherical_jn(n, q, True) - B1*spherical_jn(n, q))\n                     / (B2*spherical_yn(n, q, True) - B1*spherical_yn(n, q)))\n\n        return (-1)**n * (2*n+1) * sin(eta_n) * exp(1j*eta_n)\n\n    # Estimate the number of terms to use in the summation\n    n_max = round(q+10)\n    tol = 1e-10  # somewhat arbitrary\n    while abs(S(n_max)) > tol:\n        n_max += 10\n\n    if n_max > 200:\n        warn('TS results may be inaccurate because the modal series required a large '\n             f'number ({n_max}) of terms to converge.')\n\n    n = range(n_max)\n\n    f_inf = -2.0/q * sum(map(S, n))\n\n    return 10*log10(a**2 * abs(f_inf)**2 / 4.0)\n
"},{"location":"api_reference/#echosms.ESModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/esmodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n    super()._present_and_in(p, ['boundary_type'], self.boundary_types)\n    super()._present_and_positive(p, ['medium_rho', 'medium_c', 'a', 'f',\n                                      'target_longitudinal_c',\n                                      'target_transverse_c', 'target_rho'])\n
"},{"location":"api_reference/#kamodel","title":"KAModel","text":"

Bases: ScatterModelBase

Kirchhoff approximation (KA) scattering model.

This class calculates acoustic scatter from arbitrary surfaces.

Source code in src/echosms/kamodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'Kirchhoff approximation'\n    self.short_name = 'ka'\n    self.analytical_type = 'approximate'\n    self.boundary_types = ['pressure release']\n    self.shapes = ['closed surfaces']\n    self.max_ka = 20  # [1]\n    self.no_expand_parameters = ['mesh']\n
"},{"location":"api_reference/#echosms.KAModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.KAModel.calculate_ts_single","title":"calculate_ts_single(medium_c, theta, phi, f, mesh, boundary_type, validate_parameters=True, **kwargs)","text":"

Calculate the scatter using the ka model for one set of parameters.

Parameters:

Name Type Description Default medium_c float

Sound speed in the fluid medium surrounding the target [m/s].

required theta float

Pitch angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required phi float

Roll angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required f float

Frequency to calculate the scattering at [Hz].

required mesh Any

The triangular mesh that defines the scattering surface. This parameter must provide attributes with names of:

  • triangles_center (the position of the centre of each triangular face [m]),
  • face_normals (the outward-pointing unit normals for each triangular face),
  • area_faces (the area of each triangular face [m\u00b2]).

A suitable library for creating and manipulating triangular meshes is trimesh.

required boundary_type str

The boundary type. Supported types are given in the boundary_types class variable.

required validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) of the target [dB].

Notes

The class implements the code in Foote (1985).

References

Foote, K. G. (1985). Rather-high-frequency sound scattering of swimbladdered fish. The Journal of the Acoustical Society of America, 78(2), 688\u2013700. https://doi.org/10.1121/1.392438

Source code in src/echosms/kamodel.py 
def calculate_ts_single(self, medium_c, theta, phi, f, mesh,\n                        boundary_type, validate_parameters=True, **kwargs) -> float:\n    \"\"\"\n    Calculate the scatter using the ka model for one set of parameters.\n\n    Parameters\n    ----------\n    medium_c : float\n        Sound speed in the fluid medium surrounding the target [m/s].\n    theta : float\n        Pitch angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    phi : float\n        Roll angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    f : float\n        Frequency to calculate the scattering at [Hz].\n    mesh : Any\n        The triangular mesh that defines the scattering surface. This parameter must provide\n        attributes with names of:\n\n        - `triangles_center` (the position of the centre of each triangular face [m]),\n        - `face_normals` (the outward-pointing unit normals for each triangular face),\n        - `area_faces` (the area of each triangular face [m\u00b2]).\n\n        A suitable library for creating and manipulating triangular meshes\n        is [trimesh](https://trimesh.org).\n    boundary_type : str\n        The boundary type. Supported types are given in the `boundary_types` class variable.\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) of the target [dB].\n\n    Notes\n    -----\n    The class implements the code in Foote (1985).\n\n    References\n    ----------\n    Foote, K. G. (1985). Rather-high-frequency sound scattering of swimbladdered fish.\n    The Journal of the Acoustical Society of America, 78(2), 688\u2013700.\n    <https://doi.org/10.1121/1.392438>\n\n    \"\"\"\n    if validate_parameters:\n        p = {'medium_c': medium_c, 'theta': theta, 'phi': phi, 'f': f, 'mesh': mesh}\n        self.validate_parameters(p)\n\n    if boundary_type not in self.boundary_types:\n        raise ValueError(f'The {self.long_name} model does not support '\n                         f'a model type of \"{boundary_type}\".')\n\n    # This model keeps the organism fixed and varies the incident wave vector. So need\n    # to convert the theta and phi echoSMs coordinate sytem Tait-Bryan angles\n    # into an (x,y,z) vector.\n\n    # Acoustic wave incident vector and its' norm\n    rot = R.from_euler('ZYX', (0, theta-90, -phi), degrees=True)\n    k_norm = rot.as_matrix() @ np.array([[0, 0, 1]]).T\n    k = k_norm * wavenumber(medium_c, f)\n\n    r = mesh.triangles_center  # position vector of each surface element\n    dS = mesh.area_faces.reshape((-1, 1))  # [m^2]\n\n    kn_nn = mesh.face_normals @ k_norm\n\n    fbs = 1./wavelength(medium_c, f)\\\n        * np.sum(np.exp(2j*r @ k) * np.heaviside(kn_nn, 0.5) * kn_nn * dS)\n\n    return 10*log10(abs(fbs)**2)  # ts\n
"},{"location":"api_reference/#echosms.KAModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/kamodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n    super()._present_and_in(p, ['boundary_type'], self.boundary_types)\n    super()._present_and_positive(p, ['medium_c', 'f'])\n
"},{"location":"api_reference/#mssmodel","title":"MSSModel","text":"

Bases: ScatterModelBase

Modal series solution (MSS) scattering model.

This class calculates acoustic scatter from spheres and shells with various boundary conditions, as listed in the boundary_types class attribute.

Source code in src/echosms/mssmodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'modal series solution'\n    self.short_name = 'mss'\n    self.analytical_type = 'exact'\n    self.boundary_types = ['fixed rigid', 'pressure release', 'fluid filled',\n                           'fluid shell fluid interior',\n                           'fluid shell pressure release interior']\n    self.shapes = ['sphere']\n    self.max_ka = 20  # [1]\n
"},{"location":"api_reference/#echosms.MSSModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.MSSModel.calculate_ts_single","title":"calculate_ts_single(medium_c, medium_rho, a, f, boundary_type, target_c=None, target_rho=None, shell_c=None, shell_rho=None, shell_thickness=None, validate_parameters=True, **kwargs)","text":"

Calculate the scatter using the mss model for one set of parameters.

Parameters:

Name Type Description Default medium_c float

Sound speed in the fluid medium surrounding the target [m/s].

required medium_rho float

Density of the fluid medium surrounding the target [kg/m\u00b3].

required a float

Radius of the spherical target [m].

required f float

Frequency to calculate the scattering at [Hz].

required boundary_type str

The boundary type. Supported types are given in the boundary_types class variable.

required target_c float

Sound speed in the fluid inside the sphere [m/s]. Only required for boundary_type of fluid filled.

None target_rho float

Density of the fluid inside the sphere [kg/m\u00b3]. Only required for boundary_type of fluid filled.

None shell_c float

Sound speed in the spherical shell [m/s]. Only required for boundary_types that include a fluid shell.

None shell_rho float

Density in the spherical shell [kg/m\u00b3]. Only required for boundary_types that include a fluid shell.

None shell_thickness float

Thickness of the spherical shell [m]. This value is subtracted from a to give the radius of the interior sphere. Only required for boundary_types that include a fluid shell.

None validate_parameters bool

Whether to validate the model parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) of the target [dB].

Notes

The class implements the code in Section A.1 of Jech et al. (2015).

References

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. Journal of the Acoustical Society of America 138, 3742\u20133764. https://doi.org/10.1121/1.4937607

Source code in src/echosms/mssmodel.py 
def calculate_ts_single(self, medium_c, medium_rho, a, f, boundary_type,\n                        target_c=None, target_rho=None,\n                        shell_c=None, shell_rho=None, shell_thickness=None,\n                        validate_parameters=True,\n                        **kwargs) -> float:\n    \"\"\"\n    Calculate the scatter using the mss model for one set of parameters.\n\n    Parameters\n    ----------\n    medium_c : float\n        Sound speed in the fluid medium surrounding the target [m/s].\n    medium_rho : float\n        Density of the fluid medium surrounding the target [kg/m\u00b3].\n    a : float\n        Radius of the spherical target [m].\n    f : float\n        Frequency to calculate the scattering at [Hz].\n    boundary_type : str\n        The boundary type. Supported types are given in the `boundary_types` class variable.\n    target_c : float, optional\n        Sound speed in the fluid inside the sphere [m/s].\n        Only required for `boundary_type` of ``fluid filled``.\n    target_rho : float, optional\n        Density of the fluid inside the sphere [kg/m\u00b3].\n        Only required for `boundary_type` of ``fluid filled``.\n    shell_c : float, optional\n        Sound speed in the spherical shell [m/s].\n        Only required for `boundary_type`s that include a fluid shell.\n    shell_rho : float, optional\n        Density in the spherical shell [kg/m\u00b3].\n        Only required for `boundary_type`s that include a fluid shell.\n    shell_thickness : float, optional\n        Thickness of the spherical shell [m]. This value is subtracted from ``a`` to give\n        the radius of the interior sphere.\n        Only required for `boundary_type`s that include a fluid shell.\n    validate_parameters : bool\n        Whether to validate the model parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) of the target [dB].\n\n    Notes\n    -----\n    The class implements the code in Section A.1 of Jech et al. (2015).\n\n    References\n    ----------\n    Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N.,\n    Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015.\n    Comparisons among ten models of acoustic backscattering used in aquatic ecosystem\n    research. Journal of the Acoustical Society of America 138, 3742\u20133764.\n    <https://doi.org/10.1121/1.4937607>\n    \"\"\"\n    if validate_parameters:\n        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'f': f,\n             'boundary_type': boundary_type, 'target_c': target_c, 'target_rho': target_rho,\n             'shell_c': shell_c, 'shell_rho': shell_rho, 'shell_thickness': shell_thickness}\n        self.validate_parameters(p)\n\n    k0 = wavenumber(medium_c, f)\n    ka = k0*a\n    n = np.arange(0, round(ka+20))\n\n    match boundary_type:\n        case 'fixed rigid':\n            A = list(map(lambda x: -spherical_jn(x, ka, True) / h1(x, ka, True), n))\n        case 'pressure release':\n            A = list(map(lambda x: -spherical_jn(x, ka) / h1(x, ka), n))\n        case 'fluid filled':\n            k1a = wavenumber(target_c, f)*a\n            gh = target_rho/medium_rho * target_c/medium_c\n\n            def Cn_fr(n):\n                return\\\n                    ((spherical_jn(n, k1a, True)*spherical_yn(n, ka))\n                        / (spherical_jn(n, k1a)*spherical_jn(n, ka, True))\n                        - gh*(spherical_yn(n, ka, True)/spherical_jn(n, ka, True)))\\\n                    / ((spherical_jn(n, k1a, True)*spherical_jn(n, ka))\n                       / (spherical_jn(n, k1a)*spherical_jn(n, ka, True))-gh)\n\n            A = -1/(1 + 1j*np.asarray(list(map(Cn_fr, n)), dtype=complex))\n        case 'fluid shell fluid interior':\n            b = a - shell_thickness\n\n            g21 = shell_rho / medium_rho\n            h21 = shell_c / medium_c\n            g32 = target_rho / shell_rho\n            h32 = target_c / shell_c\n\n            k1a = wavenumber(medium_c, f) * a\n            k2 = wavenumber(shell_c, f)\n            k3b = wavenumber(target_c, f) * b\n\n            def Cn_fsfi(n):\n                (b1, b2, a11, a21, a12, a22, a32, a13, a23, a33) =\\\n                    MSSModel.__eqn9(n, k1a, g21, h21, k2*a, k2*b, k3b, h32, g32)\n                return (b1*a22*a33 + a13*b2*a32 - a12*b2*a33 - b1*a23*a32)\\\n                    / (a11*a22*a33 + a13*a21*a32 - a12*a21*a33 - a11*a23*a32)\n\n            A = list(map(Cn_fsfi, n))\n        case 'fluid shell pressure release interior':\n            b = a - shell_thickness\n\n            g21 = shell_rho / medium_rho\n            h21 = shell_c / medium_c\n\n            k1a = wavenumber(medium_c, f) * a\n            k2 = wavenumber(shell_c, f)\n            ksa = k2 * a  # ksa is used in the paper, but isn't that the same as k2a?\n\n            def Cn_fspri(n):\n                (b1, b2, d1, d2, a11, a21) = MSSModel.__eqn10(n, k1a, g21, h21, ksa, k2*a, k2*b)\n                return (b1*d2-d1*b2) / (a11*d2-d1*a21)\n\n            A = list(map(Cn_fspri, n))\n        case _:\n            raise ValueError(f'The {self.long_name} model does not support '\n                             f'a model type of \"{boundary_type}\".')\n\n    fbs = -1j/k0 * np.sum((-1)**n * (2*n+1) * A)\n    return 20*log10(abs(fbs))  # ts\n
"},{"location":"api_reference/#echosms.MSSModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/mssmodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n\n    \"\"\"\n    p = as_dict(params)\n    super()._present_and_in(p, ['boundary_type'], self.boundary_types)\n    super()._present_and_positive(p, ['medium_rho', 'a', 'f'])\n\n    for bt in np.atleast_1d(p['boundary_type']):\n        match bt:\n            case 'fluid filled':\n                super()._present_and_positive(p, ['target_c', 'target_rho'])\n            case 'fluid shell fluid interior':\n                super()._present_and_positive(p, ['target_c', 'target_rho', 'shell_c',\n                                                  'shell_rho', 'shell_thickness'])\n            case 'fluid shell pressure release interior':\n                super()._present_and_positive(p, ['shell_c', 'shell_rho', 'shell_thickness'])\n
"},{"location":"api_reference/#psmsmodel","title":"PSMSModel","text":"

Bases: ScatterModelBase

Prolate spheroidal modal series (PSMS) scattering model.

Note

The fluid filled boundary type implementation is currently only accurate for weakly scattering interiors. Support for strongly scattering (e.g., gas-filled) will come later.

Source code in src/echosms/psmsmodel.py 
def __init__(self):\n    super().__init__()\n    self.long_name = 'prolate spheroidal modal series'\n    self.short_name = 'psms'\n    self.analytical_type = 'exact'\n    self.boundary_types = ['fixed rigid', 'pressure release', 'fluid filled']\n    self.shapes = ['prolate spheroid']\n    self.max_ka = 10  # [1]\n
"},{"location":"api_reference/#echosms.PSMSModel.calculate_ts","title":"calculate_ts(data, expand=False, inplace=False, multiprocess=False)","text":"

Calculate the target strength (TS) for many parameters.

Parameters:

Name Type Description Default data Pandas DataFrame, Xarray DataArray or dict

Requirements for the different input data types are:

  • DataFrame: column names must match the function parameter names in calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.
  • DataArray: dimension names must match the function parameter names in calculate_ts_single(). TS values will be calculated for all combinations of the coordinate variables.
  • dict: keys must match the function parameters in calculate_ts_single(). TS values will be calculated for all combinations of the dict values.
required multiprocess bool

Split the ts calculation across CPU cores. Multiprocessing is currently provided by mapply with little customisation. For more sophisticated uses it may be preferred to use a multiprocessing package of your choice directly on the calculate_ts_single() method. See the code in this method (calculate_ts()) for an example.

False expand bool

Only applicable if data is a dict. If True, will use as_dataframe() to expand the dict into a DataFrame with one column per dict key and return that, adding a column named ts for the results.

False inplace bool

Only applicable if data is a DataFrame. If True, the results will be added to the input DataFrame in a column named ts. If a ts column already exists, it is overwritten.

False

Returns:

Type Description None, list[float], Series, or DataFrame

The return type and value are determined by the type of the input variable (data) and the expand and inplace parameters:

  • dict input and expand=False returns a list of floats.
  • dict input and expand=True returns a DataFrame.
  • DataFrame input and inplace=False returns a Series.
  • DataFrame input and inplace=True modifies data and returns None.
  • DataArray input always modifies data and returns None.
Source code in src/echosms/scattermodelbase.py 
def calculate_ts(self, data, expand=False, inplace=False, multiprocess=False):\n    \"\"\"Calculate the target strength (TS) for many parameters.\n\n    Parameters\n    ----------\n    data : Pandas DataFrame, Xarray DataArray or dict\n        Requirements for the different input data types are:\n\n        - **DataFrame**: column names must match the function parameter names in\n          calculate_ts_single(). One TS value will be calculated for each row in the DataFrame.\n        - **DataArray**: dimension names must match the function parameter names in\n          calculate_ts_single(). TS values will be calculated for all combinations of the\n          coordinate variables.\n        - **dict**: keys must match the function parameters in calculate_ts_single().\n          TS values will be calculated for all combinations of the dict values.\n\n    multiprocess : bool\n        Split the ts calculation across CPU cores. Multiprocessing is currently provided by\n        [mapply](https://github.com/ddelange/mapply) with little customisation. For more\n        sophisticated uses it may be preferred to use a multiprocessing package of your choice\n        directly on the `calculate_ts_single()` method. See the code in this method\n        (`calculate_ts()`) for an example.\n\n    expand : bool\n        Only applicable if `data` is a dict. If `True`, will use\n        [`as_dataframe()`][echosms.utils.as_dataframe]\n        to expand the dict into a DataFrame with one column per dict key\n        and return that, adding a column named `ts` for the results.\n\n    inplace : bool\n        Only applicable if `data` is a DataFrame. If `True`, the results\n        will be added to the input DataFrame in a column named `ts`. If a `ts` column\n        already exists, it is overwritten.\n\n    Returns\n    -------\n    : None, list[float], Series, or DataFrame\n        The return type and value are determined by the type of the input variable (`data`) and\n        the `expand` and `inplace` parameters:\n\n        - dict input and `expand=False` returns a list of floats.\n        - dict input and `expand=True` returns a DataFrame.\n        - DataFrame input and `inplace=False` returns a Series.\n        - DataFrame input and `inplace=True` modifies `data` and returns `None`.\n        - DataArray input always modifies `data` and returns `None`.\n\n    \"\"\"\n    match data:\n        case dict():\n            data_df = as_dataframe(data, self.no_expand_parameters)\n        case pd.DataFrame():\n            data_df = data\n        case xr.DataArray():\n            data_df = data.to_dataframe().reset_index()\n            data_df.attrs = data.attrs\n        case _:\n            raise ValueError(f'Data type of {type(data)} is not supported'\n                             ' (only dictionaries, Pandas DataFrames and'\n                             ' Xarray DataArrays are).')\n\n    self.validate_parameters(data_df)\n\n    # Get the non-expandable model parameters\n    p = data_df.attrs['parameters'] if 'parameters' in data_df.attrs else {}\n\n    # Note: the args argument in the apply call below requires a tuple. data_df.attrs is a\n    # dict and the default behaviour is to make a tuple using the dict keys. The trailing comma\n    # and parenthesis instead causes the tuple to have one entry of the dict.\n\n    if multiprocess:\n        from mapply.mapply import mapply\n        ts = mapply(data_df, self.__ts_helper, args=(p,), axis=1)\n    else:  # this uses just one CPU\n        ts = data_df.apply(self.__ts_helper, args=(p,), axis=1)\n\n    match data:\n        case dict() if expand:\n            data_df['ts'] = ts\n            return data_df\n        case dict():\n            return ts.to_list()\n        case pd.DataFrame() if inplace:\n            data_df['ts'] = ts\n            return None\n        case pd.DataFrame():\n            return ts.rename('ts', inplace=True)\n        case xr.DataArray():\n            data.values = ts.to_numpy().reshape(data.shape)\n            return None\n        case _:\n            raise AssertionError('This code should never be reached - unsupported input data '\n                                 f'type of {type(data)}.')\n
"},{"location":"api_reference/#echosms.PSMSModel.calculate_ts_single","title":"calculate_ts_single(medium_c, medium_rho, a, b, theta, f, boundary_type, target_c=None, target_rho=None, validate_parameters=True)","text":"

Prolate spheroid modal series (PSMS) solution model.

Parameters:

Name Type Description Default medium_c float

Sound speed in the fluid medium surrounding the target [m/s].

required medium_rho float

Density of the fluid medium surrounding the target [kg/m\u00b3].

required a float

Prolate spheroid major axis radius [m].

required b float

Prolate spheroid minor axis radius [m].

required theta float

Pitch angle to calculate the scattering as per the echoSMs coordinate system [\u00b0].

required f float

Frequency to calculate the scattering at [Hz].

required boundary_type str

The model type. Supported model types are given in the boundary_types class variable.

required target_c float

Sound speed in the fluid inside the target [m/s]. Only required for boundary_type of fluid filled.

None target_rho float

Density of the fluid inside the target [kg/m\u00b3]. Only required for boundary_type of fluid filled.

None validate_parameters bool

Whether to validate the input parameters.

True

Returns:

Type Description float

The target strength (re 1 m\u00b2) of the target [dB].

Notes

The backscattered target strength of a pressure release or fluid-filled prolate spheroid is calculated using the PSMS method of Furusawa (1988) and corrections in Furusawa et al. (1994).

References

Furusawa, M. (1988). \"Prolate spheroidal models for predicting general trends of fish target strength,\" J. Acoust. Soc. Jpn. 9, 13-24. Furusawa, M., Miyanohana, Y., Ariji, M., and Sawada, Y. (1994). \u201cPrediction of krill target strength by liquid prolate spheroid model,\u201d Fish. Sci., 60, 261-265.

Source code in src/echosms/psmsmodel.py 
def calculate_ts_single(self, medium_c, medium_rho, a, b, theta, f, boundary_type,\n                        target_c=None, target_rho=None, validate_parameters=True):\n    \"\"\"Prolate spheroid modal series (PSMS) solution model.\n\n    Parameters\n    ----------\n    medium_c : float\n        Sound speed in the fluid medium surrounding the target [m/s].\n    medium_rho : float\n        Density of the fluid medium surrounding the target [kg/m\u00b3].\n    a : float\n        Prolate spheroid major axis radius [m].\n    b : float\n        Prolate spheroid minor axis radius [m].\n    theta : float\n        Pitch angle to calculate the scattering as per the echoSMs\n        [coordinate system](https://ices-tools-dev.github.io/echoSMs/\n        conventions/#coordinate-systems) [\u00b0].\n    f : float\n        Frequency to calculate the scattering at [Hz].\n    boundary_type : str\n        The model type. Supported model types are given in the `boundary_types` class variable.\n    target_c : float\n        Sound speed in the fluid inside the target [m/s].\n        Only required for `boundary_type` of ``fluid filled``.\n    target_rho : float\n        Density of the fluid inside the target [kg/m\u00b3].\n        Only required for `boundary_type` of ``fluid filled``.\n    validate_parameters : bool\n        Whether to validate the input parameters.\n\n    Returns\n    -------\n    : float\n        The target strength (re 1 m\u00b2) of the target [dB].\n\n    Notes\n    -----\n    The backscattered target strength of a pressure release or fluid-filled prolate spheroid\n    is calculated using the PSMS method of Furusawa (1988) and corrections in\n    Furusawa et al. (1994).\n\n    References\n    ----------\n    Furusawa, M. (1988). \"Prolate spheroidal models for predicting general\n        trends of fish target strength,\" J. Acoust. Soc. Jpn. 9, 13-24.\n    Furusawa, M., Miyanohana, Y., Ariji, M., and Sawada, Y. (1994).\n        \u201cPrediction of krill target strength by liquid prolate spheroid\n        model,\u201d Fish. Sci., 60, 261-265.\n    \"\"\"\n    if validate_parameters:\n        p = {'medium_c': medium_c, 'medium_rho': medium_rho, 'a': a, 'b': b,\n             'theta': theta, 'f': f, 'boundary_type': boundary_type,\n             'target_c': target_c, 'target_rho': target_rho}\n        self.validate_parameters(p)\n\n    if boundary_type not in self.boundary_types:\n        raise ValueError(f'The {self.long_name} model does not support '\n                         f'a model type of \"{boundary_type}\".')\n\n    xim = (1.0 - (b/a)**2)**(-.5)\n    q = a/xim  # semi-focal length\n\n    km = wavenumber(medium_c, f)\n    hm = km*q\n\n    if boundary_type == 'fluid filled':\n        g = target_rho / medium_rho\n        ht = wavenumber(target_c, f)*q\n\n    # Phi, the port/starboard angle is fixed for this code\n    phi_inc = np.pi  # incident direction\n    phi_sca = np.pi + phi_inc  # scattered direction\n\n    theta_inc = np.deg2rad(theta)  # incident direction\n    theta_sca = np.pi - theta_inc  # scattered direction\n\n    # Approximate limits on the summations\n    m_max = int(np.ceil(2*km*b))\n    n_max = int(m_max + np.ceil(hm/2))\n\n    f_sca = 0.0\n    for m in range(m_max+1):\n        epsilon_m = Neumann(m)\n        cos_term = np.cos(m*(phi_sca - phi_inc))\n        for n in range(m, n_max+1):\n            Smn_inc, _ = pro_ang1(m, n, hm, np.cos(theta_inc))\n            Smn_sca, _ = pro_ang1(m, n, hm, np.cos(theta_sca))\n            # The Meixner-Sch\u00e4fke normalisation scheme for the angular function of the first\n            # kind. Refer to eqn 21.7.11 in Abramowitz, M., and Stegun, I. A. (1964).\n            # Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables\n            # (Dover, New York), 10th ed.\n            N_mn = 2/(2*n+1) * factorial(n+m) / factorial(n-m)\n\n            R1m, dR1m = pro_rad1(m, n, hm, xim)\n            R2m, dR2m = pro_rad2(m, n, hm, xim)\n\n            match boundary_type:\n                case 'fluid filled':\n                    # Note: we can implement the simpler equations if impedances are\n                    # similar between the medium and the target. The gas-filled\n                    # condition does not meet that, so we have two paths here. The simplified\n                    # equations are quicker, so it is worth to do.\n                    if (abs(1.0-target_c/medium_c) <= 0.01) and (abs(1.0-g) <= 0.01):\n                        Amn = PSMSModel._fluidfilled_approx(m, n, hm, ht, xim, g)\n                    else:\n                        Amn = PSMSModel._fluidfilled_exact(m, n, hm, ht, xim, g, theta_inc)\n                case 'pressure release':\n                    Amn = -R1m/(R1m + 1j*R2m)\n                case 'fixed rigid':\n                    Amn = -dR1m/(dR1m + 1j*dR2m)\n\n            f_sca += epsilon_m * (Smn_inc / N_mn) * Smn_sca * Amn * cos_term\n\n    return 20*np.log10(np.abs(-2j / km * f_sca))\n
"},{"location":"api_reference/#echosms.PSMSModel.validate_parameters","title":"validate_parameters(params)","text":"

Validate the model parameters.

See here for calling details.

Source code in src/echosms/psmsmodel.py 
def validate_parameters(self, params):\n    \"\"\"Validate the model parameters.\n\n    See [here][echosms.ScatterModelBase.validate_parameters] for calling details.\n    \"\"\"\n\n    p = as_dict(params)\n    super()._present_and_in(p, ['boundary_type'], self.boundary_types)\n    super()._present_and_positive(p, ['medium_c', 'medium_rho', 'a', 'b', 'f'])\n\n    for bt in np.atleast_1d(p['boundary_type']):\n        match bt:\n            case 'fluid filled':\n                super()._present_and_positive(p, ['target_c', 'target_rho'])\n
"},{"location":"api_reference/#referencemodels","title":"ReferenceModels","text":"

Provide access to reference scattering model parameters.

Reference models are the models and parameters defined in Jech et al. (2015). The parameters are stored in a TOML-formatted file in the echoSMs repository and this class provides easy access to the data in that file. Additional reference models may be defined in the future and added to the TOML file (for example, entries have been added for all known calibration sphere sizes).

Attributes:

Name Type Description definitions dict

A dict representation of the target definitions.toml file.

Raises:

Type Description TOMLDecodeError

If the target definitions.toml file is not valid TOML.

KeyError

If the target definitions.toml file has multiple target entries with the same name.

