Pep8ify code.

src/normi/__init__.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 """Normalized mutual information"""
+
 __all__ = ['NormalizedMI']
 
 NORMS = {'joint', 'geometric', 'arithmetic', 'min', 'max'}

src/normi/__main__.py

@@ -5,6 +5,7 @@
 All rights reserved.
 
 """
+
 import click
 import numpy as np
 
@@ -96,7 +97,13 @@
     help='Activate verbose mode.',
 )
 def main(
-    input_file, output_basename, norm, inv_measure, n_dims, precision, verbose,
+    input_file,
+    output_basename,
+    norm,
+    inv_measure,
+    n_dims,
+    precision,
+    verbose,
 ):
     # load file
     if verbose:

src/normi/_estimators.py

@@ -6,6 +6,7 @@
 All rights reserved.
 
 """
+
 __all__ = ['NormalizedMI']  # noqa: WPS410
 
 import numpy as np
@@ -32,7 +33,7 @@
     PositiveFloat,
     PositiveInt,
     PositiveMatrix,
-    ArrayLikePositiveInt
+    ArrayLikePositiveInt,
 )
 
 
@@ -221,7 +222,8 @@ def _reset(self) -> None:
 
     @beartype
     def nmi(
-        self, normalize_method: Optional[NormString] = None,
+        self,
+        normalize_method: Optional[NormString] = None,
     ) -> NormalizedMatrix:
         """Return the normalized mutual information matrix.
 
@@ -265,11 +267,13 @@ def nmi(
 
     @beartype
     def _kraskov_estimator(
-        self, X: List[Float2DArray],
+        self,
+        X: List[Float2DArray],
     ) -> Tuple[PositiveMatrix, FloatMatrix, FloatMatrix, FloatMatrix]:
         """Estimate the mutual information and entropies matrices."""
         mi: PositiveMatrix = np.empty(  # noqa: WPS317
-            (self._n_features, self._n_features), dtype=self._dtype,
+            (self._n_features, self._n_features),
+            dtype=self._dtype,
         )
         hxy: FloatMatrix = np.empty_like(mi)
         hx: FloatMatrix = np.empty_like(mi)
@@ -286,7 +290,7 @@ def _kraskov_estimator(
             hxy[idx_i, idx_i] = 1
             hx[idx_i, idx_i] = 1
             hy[idx_i, idx_i] = 1
-            for idx_j, xj in enumerate(X[idx_i + 1:], idx_i + 1):
+            for idx_j, xj in enumerate(X[idx_i + 1 :], idx_i + 1):
                 mi_ij, hxy_ij, hx_ij, hy_ij = kraskov_estimator(
                     xi,
                     xj,
@@ -337,9 +341,7 @@ def _scale_nearest_neighbor_distance(
     if invariant_measure == 'radius':
         return radii / np.mean(radii)
     elif invariant_measure == 'volume':
-        return radii / (
-            np.mean(radii ** n_dims) ** (1 / n_dims)
-        )
+        return radii / (np.mean(radii**n_dims) ** (1 / n_dims))
     elif invariant_measure == 'kraskov':
         return radii
     # This should never be reached
@@ -399,7 +401,9 @@ def kraskov_estimator(
     # Here we rely on NearestNeighbors to select the fastest algorithm.
     tree = KDTree(xy)
     radii: FloatArray = tree.query(
-        xy, k=n_neighbors + 1, **kdtree_kwargs,
+        xy,
+        k=n_neighbors + 1,
+        **kdtree_kwargs,
     )[0][:, 1:]  # neglect self count
     # take next smaller radii
     radii: FloatArray = np.nextafter(radii[:, -1], 0)
@@ -415,8 +419,12 @@ def kraskov_estimator(
     ny: FloatArray
     nx, ny = [
         KDTree(z).query_ball_point(
-            z, r=radii, return_length=True, **kdtree_kwargs,
-        ) - 1  # fix self count
+            z,
+            r=radii,
+            return_length=True,
+            **kdtree_kwargs,
+        )
+        - 1  # fix self count
         for z in (x, y)
     ]
 

src/normi/_typing.py

@@ -9,9 +9,11 @@
 All rights reserved.
 
 """
+
 import numpy as np
 from beartype.typing import List, Union
 from beartype.vale import Is, IsAttr, IsEqual
+
 from normi import INVMEASURES, NORMS
 
 try:  # for python <= 3.8 use typing_extensions
@@ -29,23 +31,27 @@ def _get_resolution(x):
 
 def _allclose(x, y) -> bool:
     """Wrapper around np.allclose with dtype dependent atol."""
-    atol = np.max([
-        _get_resolution(x),
-        _get_resolution(y),
-        # default value of numpy
-        1e-8,
-    ])
+    atol = np.max(
+        [
+            _get_resolution(x),
+            _get_resolution(y),
+            # default value of numpy
+            1e-8,
+        ]
+    )
     return np.allclose(x, y, atol=atol)
 
 
 class NDim:
     """Class for creating Validators checking for desired dimensions."""
+
     def __class_getitem__(self, ndim):
         return IsAttr['ndim', IsEqual[ndim]]
 
 
 class DType:
     """Class for creating Validators checking for desired dtype."""
+
     def __class_getitem__(self, dtype):
         return Is[lambda arr: np.issubdtype(arr.dtype, dtype)]
 

src/normi/_utils.py

@@ -6,6 +6,7 @@
 All rights reserved.
 
 """
+
 __all__ = ['savetxt']  # noqa: WPS410
 
 import datetime
@@ -37,8 +38,7 @@ def _get_rui() -> str:
     }
 
     return (
-        'This file was generated by nmi:\n{args}' +
-        '\n\n{date}, {user}@{pc}'
+        'This file was generated by nmi:\n{args}\n\n{date}, {user}@{pc}'
     ).format(**rui)