swev-id: scikit-learn__scikit-learn-13779 - skip None estimators in sample_weight validation

sklearn/ensemble/tests/test_voting.py

@@ -19,7 +19,7 @@
 from sklearn.svm import SVC
 from sklearn.multiclass import OneVsRestClassifier
 from sklearn.neighbors import KNeighborsClassifier
-from sklearn.base import BaseEstimator, ClassifierMixin
+from sklearn.base import BaseEstimator, ClassifierMixin, RegressorMixin
 from sklearn.dummy import DummyRegressor
 
 
@@ -348,6 +348,109 @@ def test_sample_weight():
     assert_raise_message(ValueError, msg, eclf3.fit, X, y, sample_weight)
 
 
+def test_none_estimator_with_sample_weight_classifier():
+    class SupportsSampleWeightClassifier(ClassifierMixin, BaseEstimator):
+        def fit(self, X, y, sample_weight=None):
+            self.sample_weight_ = sample_weight
+            return self
+
+        def predict(self, X):
+            return np.zeros(len(X), dtype=int)
+
+    class AnotherClassifier(SupportsSampleWeightClassifier):
+        pass
+
+    X_local = np.arange(6).reshape(-1, 1)
+    y_local = np.array([0, 1, 0, 1, 0, 1])
+    sample_weight = np.ones_like(y_local, dtype=float)
+
+    voter = VotingClassifier(
+        estimators=[
+            ('to_drop', SupportsSampleWeightClassifier()),
+            ('keep', AnotherClassifier())
+        ],
+        voting='hard'
+    )
+    voter.fit(X_local, y_local, sample_weight=sample_weight)
+    voter.set_params(to_drop=None)
+
+    voter.fit(X_local, y_local, sample_weight=sample_weight)
+
+    assert len(voter.estimators_) == 1
+    assert isinstance(voter.estimators_[0], AnotherClassifier)
+    assert np.array_equal(voter.estimators_[0].sample_weight_, sample_weight)
+
+
+def test_none_estimator_with_sample_weight_regressor():
+    class SupportsSampleWeightRegressor(RegressorMixin, BaseEstimator):
+        def fit(self, X, y, sample_weight=None):
+            self.sample_weight_ = sample_weight
+            return self
+
+        def predict(self, X):
+            return np.zeros(len(X), dtype=float)
+
+    class AnotherRegressor(SupportsSampleWeightRegressor):
+        pass
+
+    X_local = np.arange(6).reshape(-1, 1)
+    y_local = np.arange(6, dtype=float)
+    sample_weight = np.ones_like(y_local, dtype=float)
+
+    voter = VotingRegressor([
+        ('to_drop', SupportsSampleWeightRegressor()),
+        ('keep', AnotherRegressor())
+    ])
+    voter.fit(X_local, y_local, sample_weight=sample_weight)
+    voter.set_params(to_drop=None)
+
+    voter.fit(X_local, y_local, sample_weight=sample_weight)
+
+    assert len(voter.estimators_) == 1
+    assert isinstance(voter.estimators_[0], AnotherRegressor)
+    assert np.array_equal(voter.estimators_[0].sample_weight_, sample_weight)
+
+
+def test_sample_weight_mixed_support_with_none():
+    class SupportsSampleWeightClassifier(ClassifierMixin, BaseEstimator):
+        def fit(self, X, y, sample_weight=None):
+            self.sample_weight_ = sample_weight
+            return self
+
+        def predict(self, X):
+            return np.zeros(len(X), dtype=int)
+
+    class NoSampleWeightClassifier(ClassifierMixin, BaseEstimator):
+        def fit(self, X, y):
+            return self
+
+        def predict(self, X):
+            return np.zeros(len(X), dtype=int)
+
+    X_local = np.arange(6).reshape(-1, 1)
+    y_local = np.array([0, 1, 0, 1, 0, 1])
+    sample_weight = np.ones_like(y_local, dtype=float)
+
+    voter = VotingClassifier(
+        estimators=[
+            ('to_drop', SupportsSampleWeightClassifier()),
+            ('no_sw', NoSampleWeightClassifier())
+        ],
+        voting='hard'
+    )
+    voter.set_params(to_drop=None)
+
+    msg = "Underlying estimator 'no_sw' does not support sample weights."
+    assert_raise_message(
+        ValueError,
+        msg,
+        voter.fit,
+        X_local,
+        y_local,
+        sample_weight=sample_weight
+    )
+
+
 def test_sample_weight_kwargs():
     """Check that VotingClassifier passes sample_weight as kwargs"""
     class MockClassifier(BaseEstimator, ClassifierMixin):

sklearn/ensemble/voting.py

@@ -78,6 +78,8 @@ def fit(self, X, y, sample_weight=None):
 
         if sample_weight is not None:
             for name, step in self.estimators:
+                if step is None:
+                    continue
                 if not has_fit_parameter(step, 'sample_weight'):
                     raise ValueError('Underlying estimator \'%s\' does not'
                                      ' support sample weights.' % name)