References

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. Journal of the Acoustical Society of America 138, 3742\u20133764. https://doi.org/10.1121/1.4937607

Source code in src/echosms/referencemodels.py 
def __init__(self):\n    self.defs_filename = Path(__file__).parent/Path('resources')/Path('target definitions.toml')\n\n    self.definitions = []\n\n    with open(self.defs_filename, 'rb') as f:\n        try:\n            self.definitions = tomllib.load(f)\n        except tomllib.TOMLDecodeError as e:\n            raise SyntaxError(f'Error while parsing file \"{self.defs_filename.name}\"') from e\n\n    # Flag duplicate target names\n    pda = pd.Series(self.names())\n    duplicates = list(pda[pda.duplicated()])\n    if duplicates:\n        raise KeyError(f'The \"{self.defs_filename.name}\" file has multiple targets '\n                       f'with the same name: '+', '.join(duplicates))\n\n    # Substitute parameters names in the target section by the values of those\n    # parameters.\n    for t in self.definitions['target']:\n        for k, v in t.items():\n            try:\n                t[k] = self.definitions['parameters'][v]\n            except (KeyError, TypeError):\n                pass\n
"},{"location":"api_reference/#echosms.ReferenceModels.names","title":"names()","text":"

Names of all model definitions.

Returns:

Type Description iterable of str

All model names in the target definitions.toml file.

Source code in src/echosms/referencemodels.py 
def names(self):\n    \"\"\"Names of all model definitions.\n\n    Returns\n    -------\n    : iterable of str\n        All model names in the ``target definitions.toml`` file.\n    \"\"\"\n    return [n['name'] for n in self.definitions['target']]\n
"},{"location":"api_reference/#echosms.ReferenceModels.parameters","title":"parameters(name)","text":"

Model parameters for a particular model.

Model parameters are a subset of the model specification where the metadata items have been removed.

Parameters:

Name Type Description Default name str

The name of a model in the target definitions.toml file.

required

Returns:

Type Description dict

The model parameters for the requested model or an empty set if no model with that name.

Source code in src/echosms/referencemodels.py 
def parameters(self, name):\n    \"\"\"Model parameters for a particular model.\n\n    Model parameters are a subset of the model specification where the metadata items have\n    been removed.\n\n    Parameters\n    ----------\n    name : str\n        The name of a model in the ``target definitions.toml`` file.\n\n    Returns\n    -------\n    : dict\n        The model parameters for the requested model or an empty set if no model with that name.\n\n    \"\"\"\n    s = self.specification(name)\n\n    if not s:\n        return []\n\n    # Remove the entries that are not parameters\n    p = s.copy()\n    for k in ['name', 'shape', 'description', 'source', 'benchmark_model']:\n        p.pop(k, None)\n    return p\n
"},{"location":"api_reference/#echosms.ReferenceModels.specification","title":"specification(name)","text":"

Model definitions for a particular model.

Parameters:

Name Type Description Default name str

The name of a model in the target definitions.toml file.

required

Returns:

Type Description dict

The model definitions for the requested model or an empty set if no model with that name.

Source code in src/echosms/referencemodels.py 
def specification(self, name):\n    \"\"\"Model definitions for a particular model.\n\n    Parameters\n    ----------\n    name : str\n        The name of a model in the ``target definitions.toml`` file.\n\n    Returns\n    -------\n    : dict\n        The model definitions for the requested model or an empty set if no model\n        with that name.\n    \"\"\"\n    s = [t for t in self.definitions['target'] if t['name'] == name]\n    if not s:\n        return s\n\n    return s[0]\n
"},{"location":"api_reference/#benchmarkdata","title":"BenchmarkData","text":"

Convenient interface to the benchmark dataset.

This dataset contains the TS results from Jech et al. (2015).

Attributes:

Name Type Description angle_dataset Pandas DataFrame

The angle dataset from the benchmark model runs.

freq_dataset Pandas DataFrame

The frequency dataset from the benchmark model runs.

Notes

The column names in the source benchmark files have been changed to be the same as those used in the ReferenceModels model definitions.

References

Jech, J.M., Horne, J.K., Chu, D., Demer, D.A., Francis, D.T.I., Gorska, N., Jones, B., Lavery, A.C., Stanton, T.K., Macaulay, G.J., Reeder, D.B., Sawada, K., 2015. Comparisons among ten models of acoustic backscattering used in aquatic ecosystem research. Journal of the Acoustical Society of America 138, 3742-3764. https://doi.org/10.1121/1.4937607

Source code in src/echosms/benchmarkdata.py 
def __init__(self):\n\n    data_directory = Path(__file__).parent/Path('resources')/Path('BenchMark_Data')\n\n    angle_data_file = data_directory/'Benchmark_Angle_TS.csv'\n    freq_data_file = data_directory/'Benchmark_Frequency_TS.csv'\n\n    self.angle_dataset = pd.read_csv(angle_data_file)\n    self.freq_dataset = pd.read_csv(freq_data_file)\n\n    # Change the column names to match the reference model names used in ReferenceModels\n    self.angle_dataset.rename(columns=BenchmarkData.a_rename, inplace=True)\n    self.freq_dataset.rename(columns=BenchmarkData.f_rename, inplace=True)\n
"},{"location":"api_reference/#utilities","title":"Utilities","text":"

Miscellaneous utility functions.

"},{"location":"api_reference/#echosms.utils.Neumann","title":"Neumann(m)","text":"

Neumann number.

Parameters:

Name Type Description Default m int

The input integer.

required

Returns:

Type Description int

The Neumann number.

Source code in src/echosms/utils.py 
def Neumann(m: int) -> int:\n    \"\"\"Neumann number.\n\n    Parameters\n    ----------\n    m :\n        The input integer.\n\n    Returns\n    -------\n    :\n        The Neumann number.\n    \"\"\"\n    if m == 0:\n        return 1\n    return 2\n
"},{"location":"api_reference/#echosms.utils.as_dataarray","title":"as_dataarray(params, no_expand=[])","text":"

Convert model parameters from dict form to a Xarray DataArray.

Parameters:

Name Type Description Default params dict

The model parameters.

required no_expand list

Key values of the non-expandable model parameters in params.

[]

Returns:

Type Description DataArray

Returns a multi-dimensional DataArray generated from the Cartesian product of all expandable items in the input dict. Non-expandable items are added to the DataArray attrs property. Expandable items are those that can be sensibly expanded into DataArray coordinates. Not all models have non-expandable items. The array is named ts, the values are initialised to nan, the dimension names are the dict keys, and the coordinate variables are the dict values.

Source code in src/echosms/utils.py 
def as_dataarray(params: dict, no_expand: list = []) -> xr.DataArray:\n    \"\"\"Convert model parameters from dict form to a Xarray DataArray.\n\n    Parameters\n    ----------\n    params :\n        The model parameters.\n\n    no_expand :\n        Key values of the non-expandable model parameters in `params`.\n\n    Returns\n    -------\n    :\n        Returns a multi-dimensional DataArray generated from the Cartesian product of all\n        expandable items in the input dict. Non-expandable items are added to the DataArray\n        attrs property. Expandable items are those that can be sensibly expanded into\n        DataArray coordinates. Not all models have non-expandable items.\n        The array is named `ts`, the values are initialised to `nan`, the\n        dimension names are the dict keys, and the coordinate variables are the dict values.\n\n    \"\"\"\n    expand, nexpand = split_dict(params, no_expand)\n\n    # Convert scalars to iterables so xarray is happy\n    for k, v in expand.items():\n        if not isinstance(v, Iterable) or isinstance(v, str):\n            expand[k] = [v]\n\n    sz = [len(v) for k, v in expand.items()]\n    return xr.DataArray(data=np.full(sz, np.nan), coords=expand, name='ts',\n                        attrs={'units': 'dB', 'dB_reference': '1 m^2',\n                               'parameters': nexpand})\n
"},{"location":"api_reference/#echosms.utils.as_dataframe","title":"as_dataframe(params, no_expand=[])","text":"

Convert model parameters from dict form to a Pandas DataFrame.

Parameters:

Name Type Description Default params dict

The model parameters.

required no_expand list

Key values of the non-expandable model parameters in params.

[]

Returns:

Type Description DataFrame

Returns a Pandas DataFrame generated from the Cartesian product of all expandable items in the input dict. DataFrame column names are obtained from the dict keys. Non-expandable items are added to the DataFrame attrs property. Expandable items are those that can be sensibly expanded into DataFrame columns. Not all models have non-expandable items.

Source code in src/echosms/utils.py 
def as_dataframe(params: dict, no_expand: list = []) -> pd.DataFrame:\n    \"\"\"Convert model parameters from dict form to a Pandas DataFrame.\n\n    Parameters\n    ----------\n    params :\n        The model parameters.\n\n    no_expand :\n        Key values of the non-expandable model parameters in `params`.\n\n    Returns\n    -------\n    :\n        Returns a Pandas DataFrame generated from the Cartesian product of all expandable\n        items in the input dict. DataFrame column names are obtained from the dict keys.\n        Non-expandable items are added to the DataFrame attrs property. Expandable items are\n        those that can be sensibly expanded into DataFrame columns. Not all models have\n        non-expandable items.\n\n    \"\"\"\n    expand, nexpand = split_dict(params, no_expand)\n\n    # Use meshgrid to do the Cartesian product then create a Pandas DataFrame from that, having\n    # flattened the multidimensional arrays and using a dict to provide column names.\n    # This preserves the differing dtypes in each column compared to other ways of\n    # constructing the DataFrame).\n    df = pd.DataFrame({k: t.flatten()\n                       for k, t in zip(expand.keys(), np.meshgrid(*tuple(expand.values())))})\n    df.attrs = {'parameters': nexpand}\n    return df\n
"},{"location":"api_reference/#echosms.utils.as_dict","title":"as_dict(params)","text":"

Convert model parameters from DataFrame or DataArray to dict.

Parameters:

Name Type Description Default params dict | DataFrame | DataArray

The model parameters

required

Raises:

Type Description TypeError:

If the input data type is not supported.

Returns:

Type Description dict

A dict containing the model parameters.

Source code in src/echosms/utils.py 
def as_dict(params: dict | pd.DataFrame | xr.DataArray) -> dict:\n    \"\"\"Convert model parameters from DataFrame or DataArray to dict.\n\n    Parameters\n    ----------\n    params:\n        The model parameters\n\n    Raises\n    ------\n    TypeError:\n        If the input data type is not supported.\n\n    Returns\n    -------\n    :\n        A dict containing the model parameters.\n    \"\"\"\n    if isinstance(params, dict):\n        return params\n\n    # Get the non-expandable model parameters\n    p = params.attrs['parameters'] if 'parameters' in params.attrs else {}\n\n    if isinstance(params, xr.DataArray):\n        return dict(zip(params.coords, params.indexes.values())) | p\n    elif isinstance(params, pd.DataFrame):\n        return params.to_dict(orient='list') | p\n\n    raise TypeError('Only dict, DataFrame, or DataArray are accepted.')\n
"},{"location":"api_reference/#echosms.utils.h1","title":"h1(n, z, derivative=False)","text":"

Spherical Hankel function of the first kind or its' derivative.

Parameters:

Name Type Description Default n int

Order (n \u2265 0).

required z float

Argument of the Hankel function.

required derivative

if True, the value of the derivative (rather than the function itself) is returned.

False

Returns:

Type Description complex

Value of the spherical Hankel function

Raises:

Type Description ValueError

For negative n values.

Notes

The value of the Hankel function is calculated from spherical Bessel functions [1].

The derivative is computed from spherical Hankel functions [2].

References

[1] https://dlmf.nist.gov/10.47.E10

[2] https://dlmf.nist.gov/10.51.E2

Source code in src/echosms/utils.py 
def h1(n: int, z: float, derivative=False) -> complex:\n    \"\"\"Spherical Hankel function of the first kind or its' derivative.\n\n    Parameters\n    ----------\n    n :\n        Order (n \u2265 0).\n    z :\n        Argument of the Hankel function.\n    derivative :\n        if True, the value of the derivative (rather than the function itself) is returned.\n\n    Returns\n    -------\n    :\n        Value of the spherical Hankel function\n\n    Raises\n    ------\n    ValueError\n        For negative n values.\n\n    Notes\n    -----\n    The value of the Hankel function is calculated from spherical Bessel functions [1].\n\n    The derivative is computed from spherical Hankel functions [2].\n\n    References\n    ----------\n    [1] <https://dlmf.nist.gov/10.47.E10>\n\n    [2] <https://dlmf.nist.gov/10.51.E2>\n    \"\"\"\n    if n < 0:\n        raise ValueError('Negative n values are not supported for spherical Hankel functions.')\n\n    if not derivative:\n        return spherical_jn(n, z) + 1j*spherical_yn(n, z)\n    return -h1(n+1, z) + (n/z) * h1(n, z)\n
"},{"location":"api_reference/#echosms.utils.pro_ang1","title":"pro_ang1(m, n, c, eta, norm=False)","text":"

Prolate spheroidal angular function of the first kind and derivative.

Calculates the prolate spheroidal angular function of the first kind and its' derivative with respect to eta.

Parameters:

Name Type Description Default m int

The order parameter (\u2265 0)

required n int

The degree parameter (\u2265 m).

required c float

The size parameter.

required eta float

The angular coordinate, \u03b7, where |\u03b7| \u2264 1.

required norm

If False, returned values are not normalised (i.e., the Meixner-Sch\u00e4fke normlalisation scheme is used). For large m this norm becomes very large. If True the returned values are scaled by the square root of the normalisation of the corresponding Legendre function. This avoids the large values that occur when norm is False.

False

Returns:

Type Description tuple[float, float]

The value of the prolate spheroidal angular function and its' derivative.

Notes

This method uses the prolate spheroidal wave function code for non complex arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on github. This code is in Fortran90 and was interfaced to Python using numpy.f2py then wrapped with the current method to provide a similar calling convention as the Scipy function of the same name.

