changed default_encoders to categorical_encoders

ISLP/models/columns.py

@@ -9,7 +9,6 @@
 from sklearn.utils.validation import check_is_fitted
 from sklearn.exceptions import NotFittedError
 
-
 class Column(NamedTuple):
 
     """
@@ -125,10 +124,7 @@ def _get_column_info(X,
                      columns,
                      is_categorical,
                      is_ordinal,
-                     default_encoders={
-                         'ordinal': OrdinalEncoder(),
-                         'categorical': OneHotEncoder()
-                         }
+                     categorical_encoders={}
                      ):
 
 
@@ -151,12 +147,18 @@ def _get_column_info(X,
         if is_categorical[i]:
             if is_ordinal[i]:
                 Xcol = _get_column(col, X) 
-                encoder = clone(default_encoders['ordinal'])
+                if col not in categorical_encoders:
+                    encoder = clone(categorical_encoders['ordinal'])
+                else:
+                    encoder = categorical_encoders[col]
                 encoder.fit(Xcol)
                 columns = ['{0}'.format(col)]
             else:
                 Xcol = _get_column(col, X) 
-                encoder = clone(default_encoders['categorical'])
+                if col not in categorical_encoders:
+                    encoder = clone(categorical_encoders['categorical'])
+                else:
+                    encoder = categorical_encoders[col]
                 cols = encoder.fit_transform(Xcol)
                 if hasattr(encoder, 'columns_'):
                     columns_ = encoder.columns_
@@ -181,7 +183,6 @@ def _get_column_info(X,
 # https://github.com/scikit-learn/scikit-learn/blob/2beed55847ee70d363bdbfe14ee4401438fba057/sklearn/ensemble/_hist_gradient_boosting/gradient_boosting.py
 # max_bins is ignored
 
-
 def _check_categories(categorical_features, X):
     """Check and validate categorical features in X
 

ISLP/models/model_spec.py

@@ -154,6 +154,7 @@ def fit(self, X, y=None):
             raise ValueError('method must be one of ["drop", "sum", None] or a callable' +
                              'that returns a contrast matrix and column names given the number' +
                              ' of levels')
+
         return self
 
     def transform(self, X):
@@ -171,22 +172,23 @@ def transform(self, X):
 
 class ModelSpec(TransformerMixin, BaseEstimator):
 
-    '''
-
-    Parameters
+    '''Parameters
     ----------
 
     terms : sequence (optional)
+
         Sequence of sets whose
         elements are columns of *X* when fit.
         For :py:class:`pd.DataFrame` these can be column
         names.
 
     intercept : bool (optional)
+
         Include a column for intercept?
 
     categorical_features : array-like of {bool, int} of shape (n_features) 
             or shape (n_categorical_features,), default=None.
+
         Indicates the categorical features. Will be ignored if *X* is a :py:class:`pd.DataFrame`
         or :py:class:`pd.Series`.
 
@@ -195,25 +197,31 @@ class ModelSpec(TransformerMixin, BaseEstimator):
         - integer array-like : integer indices indicating categorical
           features.
 
-    default_encoders : dict
-        Dictionary whose keys are elements of *terms* and values
-        are transforms to be applied to the associate columns in the model matrix
-        by running the *fit_transform* method when *fit* is called and overwriting
-        these values in the dictionary.
+    categorical_encoders : dict
+
+        Dictionary whose keys are elements of *terms* that represent
+        **categorical variables**. Its values are transforms to be
+        applied to the associate columns in the model matrix by
+        running the *fit_transform* method when *fit* is called and
+        overwriting these values in the dictionary.
+
     '''
 
     def __init__(self,
                  terms=[],
                  intercept=True,
                  categorical_features=None,
-                 default_encoders={'categorical': Contrast(method='drop'),
-                                   'ordinal': OrdinalEncoder()}
+                 categorical_encoders={}
                  ):
 
         self.intercept = intercept
         self.terms = terms
         self.categorical_features = categorical_features
-        self.default_encoders = default_encoders
+
+        self.categorical_encoders = categorical_encoders
+        self.categorical_encoders_ = {'ordinal': OrdinalEncoder(),
+                                      'categorical': Contrast(method='drop')}
+        self.categorical_encoders_.update(**categorical_encoders)
 
     def fit(self, X, y=None):
 
