ENH: add a new DataConversionWarning

for implicit conversions such as raveling the y

Author: GaelVaroquaux

sklearn/tests/test_common.py

@@ -43,6 +43,7 @@
 from sklearn.svm.base import BaseLibSVM
 
 from sklearn.cross_validation import train_test_split
+from sklearn.utils.validation import DataConversionWarning
 
 dont_test = ['SparseCoder', 'EllipticEnvelope', 'EllipticEnvelop',
              'DictVectorizer', 'LabelBinarizer', 'LabelEncoder',
@@ -495,7 +496,6 @@ def test_clustering():
             continue
         # catch deprecation and neighbors warnings
         with warnings.catch_warnings(record=True):
-            warnings.simplefilter("always")
             alg = Alg()
             if hasattr(alg, "n_clusters"):
                 alg.set_params(n_clusters=3)
@@ -663,9 +663,12 @@ def test_classifiers_input_shapes():
 
         set_random_state(classifier)
         classifier.fit(X, y[:, np.newaxis])
+        # Check that when a 2D y is given, a DataConversionWarning is
+        # raised
         with warnings.catch_warnings(record=True) as w:
+            warnings.simplefilter("always", DataConversionWarning)
             classifier.fit(X, y[:, np.newaxis])
-            print(w)
+        assert_equal(len(w), 1)
         assert_array_equal(y_pred, classifier.predict(X))
 
 

sklearn/utils/validation.py

@@ -12,6 +12,14 @@
 from .fixes import safe_copy
 
 
+class DataConversionWarning(UserWarning):
+    "A warning on implicit data conversions happening in the code"
+    pass
+
+
+warnings.simplefilter("always", DataConversionWarning)
+
+
 def _assert_all_finite(X):
     """Like assert_all_finite, but only for ndarray."""
     if (X.dtype.char in np.typecodes['AllFloat'] and not np.isfinite(X.sum())
@@ -250,7 +258,8 @@ def column_or_1d(y, warn=False):
         if warn:
             warnings.warn("A column-vector y was passed when a 1d array was"
                           " expected. Please change the shape of y to "
-                          "(n_samples, ), for example using ravel().")
+                          "(n_samples, ), for example using ravel().",
+                          DataConversionWarning, stacklevel=2)
         return np.ravel(y)
 
     raise ValueError("bad input shape {0}".format(shape))