References

Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal radial functions of the first kind and their first derivatives. Quarterly of Applied Mathematics, 60(3), 589-599. https://doi.org/10.1090/qam/1914443

Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied Mathematics, 62(3), 493-507. https://doi.org/10.1090/qam/2086042

Source code in src/echosms/utils.py 
def pro_ang1(m: int, n: int, c: float, eta: float, norm=False) -> tuple[float, float]:\n    \"\"\"Prolate spheroidal angular function of the first kind and derivative.\n\n    Calculates the prolate spheroidal angular function of the first kind and its'\n    derivative with respect to `eta`.\n\n    Parameters\n    ----------\n    m :\n        The order parameter (\u2265 0)\n    n :\n        The degree parameter (\u2265 `m`).\n    c :\n        The size parameter.\n    eta :\n        The angular coordinate, \u03b7, where |\u03b7| \u2264 1.\n    norm :\n        If `False`, returned values are not normalised (i.e., the Meixner-Sch\u00e4fke normlalisation\n        scheme is used). For large `m` this norm becomes very large. If `True` the returned values\n        are scaled by the square root of the normalisation of the corresponding Legendre function.\n        This avoids the large values that occur when `norm` is `False`.\n\n    Returns\n    -------\n    :\n        The value of the prolate spheroidal angular function and its' derivative.\n\n    Notes\n    -----\n    This method uses the prolate spheroidal wave function code for non complex\n    arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on\n    [github](https://github.com/MathieuandSpheroidalWaveFunctions). This code is in Fortran90\n    and was interfaced to Python using `numpy.f2py` then wrapped with the current method to\n    provide a similar calling convention as the Scipy function of the same name.\n\n    References\n    ----------\n    Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal\n    radial functions of the first kind and their first derivatives. Quarterly of Applied\n    Mathematics, 60(3), 589-599. <https://doi.org/10.1090/qam/1914443>\n\n    Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal\n    Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied\n    Mathematics, 62(3), 493-507. <https://doi.org/10.1090/qam/2086042>\n    \"\"\"\n    if m < 0:\n        raise ValueError('The m parameter must be positive.')\n    if n < m:\n        raise ValueError('The n parameter must be greater than or equal to the m parameter.')\n    if abs(eta) > 1.0:\n        raise ValueError('The eta parameter must be less than or equal to 1')\n\n    a = prolate_swf.profcn(c=c, m=m, lnum=n-m+2, x1=0.0, ioprad=0, iopang=2,\n                           iopnorm=int(norm), arg=[eta])\n    p = swf_t._make(a)\n    s = p.s1c * np.float_power(10.0, p.is1e)\n    sp = p.s1dc * np.float_power(10.0, p.is1de)\n\n    return s[n-m][0], sp[n-m][0]\n
"},{"location":"api_reference/#echosms.utils.pro_rad1","title":"pro_rad1(m, n, c, xi)","text":"

Prolate spheroidal radial function of the first kind and derivative.

Calculates the prolate spheroidal radial function of the first kind and its' derivative with respect to xi.

Parameters:

Name Type Description Default m int

The order parameter (\u2265 0).

required n int

The degree parameter (\u2265 m).

required c float

The size parameter.

required xi float

The radial coordinate, \u03be, where \u03be \u2265 1.

required

Returns:

Type Description tuple[float, float]

The value of the prolate spheroidal radial function and its' derivative.

Notes

This method uses the prolate spheroidal wave function code for non complex arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on github. This code is in Fortran90 and was interfaced to Python using numpy.f2py then wrapped with the current method to provide a similar calling convention as the Scipy function of the same name.

References

Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal radial functions of the first kind and their first derivatives. Quarterly of Applied Mathematics, 60(3), 589-599. https://doi.org/10.1090/qam/1914443

Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied Mathematics, 62(3), 493-507. https://doi.org/10.1090/qam/2086042

Source code in src/echosms/utils.py 
def pro_rad1(m: int, n: int, c: float, xi: float) -> tuple[float, float]:\n    \"\"\"Prolate spheroidal radial function of the first kind and derivative.\n\n    Calculates the prolate spheroidal radial function of the first kind and its'\n    derivative with respect to `xi`.\n\n    Parameters\n    ----------\n    m :\n        The order parameter (\u2265 0).\n    n :\n        The degree parameter (\u2265 `m`).\n    c :\n        The size parameter.\n    xi :\n        The radial coordinate, \u03be, where \u03be \u2265 1.\n\n    Returns\n    -------\n    :\n        The value of the prolate spheroidal radial function and its' derivative.\n\n    Notes\n    -----\n    This method uses the prolate spheroidal wave function code for non complex\n    arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on\n    [github](https://github.com/MathieuandSpheroidalWaveFunctions). This code is in Fortran90\n    and was interfaced to Python using `numpy.f2py` then wrapped with the current method to\n    provide a similar calling convention as the Scipy function of the same name.\n\n    References\n    ----------\n    Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal\n    radial functions of the first kind and their first derivatives. Quarterly of Applied\n    Mathematics, 60(3), 589-599. <https://doi.org/10.1090/qam/1914443>\n\n    Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal\n    Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied\n    Mathematics, 62(3), 493-507. <https://doi.org/10.1090/qam/2086042>\n    \"\"\"\n    if m < 0:\n        raise ValueError('The m parameter must be positive.')\n    if n < m:\n        raise ValueError('The n parameter must be greater than or equal to the m parameter.')\n    if xi < 1.0:\n        raise ValueError('The xi parameter must be greater than or equal to 1')\n\n    a = prolate_swf.profcn(c=c, m=m, lnum=n-m+2, x1=xi-1.0, ioprad=1, iopang=0, iopnorm=0, arg=[0])\n    p = swf_t._make(a)\n    s = p.r1c * np.float_power(10.0, p.ir1e)\n    sp = p.r1dc * np.float_power(10.0, p.ir1de)\n\n    return s[n-m], sp[n-m]\n
"},{"location":"api_reference/#echosms.utils.pro_rad2","title":"pro_rad2(m, n, c, xi)","text":"

Prolate spheroidal radial function of the second kind and derivative.

Calculates the prolate spheroidal radial function of the second kind and its' derivative with respect to xi.

Parameters:

Name Type Description Default m int

The order parameter (\u2265 0).

required n int

The degree parameter (\u2265 m).

required c float

The size parameter.

required xi float

The radial coordinate, \u03be, where \u03be \u2265 1.

required

Returns:

Type Description tuple[float, float]

The value of the prolate spheroidal radial function and its' derivative.

Notes

This method uses the prolate spheroidal wave function code for non complex arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on github. This code is in Fortran90 and was interfaced to Python using numpy.f2py then wrapped with the current method to provide a similar calling convention as the Scipy function of the same name.

References

Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal radial functions of the first kind and their first derivatives. Quarterly of Applied Mathematics, 60(3), 589-599. https://doi.org/10.1090/qam/1914443

Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied Mathematics, 62(3), 493-507. https://doi.org/10.1090/qam/2086042

Source code in src/echosms/utils.py 
def pro_rad2(m: int, n: int, c: float, xi: float) -> tuple[float, float]:\n    \"\"\"Prolate spheroidal radial function of the second kind and derivative.\n\n    Calculates the prolate spheroidal radial function of the second kind and its'\n    derivative with respect to `xi`.\n\n    Parameters\n    ----------\n    m :\n        The order parameter (\u2265 0).\n    n :\n        The degree parameter (\u2265 `m`).\n    c :\n        The size parameter.\n    xi :\n        The radial coordinate, \u03be, where \u03be \u2265 1.\n\n    Returns\n    -------\n    :\n        The value of the prolate spheroidal radial function and its' derivative.\n\n    Notes\n    -----\n    This method uses the prolate spheroidal wave function code for non complex\n    arguments (van Buren & Boisvert, 2002, and van Buren & Boisvert, 2024), available on\n    [github](https://github.com/MathieuandSpheroidalWaveFunctions). This code is in Fortran90\n    and was interfaced to Python using `numpy.f2py` then wrapped with the current method to\n    provide a similar calling convention as the Scipy function of the same name.\n\n    References\n    ----------\n    Van Buren, A. L., & Boisvert, J. E. (2002). Accurate calculation of prolate spheroidal\n    radial functions of the first kind and their first derivatives. Quarterly of Applied\n    Mathematics, 60(3), 589-599. <https://doi.org/10.1090/qam/1914443>\n\n    Van Buren, A. L., & Boisvert, J. E. (2004). Improved Calculation of Prolate Spheroidal\n    Radial Functions of the Second Kind and Their First Derivatives. Quarterly of Applied\n    Mathematics, 62(3), 493-507. <https://doi.org/10.1090/qam/2086042>\n    \"\"\"\n    if m < 0:\n        raise ValueError('The m parameter must be positive.')\n    if n < m:\n        raise ValueError('The n parameter must be greater than or equal to the m parameter.')\n    if xi < 1.0:\n        raise ValueError('The xi parameter must be greater than or equal to 1')\n\n    ioprad = 1 if xi-1.0 < 1e-10 else 2\n\n    # Add +2 to lnum instead of +1 as it exposes a bug in the Fortran code - if n = 0, zeros\n    # are returned instead of the correct value.\n    a = prolate_swf.profcn(c=c, m=m, lnum=n-m+2, x1=xi-1.0,\n                           ioprad=ioprad, iopang=0, iopnorm=0, arg=[0])\n    p = swf_t._make(a)\n\n    if ioprad == 1:\n        s = np.inf\n        sp = np.inf\n    else:\n        s = p.r2c * np.float_power(10.0, p.ir2e)\n        sp = p.r2dc * np.float_power(10.0, p.ir2de)\n\n    return s[n-m], sp[n-m]\n
"},{"location":"api_reference/#echosms.utils.spherical_jnpp","title":"spherical_jnpp(n, z)","text":"

Second derivative of the spherical Bessel function.

Parameters:

Name Type Description Default n int

Order (n \u2265 0)

required z float

Argument of the Bessel function.

required

Returns:

Type Description float

The second derivative of the spherical Bessel function.

Source code in src/echosms/utils.py 
def spherical_jnpp(n: int, z: float) -> float:\n    \"\"\"Second derivative of the spherical Bessel function.\n\n    Parameters\n    ----------\n    n :\n        Order (n \u2265 0)\n    z :\n        Argument of the Bessel function.\n\n    Returns\n    -------\n    :\n        The second derivative of the spherical Bessel function.\n\n    \"\"\"\n    return 1./z**2 * ((n**2-n-z**2)*spherical_jn(n, z) + 2.*z*spherical_jn(n+1, z))\n
"},{"location":"api_reference/#echosms.utils.split_dict","title":"split_dict(d, s)","text":"

Split a dict into two dicts based on a list of keys.

Parameters:

Name Type Description Default d dict

Dict to be split.

required s list

List of dict keys to use for splitting d.

required

Returns:

Type Description tuple(dict, dict)

The input dict split into two dicts based on the keys in s. The first tuple item contains the items that do not have keys in s.

Source code in src/echosms/utils.py 
def split_dict(d: dict, s: list) -> tuple[dict, dict]:\n    \"\"\"Split a dict into two dicts based on a list of keys.\n\n    Parameters\n    ----------\n    d : dict\n        Dict to be split.\n\n    s: list\n        List of dict keys to use for splitting `d`.\n\n    Returns\n    -------\n    : tuple(dict, dict)\n        The `input` dict split into two dicts based on the keys in `s`. The first tuple item\n        contains the items that do not have keys in `s`.\n    \"\"\"\n    contains = {k: v for k, v in d.items() if k in s}\n    ncontains = {k: v for k, v in d.items() if k not in s}\n    return ncontains, contains\n
"},{"location":"api_reference/#echosms.utils.wavelength","title":"wavelength(c, f)","text":"

Calculate the acoustic wavelength.

Parameters:

Name Type Description Default c float

Sound speed [m/s]

required f float

Frequency [Hz]

required

Returns:

Type Description float

The acoustic wavelength [m].

Source code in src/echosms/utils.py 
def wavelength(c: float, f: float) -> float:\n    \"\"\"Calculate the acoustic wavelength.\n\n    Parameters\n    ----------\n    c :\n        Sound speed [m/s]\n\n    f :\n        Frequency [Hz]\n\n    Returns\n    -------\n    :\n        The acoustic wavelength [m].\n    \"\"\"\n    return c/f\n
"},{"location":"api_reference/#echosms.utils.wavenumber","title":"wavenumber(c, f)","text":"

Calculate the acoustic wavenumber.

Parameters:

Name Type Description Default c float

Sound speed [m/s]

required f float

Frequency [Hz]

required

Returns:

Type Description float

The acoustic wavenumber [m\u207b\u00b9].

Source code in src/echosms/utils.py 
def wavenumber(c: float, f: float) -> float:\n    \"\"\"Calculate the acoustic wavenumber.\n\n    Parameters\n    ----------\n    c :\n        Sound speed [m/s]\n\n    f :\n        Frequency [Hz]\n\n    Returns\n    -------\n    :\n        The acoustic wavenumber [m\u207b\u00b9].\n    \"\"\"\n    return 2*np.pi*f/c\n
"},{"location":"benchmark_data/","title":"Benchmark Data","text":"

The benchmark data for the acoustic scattering models in Jech et al. (2015) are included in the echoSMs package. These comprise target strength values for two sets of benchmarks:

  • Model runs over a range of frequencies
  • Model runs over a range of incident angles at a frequency of 38 kHz

These are provided as text files (see below), or as Pandas DataFrames via the BenchmarkData class.

"},{"location":"benchmark_data/#tsf","title":"TS(f)","text":"

This dataset contains target strength (TS re 1 m\u00b2 [dB]) as a function of acoustic frequency. The data file, Benchmark_Frequency_TS.csv, is formatted as comma-separated TS values given to a precision of two decimal places (i.e., 0.01 dB). The first row in the file contains column labels, indicating the model type for that column. NA represents TS values that were not computed.

The column names and descriptions are:

Column Name Description Frequency_kHz Acoustic frequency in kHz. TS values are given at 2 kHz increments from 12 to 400 kHz. Sphere_Rigid Benchmark values for the rigid sphere. Sphere_PressureRelease Benchmark values for the pressure release sphere. Sphere_Gas Benchmark values for the gas filled sphere. Sphere_WeaklyScattering Benchmark values for the weakly scattering sphere. ShellSphere_PressureRelease Benchmark values for the pressure release shelled sphere. ShellSphere_Gas Benchmark values for the gas filled shelled sphere. ShellSphere_WeaklyScattering Benchmark values for the weakly scattering shelled sphere. ProlateSpheroid_Rigid Benchmark values for the rigid prolate spheroid. Valid TS values were computed for 12-80 kHz ProlateSpheroid_PressureRelease Benchmark values for the pressure release prolate spheroid. Valid TS values were computed for 12-80 kHz. ProlateSpheroid_Gas Benchmark values for the gas filled prolate spheroid. No benchmark TS values were computed. ProlateSpheroid_WeaklyScattering Benchmark values for the weakly scattering prolate spheroid. Cylinder_Rigid Benchmark values for the rigid cylinder. Cylinder_PressureRelease Benchmark values for the pressure release cylinder. Cylinder_Gas Benchmark values for the gas filled cylinder. Cylinder_WeaklyScattering Benchmark values for the weakly scattering cylinder."},{"location":"benchmark_data/#ts-at-38-khz","title":"TS(\u03b8) at 38 kHz","text":"

This dataset contains target strength (TS re 1m\u00b2 [dB]) as a function of insonifying angle of incidence (\u03b8) for the prolate spheroid and cylinder shapes. The data file Benchmark_Angle_TS.csv is formatted a comma-separated Ts values given to a precision of two decimal places (i.e., 0.01 dB). Incidence angle is as per the echoSMs convention. The first row in the file contains column labels, indicating the model type for that column. NA represents TS values that were not computed.

