[MRG+1] Raise error when SparseSeries is passed into classification metrics (#7373)

nielsenmarkus11 · jnothman · commit 3700b9606fb8 · 2017-10-21T23:12:33.000+11:00
* Raise error when SparseSeries is passed into roc_curve * Changed "y_true" in second if block to "y_score" * Remove code to import pandas and add sparseseries check to 'type_of_target' function. Finally, add 'type_of_target' call to _binary_clf_curve * Remove pandas import and old comparison in roc_curve. * Add test for 'type_of_target' function * Add white space after commas * Correct other white space issues * Move type_of_target test into try clause, remove test_precision_recall_curve_pos_label since as multiclass it doesn't make sense * Add test_precision_recall_curve_pos_label back in and also add test_binary_clf_curve to test new logic in _binary_clf_curve function * Correct syntax and formatting. * Remove trailing white space * Correct validation logic * Update test_multiclass.py per @jnothman 's request. * Import SkipTest function. * Remove extra white space from line 303
diff --git a/sklearn/metrics/ranking.py b/sklearn/metrics/ranking.py
@@ -311,7 +311,13 @@ def _binary_clf_curve(y_true, y_score, pos_label=None, sample_weight=None):
     thresholds : array, shape = [n_thresholds]
         Decreasing score values.
     """
-    check_consistent_length(y_true, y_score)
+    # Check to make sure y_true is valid
+    y_type = type_of_target(y_true)
+    if not (y_type == "binary" or
+            (y_type == "multiclass" and pos_label is not None)):
+        raise ValueError("{0} format is not supported".format(y_type))
+
+    check_consistent_length(y_true, y_score, sample_weight)
     y_true = column_or_1d(y_true)
     y_score = column_or_1d(y_score)
     assert_all_finite(y_true)
diff --git a/sklearn/metrics/tests/test_ranking.py b/sklearn/metrics/tests/test_ranking.py
@@ -457,6 +457,14 @@ def test_auc_score_non_binary_class():
                              roc_auc_score, y_true, y_pred)
 
 
+def test_binary_clf_curve():
+    rng = check_random_state(404)
+    y_true = rng.randint(0, 3, size=10)
+    y_pred = rng.rand(10)
+    msg = "multiclass format is not supported"
+    assert_raise_message(ValueError, msg, precision_recall_curve,
+                         y_true, y_pred)
+
 def test_precision_recall_curve():
     y_true, _, probas_pred = make_prediction(binary=True)
     _test_precision_recall_curve(y_true, probas_pred)
diff --git a/sklearn/utils/multiclass.py b/sklearn/utils/multiclass.py
@@ -243,6 +243,10 @@ def type_of_target(y):
         raise ValueError('Expected array-like (array or non-string sequence), '
                          'got %r' % y)
 
+    sparseseries = (y.__class__.__name__ == 'SparseSeries')
+    if sparseseries:
+        raise ValueError("y cannot be class 'SparseSeries'.")
+
     if is_multilabel(y):
         return 'multilabel-indicator'
 
diff --git a/sklearn/utils/tests/test_multiclass.py b/sklearn/utils/tests/test_multiclass.py
@@ -21,6 +21,7 @@
 from sklearn.utils.testing import assert_false
 from sklearn.utils.testing import assert_raises
 from sklearn.utils.testing import assert_raises_regex
+from sklearn.utils.testing import SkipTest
 
 from sklearn.utils.multiclass import unique_labels
 from sklearn.utils.multiclass import is_multilabel
@@ -295,6 +296,14 @@ def test_type_of_target():
                ' use a binary array or sparse matrix instead.')
         assert_raises_regex(ValueError, msg, type_of_target, example)
 
+    try:
+        from pandas import SparseSeries
+    except ImportError:
+        raise SkipTest("Pandas not found")
+
+    y = SparseSeries([1, 0, 0, 1, 0])
+    msg = "y cannot be class 'SparseSeries'."
+    assert_raises_regex(ValueError, msg, type_of_target, y)
 
 def test_class_distribution():
     y = np.array([[1, 0, 0, 1],