@@ -261,7 +269,8 @@ def fit(self, X, y=None):
                                              self.columns_,
                                              np.asarray(self.is_categorical_),
                                              np.asarray(self.is_ordinal_),
-                                             default_encoders=self.default_encoders)
+                                             categorical_encoders=self.categorical_encoders_)
+
         # include each column as a Feature
         # so that their columns are built if needed
 

pyproject.toml

@@ -51,7 +51,8 @@ doc = ['Sphinx>=3.0']
 [build-system]
 requires = ["setuptools>=42",
             "wheel",
-	    "versioneer[toml]"
+	    "versioneer[toml]",
+	    "Sphinx>=1.0"
 	    ]
 build-backend = "setuptools.build_meta"
 

setup.py

@@ -16,7 +16,6 @@
 
 from setuptools import setup
 
-
 # Define extensions
 EXTS = []
 

tests/models/test_columns.py

@@ -3,6 +3,7 @@
 
 from pandas.api.types import CategoricalDtype
 from ISLP.models.columns import _get_column_info
+from ISLP.models.model_spec import Contrast
 
 def test_column_info():
 
@@ -15,5 +16,7 @@ def test_column_info():
     print(_get_column_info(df,
                            df.columns,
                            [False]*4+[True],
-                           [False]*5))
+                           [False]*5,
+                           categorical_encoders={'categorical':Contrast(method='drop')}))
+
 

tests/models/test_model_matrix.py

@@ -37,7 +37,7 @@ def test_ndarray():
     X = rng.standard_normal((50,5))
 
     M = ModelSpec(terms=[1, (3,2)],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     M.fit(X)
     MX = M.transform(X)
 
@@ -51,7 +51,7 @@ def test_dataframe1():
     D = pd.DataFrame(X, columns=['A','B','C','D','E'])
 
     M = ModelSpec(terms=['A','D',('D','E')],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     clone(M)
     MX = np.asarray(M.fit_transform(D))
 
@@ -66,7 +66,7 @@ def test_dataframe2():
     D = pd.DataFrame(X, columns=['V','B','A','D','E'])
 
     M = ModelSpec(terms=['A', 'D', 'B', ('D','E'), 'V'],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     clone(M)
 
     MX = M.fit_transform(D)
@@ -83,7 +83,7 @@ def test_dataframe3():
     D['E'] = pd.Categorical(rng.choice(range(4,8), 50, replace=True))
 
     M = ModelSpec(terms=['A', 'E', ('D','E')],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = np.asarray(M.fit_transform(D))
     M2 = clone(M)
 
@@ -105,7 +105,7 @@ def test_dataframe4():
     D['E'] = pd.Categorical(rng.choice(range(4,8), 50, replace=True))
 
     M = ModelSpec(terms=['A', 'E', ('D','E'), 'D'],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = np.asarray(M.fit_transform(D))
 
     DE = pd.get_dummies(D['E'])
@@ -129,7 +129,7 @@ def test_dataframe5():
     D['E'] = pd.Categorical(rng.choice(range(4,8), 50, replace=True))
 
     M = ModelSpec(terms=['A', 'E', ('D','E')],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = np.asarray(M.fit_transform(D))
 
     # check they agree on copy of dataframe
@@ -148,7 +148,7 @@ def test_dataframe6():
     D['E'] = pd.Categorical(rng.choice(range(4,8), 50, replace=True))
 
     M = ModelSpec(terms=['A',W,(W,'D',)],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = M.fit_transform(D)
 
     MX = np.asarray(MX)
@@ -162,7 +162,7 @@ def test_dataframe7():
     D['Eee'] = pd.Categorical(rng.choice(range(4,8), 50, replace=True))
 
     M = ModelSpec(terms=D.columns.drop(['Y','C']),
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = M.fit_transform(D)
     print(MX.columns)
     MX = np.asarray(MX)
@@ -179,7 +179,7 @@ def test_dataframe8():
     # raises a ValueError because poly will have been already fit -- need new instance of Poly
     W = Feature(('A',), 'poly(A)', poly)
     M = ModelSpec(terms=list(D.columns.drop(['Y','C'])) + [(W,'E')],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = M.fit_transform(D)
 
     print(MX.columns)
@@ -198,7 +198,7 @@ def test_dataframe9():
     W = Feature(('A',), 'poly(A)', poly)
     U = Feature(('B',), 'poly(B)', clone(poly))
     M = ModelSpec(terms=list(D.columns.drop(['Y','C'])) + [W,U],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = M.fit_transform(D)
 
     print(MX.columns)
@@ -215,7 +215,7 @@ def test_dataframe10():
     D['E'] = pd.Categorical(rng.choice(range(4,8), 50, replace=True))
 
     M = ModelSpec(terms=['A', 'E', 'C', W, (W, 'D',), U],
-                  default_encoders=default_encoders)
+                  categorical_encoders=default_encoders)
     MX = M.fit_transform(D)
     print(MX.columns)
     MX = np.asarray(MX)