The column names and descriptions are:

Column Name Description Angle_deg Angle of incidence. TS values are given at 2-degree increments from 0 to 90\u00b0. ProlateSpheroid_Rigid Benchmark values for the rigid prolate spheroid. ProlateSpheroid_PressureRelease Benchmark values for the pressure release prolate spheroid. ProlateSpheroid_Gas Benchmark values for the gas filled prolate spheroid. ProlateSpheroid_WeaklyScattering Benchmark values for the weakly scattering prolate spheroid. Cylinder_Rigid Benchmark values for the rigid cylinder. TS values for end-on (0\u00b0) incidence were not computed. Cylinder_PressureRelease Benchmark values for the pressure release cylinder. TS values for end-on (0\u00b0) incidence were not computed. Cylinder_Gas Benchmark values for the gas filled cylinder. TS values for end-on (0\u00b0) incidence were not computed. Cylinder_WeaklyScattering Benchmark values for the weakly scattering cylinder. TS values for end-on (0\u00b0) incidence were not computed."},{"location":"conventions/","title":"Conventions","text":""},{"location":"conventions/#units","title":"Units","text":"

We use SI units for the model parameters, except for angles (we use degrees instead of radians) and target strength (we use deciBels). All model code must accept inputs and produce results using the units below. If the model calculations use different units internally, the code should internally convert between them.

Parameter Units Notes length, diameter, radius, thickness, etc m density kg/m\u00b3 sound speed m/s angle \u00b0 See Coordinate systems frequency Hz target strength dB reference value is 1 m\u00b2"},{"location":"conventions/#coordinate-systems","title":"Coordinate systems","text":"

The right-handed cartesian coordinate system as defined by ISO 80000-21 is to be used, as illustrated below. The acoustic wave is defined to always travel in the positive z direction and the organism is rotated to achieve different acoustic incidence angles.

The Tait-Bryan z-y'-x'' (intrinsic) convention was chosen to represent organism rotations as it is commonly used in nautical situations. Intrinsic means that the rotations are about the axes of the rotating organism, rather than the original coordinate system. The order of rotations is z, then y, then x.

Rotations about the z-axis are yaw (\u03c8), about the y-axis are pitch (\u03b8), and about the x-axis are roll (\u0278). The definitions are such that:

  • A yaw (\u03c8) value of 0\u00b0 occurs when the organism lies along the positive x-axis (as per the illustration) and positive yaw values (as per the yellow arrow) rotate the organism's head towards the positive y-axis,
  • Pitch (\u03b8) values of 0\u00b0, 90\u00b0, and 180\u00b0 correspond to acoustic wave incidence angles of head on, dorsal, and tail on, respectively,
  • Roll (\u0278) values of \u201390\u00b0, 0\u00b0, and 90\u00b0 correspond to acoustic wave incidences onto the right (starboard), dorsal, and left (port) sides of the organism, respectively.

All model code should accept angles and produce results in this coordinate system. If the model calculations use a different coordinate system, the code should internally convert between the system given above and the version used in the code.

"},{"location":"conventions/#code-style","title":"Code style","text":"

Contributions of code should follow standardised or community-agreed styles and be provided in (or added to) a structure suitable for packaging and uploading to package libraries. For Python this includes pip and/or conda, for R this would be CRAN, for Matlab this would be a toolbox on the MATLAB File Exchange, etc.

Python code should follow PEP8 and docstrings should use PEP257 with the contents following the numpydoc style. An exception to PEP8 is made to allow lines of up to 100 characters.

  1. ISO. 2019. ISO 80000-2. Part 2: Mathematics.\u00a0\u21a9

"},{"location":"developing/","title":"Developing echoSMs","text":"

These notes are a work in progress.

This page contains notes and instructions on developing and adding new models to echoSMs.

"},{"location":"developing/#obtaining-the-source-code","title":"Obtaining the source code","text":"

The echoSMs source code is kept on github under an ICES account. Clone the repository with this URL:

https://github.com/ices-tools-dev/echoSMs.git\n
"},{"location":"developing/#generating-packages-for-pypi","title":"Generating packages for PyPI","text":"

EchoSMs is a pure Python package. The build configuration is done via a pyproject.toml file and hatchling is used to produce packages.

A github action in the echoSMS repository will generate a Python wheel and source package and upload these to PyPI. This action is triggered whenever a tagged commit occurs to the repository. The tag is used as the new version number. EchoSMs version numbers follow the semantic versioning convention.

Every commit to the echoSMs repository will generate a development package being uploaded to TestPyPI. This is used to always check that a commit does not prevent production of a package and is where a package containing the latest commit can be obtained.

"},{"location":"developing/#documentation","title":"Documentation","text":"

The echoSMs documentation is produced using mkdocs and mkdocstrings. The documentation pages are hosted by github and are regenerated after every commit to the repository using a github action.

Documentation edits can be tested locally by running:

mkdocs serve\n

in the top level of the echoSMs repository. The documentation is then available at http://127.0.0.1:8000.

"},{"location":"developing/#tests","title":"Tests","text":"

EchoSMs uses the pytest testing framework. After installing pytest, run the tests using

pytest -v\n

in the top level of the echoSMs repository.

"},{"location":"developing/#adding-a-new-scattering-model","title":"Adding a new scattering model","text":"

TBD.

"},{"location":"other_software/","title":"Other software","text":"

Other software that provides source code for acoustic scattering models of relevance to fisheries and plankton acoustics includes:

  • acousticTS: R code for calculating scattering using the DCM, DWBA, SDWBA, SDWBA_curved, KRM, MSS model, as well as that of calibration spheres.
  • Coupled BEM acoustic: Julia code that calculates the TS of three-dimensional shapes with an included object (e.g., a swimbladder).
  • scatmod: Open source acoustic scattering models for fisheries acoustics. Python and R code for fluid spheres.
  • FishAcoustics: Contains a Python module that implements the phase-tracking DWBA model.
  • KRM Model: A web page that uses the KRM model to estimate the TS of predefined or user-supplied shapes over a range of input parameters.
  • KRMr: KRM model for fish in R.
  • Liquid spheroid: Julia and C++ code to calculate the scattering by fluid prolate and oblate spheroids.
  • SDWBA Model: A web page that uses the SDWBA model to estimate the TS of predefined shapes over a range of input parameters.
  • SDWBA_TS: Matlab code that implements the SDWBA model for Antarctic krill.
  • sphereTS: Python code to calculate the TS of calibration spheres.
  • Standard sphere target strength calculator: A web page that calculates the TS of calibration spheres.
  • tetrascatt: R and C++ code that implements the DWBA model on arbitrary geometries.
  • ZooScatR: R code that implements the DWBA model.
"},{"location":"usage/","title":"Using echoSMs","text":"

EchoSMs is (currently) a Python package that implements acoustic scattering models. Each different model is a separate Python class in the echoSMs package. They all inherit from a common base class that defines a common calling convention.

"},{"location":"usage/#installation","title":"Installation","text":"

EchoSMs is available on PyPi as echosms. Install it with:

pip install echosms\n

The prolate spheroidal modal series model in echoSMs uses spheroidal wave functions. A high-accuracy implementation of these is available in the Python package spheroidalwavefunctions, as the versions provided by scipy are insufficient. This should be installed automatically when you install echosms, but note that spheroidalwavefunctions is currently only available for Linux and Windows on x86_64 CPU architectures (create an issue if you want it on a system that is not currently supported).

"},{"location":"usage/#versions","title":"Versions","text":"

The installed version of echosms can be printed with this code:

import importlib\nprint(importlib.metadata.version('echosms'))\n

The changelogs for echoSMs are available here. The latest version is always at the top of that list.

"},{"location":"usage/#model-overview","title":"Model overview","text":"

Currently, the following models are available in echoSMs:

Model type Python class name Description Deformed cylinder model DCMModel Truncated cylinders with various boundary conditions Elastic sphere ESModel Elastic spheres, such as echosounder calibration spheres Kirchhoff approximation KAModel Surfaces that are mainly convex Modal series solution MSSModel Spheres with various boundary conditions, including shells Prolate spheroidal modal series PSMSModel Prolate spheroids with various boundary conditions Phase-tracking distorted wave Born approximation PTDWBAModel Weakly scattering objects of any shape with inhomogeneous interiors

Future models will include more types of DWBA models, the Kirchhoff-ray-mode model, the Fourier matching method, and potentially finite element and boundary element models.

"},{"location":"usage/#running-a-model","title":"Running a model","text":"

Each echoSMs model expects input parameters that define the model (e.g., size, shape, material properties, etc). These can be provided in three ways:

  • A Python dictionary with an entry for each parameter,
  • A Pandas DataFrame with columns for each parameter and a row for each model run,
  • An Xarray DataArray with as many dimensions as parameters. The parameter values are in the DataArray coordinate variables.

To use a model, you need to know what parameters it requires. These are documented in the calculate_ts_single function that each model has (refer to the echoSMS API reference for details). The units for numerical parameters will always follow the echoSMs unit convention. For example, the MSSModel, when simulating the scattering from a pressure release sphere, needs the following parameters:

Name Description medium_c Sound speed in the fluid medium surrounding the target [m/s] medium_rho Density of the fluid medium surrounding the target [kg/m\u00b3] a Radius of the spherical target [m] f Frequency to calculate the scattering at [Hz] boundary_type The boundary type. Supported types are fixed rigid, pressure release, and fluid filled

The simplest way to provide these to the model is a dictionary:

    p = {'medium_rho': 1026.8,\n         'medium_c': 1477.4,\n         'a': 0.01, \n         'boundary_type': 'pressure release',\n         'f': 38000}\n

An instance of the model can then be created and the calculate_ts function called with these parameters:

    from echosms import MSSModel\n    model = MSSModel()\n    model.calculate_ts(p)\n

This will return one TS value corresponding to the parameters given. If you want to run the model for a range of parameters, the relevant dictionary items can contain multiple values:

        import numpy as np\n        p = {'medium_rho': 1026.8,\n             'medium_c': 1477.4,\n             'a': 0.01,\n             'boundary_type': 'pressure release',\n             'f', np.arange(10, 100, 1)*1000}  # [Hz]\n        model.calculate_ts(p)\n

It is also fine to have multiple items with multiple values:

        p = {'medium_rho': 1026.8,\n             'medium_c': 1477.4,\n             'a': np.arange(0.01, 0.02, 0.001),  # [m]\n             'boundary_type': ['pressure release', 'fixed rigid'],\n             'f': np.arange(10, 100, 1)*1000}  # [Hz]\n        model.calculate_ts(p)\n

The TS will be calculated for all combinations of the parameters. To do this, echoSMs expands the parameters into a Pandas DataFrame with one column for each parameter and one row for each of the combinations. It then runs the model on each row of the DataFrame. That DataFrame, with the TS included, can be returned instead of a list of TS values by using the expand option:

        model.calculate_ts(p, expand=True)\n

An introductory Jupyter notebook is available that covers the above concepts and a Python script that covers this and more is available here.

"},{"location":"usage/#using-dataframes-and-dataarrays-directly","title":"Using DataFrames and DataArrays directly","text":"

Instead of passing a dictionary to the calculate_ts function, a DataFrame or DataArray can be passed instead. The crucial aspect is that the DataFrame columns must have the same names as the parameters that the model requires. For a DataArray, the coordinate dimensions must have the same names as the model parameters.

EchoSMS provides two utility functions (as_dataframe, and as_dataarray) to convert from a dictionary representation of model parameters to a DataFrame or DataArray, or you can construct your own, or modify those returned by the as_dataframe and as_dataarray functions.

The benefit of using a DataFrame is that you have fine control over what model runs will happen - it doesn't have to be the full set of combinations of input parameters. The benefit of using a DataArray is that it is easy to extract subsets of the results for further analysis and plotting.

For a DataFrame, the number of model runs will be the number of rows in the DataFrame. For a DataArray the number of models run will be the size of the DataArray (e.g., DataArray.size())

When passing a DataFrame to a model, you can choose whether the TS results are returned as a Series or are added to the existing DataFrame (in a column called ts). Use the inplace = True parameter in the call to calculate_ts for this. When passing a DataArray to a model, the TS results are always returned in the data part of the passed in DataArray.

"},{"location":"usage/#more-complex-model-parameters","title":"More complex model parameters","text":"

Some models require parameters for which it is not sensible to duplicate them across rows in a DataFrame or as a dimension in a DataArray (e.g., the data that specifies the three-dimensional shape of a fish swimbladder). EchoSMs allows for this with the concept of non-expandable parameters - these are not expanded into DataFrame columns or DataArray dimensions and are available from the models no_expand_parameters attribute.

But, as it is very convenient to have all the model parameters in one data structure, echoSMs will store the non-expandable parameters as DataFrame or DataArray attributes. Due to a bug in the DataFrame implementation, the parameters are stored as a nested dictionary under a parameters attribute. An example of this is the PTDWBAModel:

    from echosms import PTDWBAModel, as_dataframe\n    import numpy as np\n\n    model = PTDWBAModel()\n    m = {'volume': np.full((5,5,5), 0),\n         'f': np.arange(10, 100, 1)*1000,\n         'rho': [1024, 1025],  \n         'c': [1500, 1501],\n         'voxel_size': (0.001, 0.001, 0.001),\n         'theta': 90,\n         'phi': 0}\n    m['volume'][3,3,3] = 1  # something to produce scatter\n    p = as_dataframe(m, model.no_expand_parameters)\n    model.calculate_ts(p, inplace=True)\n    print(p)\n

For the PTDWBA model, only theta and phi are expandable, so p contains just two columns. The remaining parameters are available via:

    p.attrs['parameters']\n

Note that while rho and c look like parameters that would be expanded, they are in the list of non-expandable parameters, so are not expanded. This is because the structure of the PTDWBA model means that it it not sensible to have variable parameters for rho and c.

