[MRG + 1] Issue#8062: JoblibException thrown when passing "fit_params={'sample_… (scikit-learn#8068)

xor · Sundrique · commit a7e6bc007a84 · 2017-06-14T11:42:40.000+08:00
* Issue#8062: JoblibException thrown when passing "fit_params={'sample_weights': weights}" to RandomizedSearchCV with RandomForestClassifier * Added test for issues scikit-learn#8068 and scikit-learn#8064. * Clean up with pyflakes. * Changed cryptic comment.
diff --git a/sklearn/ensemble/forest.py b/sklearn/ensemble/forest.py
@@ -245,6 +245,8 @@ def fit(self, X, y, sample_weight=None):
         # Validate or convert input data
         X = check_array(X, accept_sparse="csc", dtype=DTYPE)
         y = check_array(y, accept_sparse='csc', ensure_2d=False, dtype=None)
+        if sample_weight is not None:
+            sample_weight = check_array(sample_weight, ensure_2d=False)
         if issparse(X):
             # Pre-sort indices to avoid that each individual tree of the
             # ensemble sorts the indices.
diff --git a/sklearn/ensemble/tests/test_forest.py b/sklearn/ensemble/tests/test_forest.py
@@ -947,6 +947,12 @@ def check_class_weights(name):
     clf2.fit(iris.data, iris.target, sample_weight)
     assert_almost_equal(clf1.feature_importances_, clf2.feature_importances_)
 
+    # Using a Python 2.x list as the sample_weight parameter used to raise
+    # an exception. This test makes sure such code will now run correctly.
+    clf = ForestClassifier()
+    sample_weight = [1.] * len(iris.data)
+    clf.fit(iris.data, iris.target, sample_weight=sample_weight)
+
 
 def test_class_weights():
     for name in FOREST_CLASSIFIERS:
