Fix for AttributeError thrown when calling metrics.pairwise_distances with binary metrics and Y is None

doc/whats_new/v0.21.rst

@@ -2,6 +2,29 @@
 
 .. currentmodule:: sklearn
 
+.. _changes_0_21_1:
+
+Version 0.21.1
+==============
+
+**May 2019**
+
+
+This is a bug-fix release with some minor documentation improvements and
+enhancements to features released in 0.21.0.
+
+Changelog
+---------
+
+:mod:`sklearn.metrics`
+......................
+
+- |Fix| Fixed a bug in :class:`metrics.pairwise_distances` where it would raise
+  ``AttributeError`` for boolean metrics when ``X`` had a boolean dtype and
+  ``Y == None``.
+  :issue:`13864` by :user:`Paresh Mathur <rick2047>`.
+
+
 .. _changes_0_21:
 
 Version 0.21.0

sklearn/metrics/pairwise.py

@@ -306,7 +306,7 @@ def _euclidean_distances_upcast(X, XX=None, Y=None, YY=None):
     maxmem = max(
         ((x_density * n_samples_X + y_density * n_samples_Y) * n_features
          + (x_density * n_samples_X * y_density * n_samples_Y)) / 10,
-        10 * 2**17)
+        10 * 2 ** 17)
 
     # The increase amount of memory in 8-byte blocks is:
     # - x_density * batch_size * n_features (copy of chunk of X)
@@ -315,7 +315,7 @@ def _euclidean_distances_upcast(X, XX=None, Y=None, YY=None):
     # Hence x² + (xd+yd)kx = M, where x=batch_size, k=n_features, M=maxmem
     #                                 xd=x_density and yd=y_density
     tmp = (x_density + y_density) * n_features
-    batch_size = (-tmp + np.sqrt(tmp**2 + 4 * maxmem)) / 2
+    batch_size = (-tmp + np.sqrt(tmp ** 2 + 4 * maxmem)) / 2
     batch_size = max(int(batch_size), 1)
 
     x_batches = gen_batches(X.shape[0], batch_size)
@@ -900,7 +900,7 @@ def sigmoid_kernel(X, Y=None, gamma=None, coef0=1):
     K = safe_sparse_dot(X, Y.T, dense_output=True)
     K *= gamma
     K += coef0
-    np.tanh(K, K)   # compute tanh in-place
+    np.tanh(K, K)  # compute tanh in-place
     return K
 
 
@@ -933,7 +933,7 @@ def rbf_kernel(X, Y=None, gamma=None):
 
     K = euclidean_distances(X, Y, squared=True)
     K *= -gamma
-    np.exp(K, K)    # exponentiate K in-place
+    np.exp(K, K)  # exponentiate K in-place
     return K
 
 
@@ -967,7 +967,7 @@ def laplacian_kernel(X, Y=None, gamma=None):
         gamma = 1.0 / X.shape[1]
 
     K = -gamma * manhattan_distances(X, Y)
-    np.exp(K, K)    # exponentiate K in-place
+    np.exp(K, K)  # exponentiate K in-place
     return K
 
 
@@ -1545,7 +1545,8 @@ def pairwise_distances(X, Y=None, metric="euclidean", n_jobs=None, **kwds):
 
         dtype = bool if metric in PAIRWISE_BOOLEAN_FUNCTIONS else None
 
-        if dtype == bool and (X.dtype != bool or Y.dtype != bool):
+        if (dtype == bool and
+                (X.dtype != bool or (Y is not None and Y.dtype != bool))):
             msg = "Data was converted to boolean for metric %s" % metric
             warnings.warn(msg, DataConversionWarning)
 
@@ -1576,7 +1577,6 @@ def pairwise_distances(X, Y=None, metric="euclidean", n_jobs=None, **kwds):
     'yule',
 ]
 
-
 # Helper functions - distance
 PAIRWISE_KERNEL_FUNCTIONS = {
     # If updating this dictionary, update the doc in both distance_metrics()

sklearn/metrics/tests/test_pairwise.py

@@ -173,6 +173,15 @@ def test_pairwise_boolean_distance(metric):
     with pytest.warns(DataConversionWarning, match=msg):
         pairwise_distances(X, metric=metric)
 
+    # Check that the warning is raised if X is boolean by Y is not boolean:
+    with pytest.warns(DataConversionWarning, match=msg):
+        pairwise_distances(X.astype(bool), Y=Y, metric=metric)
+
+    # Check that no warning is raised if X is already boolean and Y is None:
+    with pytest.warns(None) as records:
+        pairwise_distances(X.astype(bool), metric=metric)
+    assert len(records) == 0
+
 
 def test_no_data_conversion_warning():
     # No warnings issued if metric is not a boolean distance function