If you pass the dictionary form of the parameters to a model, this treatment of non-expanding parameters is done automatically, where

    model.calculate_ts(m, expand=True)\n

returns the same results as

    p = as_dataframe(m, model.no_expand_parameters)\n    model.calculate_ts(p, inplace=True)`\n    print(p)\n
"},{"location":"usage/#multiprocessing","title":"Multiprocessing","text":"

This is an experimental feature.

The multiprocess = True parameter in the call to calculate_ts will cause echoSMs to divide the requested model runs over as many cores as your computer has. Total solution time will decrease almost linearly with the number of models runs.

"},{"location":"usage/#reference-model-definitions","title":"Reference model definitions","text":"

Jech et al., (2015) presented reference models for a range of scattering objects: spheres, spherical shells, prolate spheroids, and finite cylinders for several boundary conditions (fixed rigid, pressure release, fluid-filled) and parameters (backscatter as a function of frequency and incident angle). These model definitions are included in echoSMs via the ReferenceModels class, along with other objects, such as calibration spheres. For example, the names of all the model definitions are available with:

    from echosms import ReferenceModels\n    rm = ReferenceModels()\n    rm.names()\n

which returns:

['fixed rigid sphere',\n 'pressure release sphere',\n 'gas filled sphere',\n 'weakly scattering sphere',\n 'spherical fluid shell with pressure release interior',\n 'spherical fluid shell with gas interior',\n 'spherical fluid shell with weakly scattering interior',\n 'fixed rigid prolate spheroid',\n 'pressure release prolate spheroid',\n 'gas filled prolate spheroid',\n 'weakly scattering prolate spheroid',\n 'fixed rigid finite cylinder',\n 'pressure release finite cylinder',\n 'gas filled finite cylinder',\n 'weakly scattering finite cylinder',\n 'WC20 calibration sphere',\n 'WC21 calibration sphere',\n 'WC22 calibration sphere',\n 'WC25 calibration sphere',\n 'WC38.1 calibration sphere',\n 'WC57.2 calibration sphere',\n 'WC60 calibration sphere',\n 'WC64 calibration sphere',\n 'Cu13.7 calibration sphere',\n 'Cu23 calibration sphere',\n 'Cu32 calibration sphere',\n 'Cu42 calibration sphere',\n 'Cu45 calibration sphere',\n 'Cu60 calibration sphere',\n 'Cu63 calibration sphere',\n 'Cu64 calibration sphere']\n

and the specification for a particular model is given by:

    rm.specification('spherical fluid shell with weakly scattering interior')\n

which returns:

{'name': 'spherical fluid shell with weakly scattering interior',\n 'shape': 'sphere',\n 'boundary_type': 'fluid shell fluid interior',\n 'description': 'A fluid spherical shell with a weakly scattering shell and interior',\n 'a': 0.01,\n 'shell_thickness': 0.001,\n 'medium_rho': 1026.8,\n 'medium_c': 1477.4,\n 'shell_rho': 1028.9,\n 'shell_c': 1480.3,\n 'target_rho': 1031.0,\n 'target_c': 1483.3,\n 'source': 'https://doi.org/10.1121/1.4937607',\n 'benchmark_model': 'mss'}\n

Note that the specification contains more information that the model itself needs, so the subset needed for running a model is available via:

    m = rm.parameters('spherical fluid shell with weakly scattering interior')\n    print(m)\n

which returns:

{'boundary_type': 'fluid shell fluid interior',\n 'a': 0.01,\n 'shell_thickness': 0.001,\n 'medium_rho': 1026.8,\n 'medium_c': 1477.4,\n 'shell_rho': 1028.9,\n 'shell_c': 1480.3,\n 'target_rho': 1031.0,\n 'target_c': 1483.3}\n

Note that the parameters() call does not return all of the parameters needed by a model. For example, f is not there and needs to be added before running a model:

    m['f'] = [38000, 40000, 42000]\n\n    from echosms import MSSModel\n    model = MSSModel()\n    model.calculate_ts(m)\n
"},{"location":"usage/#benchmark-model-ts","title":"Benchmark model TS","text":"

Jech et al., (2015) presented benchmark model runs for the reference models. The TS results from these benchmarks are available in echoSMs via the BenchMarkData class. This class is a simple wrapper around code that reads the CSV-formatted file of benchmark values into a Pandas DataFrame, whereupon they can be readily accessed like this:

    from echosms import BenchmarkData\n    bm = BenchmarkData()\n    bm.angle_dataset  # the TS as a function of angle at 38 kHz\n    bm.freq_dataset  # the TS as a function of frequency\n

The TS and frequency values for a particular benchmark are available with normal Pandas DataFrame indexing syntax. The DataFrame columns names as the same as the ReferenceModels names. For example:

    bm.freq_dataset['weakly scattering sphere']\n    bm.freq_dataset['frequency (kHz)']\n

or for the angle dataset:

    bm.angle_dataset['weakly scattering sphere']\n    bm.angle_dataset['angle (deg)']\n
"}]} \ No newline at end of file diff --git a/sitemap.xml b/sitemap.xml new file mode 100644 index 0000000..0584a98 --- /dev/null +++ b/sitemap.xml @@ -0,0 +1,31 @@ + + + + https://ices-tools-dev.github.io/echoSMs/ + 2024-10-11 + + + https://ices-tools-dev.github.io/echoSMs/api_reference/ + 2024-10-11 + + + https://ices-tools-dev.github.io/echoSMs/benchmark_data/ + 2024-10-11 + + + https://ices-tools-dev.github.io/echoSMs/conventions/ + 2024-10-11 + + + https://ices-tools-dev.github.io/echoSMs/developing/ + 2024-10-11 + + + https://ices-tools-dev.github.io/echoSMs/other_software/ + 2024-10-11 + + + https://ices-tools-dev.github.io/echoSMs/usage/ + 2024-10-11 + + \ No newline at end of file diff --git a/sitemap.xml.gz b/sitemap.xml.gz new file mode 100644 index 0000000..3615819 Binary files /dev/null and b/sitemap.xml.gz differ diff --git a/src/make_coordinate_system_figure.py b/src/make_coordinate_system_figure.py new file mode 100644 index 0000000..12f12cb --- /dev/null +++ b/src/make_coordinate_system_figure.py @@ -0,0 +1,193 @@ +# %% +"""Create the coordinate system figure for the documentation.""" +import pyvista as pv +import numpy as np +from pathlib import Path +import xml.etree.ElementTree as ET + +# These two functions came from https://github.com/pyvista/pyvista/discussions/5023 and are +# used to create arced arrows. + + +def rotate_to(a, b): + """Return a rotation matrix taking the vector A to B.""" + # Naive approach is unstable if a and b are near parallel + theta = np.arccos(np.dot(a, b)) + eps = 1e-3 + if np.absolute(theta) < eps: + return np.eye(4) + elif np.absolute(np.pi - theta) < eps: # Close to 180 degrees + # Choose the coordinate axis most orthogonal to A + x = np.zeros(3) + x[np.argmin(np.absolute(a))] = 1.0 + axis = np.cross(a, x) + axis /= np.linalg.norm(axis) + return pv.utilities.transformations.axis_angle_rotation(axis, theta, deg=False) + else: + axis = np.cross(a, b) + return pv.utilities.transformations.axis_angle_rotation(axis, theta, deg=False) + + +def semi_circular_arrow( + start_angle=0, + circ_frac=0.75, + body_axial_res=100, + body_radial_res=50, + head_radial_res=100, + circ_radius=10, + body_radius=1, + head_length=3, + head_radius_frac=1.5, + normal=None, + center=None): + """Create a semi circular arrow.""" + t = np.linspace(0, circ_frac * 2 * np.pi, body_axial_res) + start_angle + x = circ_radius * np.cos(t) + y = circ_radius * np.sin(t) + z = np.zeros(body_axial_res) + body_pts = np.column_stack([x, y, z]) + body = pv.MultipleLines(body_pts).tube(body_radius, n_sides=body_radial_res) + + # Direction the head points + dhead = body_pts[-1] - body_pts[-2] + dhead /= np.linalg.norm(dhead) + + head = pv.Cone( + center=body_pts[-1] + dhead * (head_length / 2), + direction=dhead, + height=head_length, + radius=head_radius_frac * body_radius, + resolution=head_radial_res, + ) + + arrow = body.merge(head, merge_points=False) + + if normal is not None: + arrow = arrow.transform(rotate_to((0, 0, 1), normal), inplace=False) + + if center is not None: + arrow = arrow.translate(center, inplace=False) + + return arrow + +# %% +# This code generates the figure + + +resourcesDir = Path(r'../resources') +al = 10.0 # axes length +sr = 0.01 # shaft radius for axes and arrows + +# Axes arrows +arrow_x = pv.Arrow(start=(0, 0, 0), direction=(al, 0, 0), shaft_radius=sr, tip_radius=0.02, + tip_length=0.1, scale='auto') +arrow_y = pv.Arrow(start=(0, 0, 0), direction=(0, al, 0), shaft_radius=sr, tip_radius=0.02, + tip_length=0.1, scale='auto') +arrow_z = pv.Arrow(start=(0, 0, 0), direction=(0, 0, al), shaft_radius=sr, tip_radius=0.02, + tip_length=0.1, scale='auto') + +# Angle arrows and arcs +circ_frac = 0.8 +along_axis = 0.7 +arc_pitch = semi_circular_arrow(center=(0, al*along_axis, 0), circ_frac=circ_frac, + start_angle=-0, + circ_radius=al/10, normal=(0, 1, 0), + body_radius=0.05, head_length=0.5) +arc_roll = semi_circular_arrow(center=(al*along_axis, 0, 0), circ_frac=circ_frac, + start_angle=0, + circ_radius=al/10, normal=(1, 0, 0), + body_radius=0.05, head_length=0.5) +arc_yaw = semi_circular_arrow(center=(0, 0, al*along_axis), circ_frac=circ_frac, + start_angle=0, + circ_radius=al/10, normal=(0, 0, 1), + body_radius=0.05, head_length=0.5) + +# axes labels +axes_label_pts = [[al*1.02, 0., 0.], [0., al*1.05, 0.], [0., 0., al*1.07]] +axes_label_txt = ['x', 'y', 'z'] + +# angle labels +angles_label_pts = [[al*along_axis, 0., -al/10*1.1], + [al/10*1.1, al*along_axis, 0.], + [al/10*1.1, 0., al*along_axis]] +angles_label_txt = ['φ', 'θ', 'ψ'] + +# the fish model came from: https://3dmag.org/en/market/download/item/6255/ +reader = pv.get_reader('../resources/herring.stl') +fish = reader.read() + +# centre the fish on the origin +b = fish.bounds +offset = (-((b[1]-b[0])/2 + b[0]), (b[3]-b[2])/2 + b[2], (b[5]-b[4])/2 + b[4]) +fish.translate(offset, inplace=True) + +# rotate the fish to fit our coordinate system +fish.rotate_x(180, inplace=True) +fish.rotate_z(-90, inplace=True) + +# scale the fish to length (along the z axis) +length = fish.bounds[5] - fish.bounds[4] +fish.scale(0.3*al/length, inplace=True) + +# Assemble the 3D scene +for t in [pv.themes.DocumentTheme(), pv.themes.DarkTheme()]: + + text_colour = 'black' + if t.name == 'dark': + text_colour = pv.Color([190, 193, 198]) # to match text color in the docs dark theme + t.background = pv.Color([20, 33, 41]) # to match the background in the docs dark theme + + p = pv.Plotter(window_size=[1600, 860], theme=t, off_screen=True) + p.add_mesh(fish, opacity=.9) + p.add_mesh(arrow_x, color='gray') + p.add_mesh(arrow_y, color='gray') + p.add_mesh(arrow_z, color='gray') + p.add_mesh(arc_pitch, color='green') + p.add_mesh(arc_roll, color='red') + p.add_mesh(arc_yaw, color='yellow') + + # the angle labels + p.add_point_labels(angles_label_pts, angles_label_txt, font_family='times', italic=True, + bold=False, shape=None, always_visible=True, show_points=False, + font_size=50, text_color=text_colour) + # the axes labels + p.add_point_labels(axes_label_pts, axes_label_txt, + font_size=50, italic=True, bold=False, shape=None, + always_visible=True, show_points=False, font_family='times', + text_color=text_colour) + + p.camera_position = [(9.44840097372024, 17.277196718053595, -7.056312001523225), + (2.7462545037450483, 2.6898350612751827, 1.695119545588695), + (-0.2125924945535939, -0.42901864969164194, -0.877922222908294)] + + # Unfortunately, all exports have issues... + # p.export_html('coordinate_system2.html') # loses all text + # p.export_obj('coordinate_system2.obj') # no colour? + # p.export_gltf('coordinate_system2.gltf') # loses text + # p.export_vrml('coordinate_system2.vrml') # no text?? + # p.export_vtksz(resourcesDir/'coordinate_system.vtksz') + + # no greek symbols and scale > 1 loses some of the text + # p.screenshot('coordinate_system2.png', transparent_background=True, scale=1) + + # Best option for the moment. Html would be preferrable if the text showed up + if t.name == 'dark': + savefile = resourcesDir/'coordinate_system_dark.svg' + else: + savefile = resourcesDir/'coordinate_system_light.svg' + + p.save_graphic(savefile) # raster, otherwise good + + # this generates an on-screen version. But doesn't show the greek symbols + # p.show() + + # p.close() + + # Modify the generate svg to make the labels properly italic + tree = ET.parse(savefile) + root = tree.getroot() + + for text in root.iter('{http://www.w3.org/2000/svg}text'): + text.set('font-style', 'italic') + + tree.write(savefile) diff --git a/src/make_coordinate_system_figure_vedo.py b/src/make_coordinate_system_figure_vedo.py new file mode 100644 index 0000000..58b8e07 --- /dev/null +++ b/src/make_coordinate_system_figure_vedo.py @@ -0,0 +1,71 @@ +# %% +"""Create the coordinate system figure for the documentation.""" + +# Try with vedo to perhaps get a better 3D html output than pyvista. + +import vedo as vd +import numpy as np +from pathlib import Path + +# This code generates the figure + +resourcesDir = Path(r'../resources') +al = 10.0 # axes length +sr = 0.01 # shaft radius for axes and arrows + + +# the fish model came from: https://3dmag.org/en/market/download/item/6255/ +fish = vd.Mesh(str(resourcesDir / 'herring.stl')) +fish.color('green7', alpha=0.9) + +# centre the fish on the origin, rotate, and scale +b = fish.bounds() +offset = [-((b[1]-b[0])/2 + b[0]), (b[3]-b[2])/2 + b[2], (b[5]-b[4])/2 + b[4]] +scale = 0.3*al/(b[5] - b[4]) +fish.apply_transform(vd.LinearTransform().translate(offset).rotate_x(180).rotate_z(-90)) +fish.apply_transform(vd.LinearTransform().scale(scale)) + +# Axes arrows +axes = vd.Arrows(start_pts=[[0, 0, 0], + [0, 0, 0], + [0, 0, 0]], + end_pts=[[al, 0, 0], + [0, al, 0], + [0, 0, al]], + shaft_radius=sr, head_radius=0.02, head_length=0.1, res=15) + +# Axes labels +axes_label_pts = [[al*1.02, 0., 0.], [0., al*1.05, 0.], [0., 0., al*1.07]] +axes_label_txt = ['x', 'y', 'z'] + + +# Rotation arrows +circ_frac = 0.8 +along_axis = 0.7 +radius = al/10 +tube_radius = 0.05 + +theta = np.arange(0, circ_frac*2*np.pi, 2*np.pi/100) +x = radius * np.cos(theta) +y = radius * np.sin(theta) +z = np.full(x.size, along_axis*al) + +rot_line_x = vd.Line(list(zip(z, y, x))) +rot_line_y = vd.Line(list(zip(x, z, y))) +rot_line_z = vd.Line(list(zip(x, y, z))) + +rot_arrow_x = vd.Tube(rot_line_x, r=tube_radius, c='red') +rot_arrow_y = vd.Tube(rot_line_y, r=tube_radius, c='green') +rot_arrow_z = vd.Tube(rot_line_z, r=tube_radius, c='yellow') + +# need to add arrow heads to these lines + +# Rotation labels + + +# Camera position + +plt = vd.Plotter(size=(800, 600), bg='white') +plt.show(fish, axes, rot_arrow_x, rot_arrow_y, rot_arrow_z) + +plt.export(str(resourcesDir/'coordinate_system.html')) diff --git a/tutorial.ipynb b/tutorial.ipynb new file mode 100644 index 0000000..97f4823 --- /dev/null +++ b/tutorial.ipynb @@ -0,0 +1 @@ +{"cells":[{"cell_type":"markdown","metadata":{},"source":["## Tutorial \n","\n"," \"Open\n","\n","\n","This notebook provides an introductory tutorial for echoSMs."]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"collapsed":true,"executionInfo":{"elapsed":10231,"status":"ok","timestamp":1724374592133,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"vvvd0_LuowDN","outputId":"6ebbda45-32d1-4c4b-9ea2-85dbf1326efd"},"outputs":[],"source":["!pip install matplotlib echosms"]},{"cell_type":"markdown","metadata":{"id":"K9izbBdpuj30"},"source":["## Imports\n","\n","We import the modal series solution model from echoSMs and the benchmark data and reference models."]},{"cell_type":"code","execution_count":2,"metadata":{"executionInfo":{"elapsed":2521,"status":"ok","timestamp":1724374603926,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"yzcGKsBuo2Hj"},"outputs":[],"source":["from echosms import MSSModel, BenchmarkData, ReferenceModels\n","import matplotlib.pyplot as plt\n","import numpy as np"]},{"cell_type":"markdown","metadata":{"id":"V1uNxsWfraXn"},"source":["## Reference models\n","The reference models in the Jech et al (2015) paper are available in the echoSMs package:"]},{"cell_type":"code","execution_count":3,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":2,"status":"ok","timestamp":1724374610285,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"o2J4jrvvo_OL","outputId":"1506a818-16c5-4892-ac2d-311bd4f69eb2"},"outputs":[{"name":"stdout","output_type":"stream","text":["Available reference models are:\n","\n","fixed rigid sphere\n","pressure release sphere\n","gas filled sphere\n","weakly scattering sphere\n","spherical fluid shell with pressure release interior\n","spherical fluid shell with gas interior\n","spherical fluid shell with weakly scattering interior\n","fixed rigid prolate spheroid\n","pressure release prolate spheroid\n","gas filled prolate spheroid\n","weakly scattering prolate spheroid\n","fixed rigid finite cylinder\n","pressure release finite cylinder\n","gas filled finite cylinder\n","weakly scattering finite cylinder\n","WC38.1 calibration sphere\n","Cu60 calibration sphere\n"]}],"source":["rm = ReferenceModels()\n","print('Available reference models are:\\n')\n","print('\\n'.join(rm.names()))"]},{"cell_type":"markdown","metadata":{"id":"uz4JhAIArmZA"},"source":["## Benchmark results\n","Likewise, the results from the benchmark model runs in the Jech et al (2015) paper are available in the echoSMs package."]},{"cell_type":"code","execution_count":4,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":444},"executionInfo":{"elapsed":6,"status":"ok","timestamp":1724374618002,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"HV8YqRUDpDfF","outputId":"4e851e6c-2280-4f4a-cc5a-41052994e26b"},"outputs":[{"data":{"text/html":["
\n","\n","\n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n"," \n","
Frequency_kHzSphere_RigidSphere_PressureReleaseSphere_GasSphere_WeaklyScatteringShellSphere_PressureReleaseShellSphere_GasShellSphere_WeaklyScatteringProlateSpheroid_RigidProlateSpheroid_PressureReleaseProlateSpheroid_GasProlateSpheroid_WeaklyScatteringCylinder_RigidCylinder_PressureReleaseCylinder_GasCylinder_WeaklyScattering
012-54.44-42.29-42.34-103.95-42.83-42.80-99.15-35.98-30.16NaN-87.05-38.75-35.29-35.30-89.79
114-52.20-42.92-42.93-101.62-43.40-43.44-96.79-33.83-30.02NaN-84.71-36.83-34.93-34.93-87.54
216-50.40-43.52-43.52-99.69-43.96-43.99-94.83-32.20-29.87NaN-82.78-35.45-34.56-34.56-85.73
318-48.96-44.02-44.03-98.10-44.49-44.51-93.20-30.97-29.70NaN-81.19-34.51-34.17-34.18-84.27
420-47.85-44.39-44.40-96.79-44.94-44.97-91.85-30.08-29.54NaN-79.87-33.95-33.80-33.81-83.12
...................................................
190392-45.96-46.00-46.03-103.61-46.91-46.88-108.29NaNNaNNaN-86.94-21.74-21.86-21.92-70.78
191394-45.86-46.00-45.96-100.25-46.91-46.78-107.19NaNNaNNaN-83.45-21.72-21.84-21.83-70.14
192396-45.80-46.00-46.07-98.03-46.91-46.75-106.49NaNNaNNaN-81.16-21.70-21.82-21.82-69.78
193398-45.80-46.00-45.96-96.47-46.91-46.99-106.14NaNNaNNaN-79.55-21.68-21.80-21.79-69.68
194400-45.84-46.00-46.05-95.37-46.91-46.94-106.15NaNNaNNaN-78.41-21.65-21.77-21.79-69.83
\n","

195 rows × 16 columns

\n","
"],"text/plain":[" Frequency_kHz Sphere_Rigid Sphere_PressureRelease Sphere_Gas \\\n","0 12 -54.44 -42.29 -42.34 \n","1 14 -52.20 -42.92 -42.93 \n","2 16 -50.40 -43.52 -43.52 \n","3 18 -48.96 -44.02 -44.03 \n","4 20 -47.85 -44.39 -44.40 \n",".. ... ... ... ... \n","190 392 -45.96 -46.00 -46.03 \n","191 394 -45.86 -46.00 -45.96 \n","192 396 -45.80 -46.00 -46.07 \n","193 398 -45.80 -46.00 -45.96 \n","194 400 -45.84 -46.00 -46.05 \n","\n"," Sphere_WeaklyScattering ShellSphere_PressureRelease ShellSphere_Gas \\\n","0 -103.95 -42.83 -42.80 \n","1 -101.62 -43.40 -43.44 \n","2 -99.69 -43.96 -43.99 \n","3 -98.10 -44.49 -44.51 \n","4 -96.79 -44.94 -44.97 \n",".. ... ... ... \n","190 -103.61 -46.91 -46.88 \n","191 -100.25 -46.91 -46.78 \n","192 -98.03 -46.91 -46.75 \n","193 -96.47 -46.91 -46.99 \n","194 -95.37 -46.91 -46.94 \n","\n"," ShellSphere_WeaklyScattering ProlateSpheroid_Rigid \\\n","0 -99.15 -35.98 \n","1 -96.79 -33.83 \n","2 -94.83 -32.20 \n","3 -93.20 -30.97 \n","4 -91.85 -30.08 \n",".. ... ... \n","190 -108.29 NaN \n","191 -107.19 NaN \n","192 -106.49 NaN \n","193 -106.14 NaN \n","194 -106.15 NaN \n","\n"," ProlateSpheroid_PressureRelease ProlateSpheroid_Gas \\\n","0 -30.16 NaN \n","1 -30.02 NaN \n","2 -29.87 NaN \n","3 -29.70 NaN \n","4 -29.54 NaN \n",".. ... ... \n","190 NaN NaN \n","191 NaN NaN \n","192 NaN NaN \n","193 NaN NaN \n","194 NaN NaN \n","\n"," ProlateSpheroid_WeaklyScattering Cylinder_Rigid \\\n","0 -87.05 -38.75 \n","1 -84.71 -36.83 \n","2 -82.78 -35.45 \n","3 -81.19 -34.51 \n","4 -79.87 -33.95 \n",".. ... ... \n","190 -86.94 -21.74 \n","191 -83.45 -21.72 \n","192 -81.16 -21.70 \n","193 -79.55 -21.68 \n","194 -78.41 -21.65 \n","\n"," Cylinder_PressureRelease Cylinder_Gas Cylinder_WeaklyScattering \n","0 -35.29 -35.30 -89.79 \n","1 -34.93 -34.93 -87.54 \n","2 -34.56 -34.56 -85.73 \n","3 -34.17 -34.18 -84.27 \n","4 -33.80 -33.81 -83.12 \n",".. ... ... ... \n","190 -21.86 -21.92 -70.78 \n","191 -21.84 -21.83 -70.14 \n","192 -21.82 -21.82 -69.78 \n","193 -21.80 -21.79 -69.68 \n","194 -21.77 -21.79 -69.83 \n","\n","[195 rows x 16 columns]"]},"execution_count":4,"metadata":{},"output_type":"execute_result"}],"source":["bm = BenchmarkData()\n","bmf = bm.freq_dataset # this is a Pandas DataFrame\n","bmf"]},{"cell_type":"markdown","metadata":{"id":"SreX_FuOr9US"},"source":["## Creating the model parameters\n","\n","We can now get the model parameters and results for a given model in the Jech et al (2015) paper and run the same model using the echoSMs package and compare them. First step is to get the model parameters for a model - we choose the weakly scattering sphere for this example:"]},{"cell_type":"code","execution_count":5,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":364,"status":"ok","timestamp":1724374636656,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"jynjQr3lpIBp","outputId":"c0307411-2e0c-48d0-bdb9-8b052e5e8239"},"outputs":[{"data":{"text/plain":["{'boundary_type': 'fluid filled',\n"," 'a': 0.01,\n"," 'medium_rho': 1026.8,\n"," 'medium_c': 1477.4,\n"," 'target_rho': 1028.9,\n"," 'target_c': 1480.3}"]},"execution_count":5,"metadata":{},"output_type":"execute_result"}],"source":["\n","m = rm.parameters('weakly scattering sphere')\n","m"]},{"cell_type":"markdown","metadata":{"id":"wUh0d2QTrUpq"},"source":["These parameters need to have an angle and frequency range added. We will use the frequencies from the Jech et al (2015) paper to make comparisons simplier."]},{"cell_type":"code","execution_count":6,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":571,"status":"ok","timestamp":1724373610677,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"RTXejojmpQqC","outputId":"1130042a-0942-4d05-c3aa-80429019f1b6"},"outputs":[{"data":{"text/plain":["{'boundary_type': 'fluid filled',\n"," 'a': 0.01,\n"," 'medium_rho': 1026.8,\n"," 'medium_c': 1477.4,\n"," 'target_rho': 1028.9,\n"," 'target_c': 1480.3,\n"," 'f': 0 12000.0\n"," 1 14000.0\n"," 2 16000.0\n"," 3 18000.0\n"," 4 20000.0\n"," ... \n"," 190 392000.0\n"," 191 394000.0\n"," 192 396000.0\n"," 193 398000.0\n"," 194 400000.0\n"," Name: Frequency_kHz, Length: 195, dtype: float64,\n"," 'theta': 90}"]},"execution_count":6,"metadata":{},"output_type":"execute_result"}],"source":["m['f'] = bm.freq_dataset['Frequency_kHz']*1e3\n","m['theta'] = 90\n","m"]},{"cell_type":"markdown","metadata":{"id":"FL-vhBY2qStT"},"source":["## Calculating target strength\n","\n","The reference model for a weakly scattering sphere was the model series solution, so we create an instance of that model in echoSMs and get it to calculate the target strength as per the parameters in ``m``.\n"]},{"cell_type":"code","execution_count":7,"metadata":{"id":"Di4CFovupSGx"},"outputs":[],"source":["mod = MSSModel()\n","ts = mod.calculate_ts(m)"]},{"cell_type":"markdown","metadata":{"id":"bZP9vgKAqqwI"},"source":["## Comparison to existing target strength\n","\n","These results can be compared to those from the Jech et al (2015) paper. We can also calculate the mean difference between the Jech values and those from the echoSMs calculations."]},{"cell_type":"code","execution_count":8,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":497},"executionInfo":{"elapsed":2206,"status":"ok","timestamp":1724374342091,"user":{"displayName":"Gavin Macaulay","userId":"15996873107556149664"},"user_tz":-720},"id":"cXTE9u-TpumE","outputId":"46e575e0-361a-4f1e-eebe-acc8f90dbbef"},"outputs":[{"data":{"image/png":"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","text/plain":["
"]},"metadata":{},"output_type":"display_data"}],"source":["jech_index = np.mean(np.abs(ts - bmf['Sphere_WeaklyScattering']))\n","\n","fig, axs = plt.subplots(2, 1, sharex=True)\n","\n","axs[0].plot(m['f']/1e3, ts, label='echoSMs')\n","axs[0].plot(bmf['Frequency_kHz'], bmf['Sphere_WeaklyScattering'], label='Benchmark')\n","axs[0].set_ylabel('TS re 1 m$^2$ [dB]')\n","axs[0].legend(frameon=False, fontsize=6)\n","\n","axs[1].plot(m['f']*1e-3, ts-bmf['Sphere_WeaklyScattering'])\n","axs[1].set_xlabel('Frequency [kHz]')\n","axs[1].set_ylabel(r'$\\Delta$ TS [dB]')\n","axs[1].annotate(f'{jech_index:.2f} dB', (0.05, 0.80), xycoords='axes fraction',\n"," backgroundcolor=[.8, .8, .8])\n","_ = plt.suptitle('Weakly scattering sphere')"]},{"cell_type":"markdown","metadata":{"id":"QOHyiEE-vkbr"},"source":["There is a 0.15 dB difference between the echoSMs results and those from the Jech et al (2015) paper. We don't know why (comparisons of other models and parameters give near identical results - it is just the weakly scattering models that don't agree)."]}],"metadata":{"colab":{"authorship_tag":"ABX9TyOLhNe1NqPbGufGbKTfWjEp","provenance":[{"file_id":"1EPUlnNihQmkFtk5OvXHN0B0MUKSTvMkX","timestamp":1724374399220}]},"kernelspec":{"display_name":"Python 3","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.11.9"}},"nbformat":4,"nbformat_minor":0} diff --git a/usage/index.html b/usage/index.html new file mode 100644 index 0000000..40fcbf8 --- /dev/null +++ b/usage/index.html @@ -0,0 +1,877 @@ + + + + + + + + + + + + + + + + + + + + + + + + + Using echoSMs - echoSMs + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ +
+ + + + + + +
+ + +
+ +
+ + + + + + +
+
+ + + +
+
+
+ + + + + + + +
+
+
+ + + + +
+
+ + + + + + + + + + + + + + + +

Using echoSMs

+

EchoSMs is (currently) a Python package that implements acoustic scattering models. Each different model is a separate Python class in the echoSMs package. They all inherit from a common base class that defines a common calling convention.

+

Installation

+

EchoSMs is available on PyPi as echosms. Install it with:

+
pip install echosms
+
+

The prolate spheroidal modal series model in echoSMs uses spheroidal wave functions. A high-accuracy implementation of these is available in the Python package spheroidalwavefunctions, as the versions provided by scipy are insufficient. This should be installed automatically when you install echosms, but note that spheroidalwavefunctions is currently only available for Linux and Windows on x86_64 CPU architectures (create an issue if you want it on a system that is not currently supported).

+

Versions

+

The installed version of echosms can be printed with this code:

+
import importlib
+print(importlib.metadata.version('echosms'))
+
+

The changelogs for echoSMs are available here. The latest version is always at the top of that list.

+

Model overview

+

Currently, the following models are available in echoSMs:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Model typePython class nameDescription
Deformed cylinder modelDCMModelTruncated cylinders with various boundary conditions
Elastic sphereESModelElastic spheres, such as echosounder calibration spheres
Kirchhoff approximationKAModelSurfaces that are mainly convex
Modal series solutionMSSModelSpheres with various boundary conditions, including shells
Prolate spheroidal modal seriesPSMSModelProlate spheroids with various boundary conditions
Phase-tracking distorted wave Born approximationPTDWBAModelWeakly scattering objects of any shape with inhomogeneous interiors
+

Future models will include more types of DWBA models, the Kirchhoff-ray-mode model, the Fourier matching method, and potentially finite element and boundary element models.

+

Running a model

+

Each echoSMs model expects input parameters that define the model (e.g., size, shape, material properties, etc). These can be provided in three ways:

+
    +
  • A Python dictionary with an entry for each parameter,
    • +
  • A Pandas DataFrame with columns for each parameter and a row for each model run,
    • +
  • An Xarray DataArray with as many dimensions as parameters. The parameter values are in the DataArray coordinate variables.
    • +
+

To use a model, you need to know what parameters it requires. These are documented in the calculate_ts_single function that each model has (refer to the echoSMS API reference for details). The units for numerical parameters will always follow the echoSMs unit convention. For example, the MSSModel, when simulating the scattering from a pressure release sphere, needs the following parameters:

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
NameDescription
medium_cSound speed in the fluid medium surrounding the target [m/s]
medium_rhoDensity of the fluid medium surrounding the target [kg/m³]
aRadius of the spherical target [m]
fFrequency to calculate the scattering at [Hz]
boundary_typeThe boundary type. Supported types are fixed rigid, pressure release, and fluid filled
+

The simplest way to provide these to the model is a dictionary:

+
    p = {'medium_rho': 1026.8,
+         'medium_c': 1477.4,
+         'a': 0.01, 
+         'boundary_type': 'pressure release',
+         'f': 38000}
+
+

An instance of the model can then be created and the calculate_ts function called with these parameters:

+
    from echosms import MSSModel
+    model = MSSModel()
+    model.calculate_ts(p)
+
+

This will return one TS value corresponding to the parameters given. If you want to run the model for a range of parameters, the relevant dictionary items can contain multiple values:

+
        import numpy as np
+        p = {'medium_rho': 1026.8,
+             'medium_c': 1477.4,
+             'a': 0.01,
+             'boundary_type': 'pressure release',
+             'f', np.arange(10, 100, 1)*1000}  # [Hz]
+        model.calculate_ts(p)
+
+

It is also fine to have multiple items with multiple values:

+
        p = {'medium_rho': 1026.8,
+             'medium_c': 1477.4,
+             'a': np.arange(0.01, 0.02, 0.001),  # [m]
+             'boundary_type': ['pressure release', 'fixed rigid'],
+             'f': np.arange(10, 100, 1)*1000}  # [Hz]
+        model.calculate_ts(p)
+
+

The TS will be calculated for all combinations of the parameters. To do this, echoSMs expands the parameters into a Pandas DataFrame with one column for each parameter and one row for each of the combinations. It then runs the model on each row of the DataFrame. That DataFrame, with the TS included, can be returned instead of a list of TS values by using the expand option:

+
        model.calculate_ts(p, expand=True)
+
+

An introductory Jupyter notebook is available that covers the above concepts and a Python script that covers this and more is available here.

+

Using DataFrames and DataArrays directly

+

Instead of passing a dictionary to the calculate_ts function, a DataFrame or DataArray can be passed instead. The crucial aspect is that the DataFrame columns must have the same names as the parameters that the model requires. For a DataArray, the coordinate dimensions must have the same names as the model parameters.

+

EchoSMS provides two utility functions (as_dataframe, and as_dataarray) to convert from a dictionary representation of model parameters to a DataFrame or DataArray, or you can construct your own, or modify those returned by the as_dataframe and as_dataarray functions.

+

The benefit of using a DataFrame is that you have fine control over what model runs will happen - it doesn't have to be the full set of combinations of input parameters. The benefit of using a DataArray is that it is easy to extract subsets of the results for further analysis and plotting.

+

For a DataFrame, the number of model runs will be the number of rows in the DataFrame. For a DataArray the number of models run will be the size of the DataArray (e.g., DataArray.size())

+

When passing a DataFrame to a model, you can choose whether the TS results are returned as a Series or are added to the existing DataFrame (in a column called ts). Use the inplace = True parameter in the call to calculate_ts for this. When passing a DataArray to a model, the TS results are always returned in the data part of the passed in DataArray.

+

More complex model parameters

+

Some models require parameters for which it is not sensible to duplicate them across rows in a DataFrame or as a dimension in a DataArray (e.g., the data that specifies the three-dimensional shape of a fish swimbladder). EchoSMs allows for this with the concept of non-expandable parameters - these are not expanded into DataFrame columns or DataArray dimensions and are available from the models no_expand_parameters attribute.

+

But, as it is very convenient to have all the model parameters in one data structure, echoSMs will store the non-expandable parameters as DataFrame or DataArray attributes. Due to a bug in the DataFrame implementation, the parameters are stored as a nested dictionary under a parameters attribute. An example of this is the PTDWBAModel:

+
    from echosms import PTDWBAModel, as_dataframe
+    import numpy as np
+
+    model = PTDWBAModel()
+    m = {'volume': np.full((5,5,5), 0),
+         'f': np.arange(10, 100, 1)*1000,
+         'rho': [1024, 1025],  
+         'c': [1500, 1501],
+         'voxel_size': (0.001, 0.001, 0.001),
+         'theta': 90,
+         'phi': 0}
+    m['volume'][3,3,3] = 1  # something to produce scatter
+    p = as_dataframe(m, model.no_expand_parameters)
+    model.calculate_ts(p, inplace=True)
+    print(p)
+
+

For the PTDWBA model, only theta and phi are expandable, so p contains just two columns. The remaining parameters are available via:

+
    p.attrs['parameters']
+
+

Note that while rho and c look like parameters that would be expanded, they are in the list of non-expandable parameters, so are not expanded. This is because the structure of the PTDWBA model means that it it not sensible to have variable parameters for rho and c.

+

If you pass the dictionary form of the parameters to a model, this treatment of non-expanding parameters is done automatically, where

+
    model.calculate_ts(m, expand=True)
+
+

returns the same results as

+
    p = as_dataframe(m, model.no_expand_parameters)
+    model.calculate_ts(p, inplace=True)`
+    print(p)
+
+

Multiprocessing

+

This is an experimental feature.

+

The multiprocess = True parameter in the call to calculate_ts will cause echoSMs to divide the requested model runs over as many cores as your computer has. Total solution time will decrease almost linearly with the number of models runs.

+

Reference model definitions

+

Jech et al., (2015) presented reference models for a range of scattering objects: spheres, spherical shells, prolate spheroids, and finite cylinders for several boundary conditions (fixed rigid, pressure release, fluid-filled) and parameters (backscatter as a function of frequency and incident angle). These model definitions are included in echoSMs via the ReferenceModels class, along with other objects, such as calibration spheres. For example, the names of all the model definitions are available with:

+
    from echosms import ReferenceModels
+    rm = ReferenceModels()
+    rm.names()
+
+

which returns:

+
['fixed rigid sphere',
+ 'pressure release sphere',
+ 'gas filled sphere',
+ 'weakly scattering sphere',
+ 'spherical fluid shell with pressure release interior',
+ 'spherical fluid shell with gas interior',
+ 'spherical fluid shell with weakly scattering interior',
+ 'fixed rigid prolate spheroid',
+ 'pressure release prolate spheroid',
+ 'gas filled prolate spheroid',
+ 'weakly scattering prolate spheroid',
+ 'fixed rigid finite cylinder',
+ 'pressure release finite cylinder',
+ 'gas filled finite cylinder',
+ 'weakly scattering finite cylinder',
+ 'WC20 calibration sphere',
+ 'WC21 calibration sphere',
+ 'WC22 calibration sphere',
+ 'WC25 calibration sphere',
+ 'WC38.1 calibration sphere',
+ 'WC57.2 calibration sphere',
+ 'WC60 calibration sphere',
+ 'WC64 calibration sphere',
+ 'Cu13.7 calibration sphere',
+ 'Cu23 calibration sphere',
+ 'Cu32 calibration sphere',
+ 'Cu42 calibration sphere',
+ 'Cu45 calibration sphere',
+ 'Cu60 calibration sphere',
+ 'Cu63 calibration sphere',
+ 'Cu64 calibration sphere']
+
+

and the specification for a particular model is given by:

+
    rm.specification('spherical fluid shell with weakly scattering interior')
+
+

which returns:

+
{'name': 'spherical fluid shell with weakly scattering interior',
+ 'shape': 'sphere',
+ 'boundary_type': 'fluid shell fluid interior',
+ 'description': 'A fluid spherical shell with a weakly scattering shell and interior',
+ 'a': 0.01,
+ 'shell_thickness': 0.001,
+ 'medium_rho': 1026.8,
+ 'medium_c': 1477.4,
+ 'shell_rho': 1028.9,
+ 'shell_c': 1480.3,
+ 'target_rho': 1031.0,
+ 'target_c': 1483.3,
+ 'source': 'https://doi.org/10.1121/1.4937607',
+ 'benchmark_model': 'mss'}
+
+

Note that the specification contains more information that the model itself needs, so the subset needed for running a model is available via:

+
    m = rm.parameters('spherical fluid shell with weakly scattering interior')
+    print(m)
+
+

which returns:

+
{'boundary_type': 'fluid shell fluid interior',
+ 'a': 0.01,
+ 'shell_thickness': 0.001,
+ 'medium_rho': 1026.8,
+ 'medium_c': 1477.4,
+ 'shell_rho': 1028.9,
+ 'shell_c': 1480.3,
+ 'target_rho': 1031.0,
+ 'target_c': 1483.3}
+
+

Note that the parameters() call does not return all of the parameters needed by a model. For example, f is not there and needs to be added before running a model:

+
    m['f'] = [38000, 40000, 42000]
+
+    from echosms import MSSModel
+    model = MSSModel()
+    model.calculate_ts(m)
+
+

Benchmark model TS

+

Jech et al., (2015) presented benchmark model runs for the reference models. The TS results from these benchmarks are available in echoSMs via the BenchMarkData class. This class is a simple wrapper around code that reads the CSV-formatted file of benchmark values into a Pandas DataFrame, whereupon they can be readily accessed like this:

+
    from echosms import BenchmarkData
+    bm = BenchmarkData()
+    bm.angle_dataset  # the TS as a function of angle at 38 kHz
+    bm.freq_dataset  # the TS as a function of frequency
+
+

The TS and frequency values for a particular benchmark are available with normal Pandas DataFrame indexing syntax. The DataFrame columns names as the same as the ReferenceModels names. For example:

+
    bm.freq_dataset['weakly scattering sphere']
+    bm.freq_dataset['frequency (kHz)']
+
+

or for the angle dataset:

+
    bm.angle_dataset['weakly scattering sphere']
+    bm.angle_dataset['angle (deg)']
+
+ + + + + + + + + + + + + +
+
+ + + +
+ + + +
+ + + +
+
+
+
+ + + + + + + + + + \ No newline at end of file

Python code should follow PEP8 and docstrings should use PEP257 with the contents following the numpydoc style. An exception to PEP8 is made to allow lines of up to 100 characters.