make sparse encoder map copy on write arrays as gram omp modifies in place

vighneshbirodkar · vighneshbirodkar · commit 562dffe43eb8 · 2015-12-10T10:49:32.000-05:00
diff --git a/sklearn/decomposition/dict_learning.py b/sklearn/decomposition/dict_learning.py
@@ -288,7 +288,7 @@ def sparse_encode(X, dictionary, gram=None, cov=None, algorithm='lasso_lars',
     code = np.empty((n_samples, n_components))
     slices = list(gen_even_slices(n_samples, _get_n_jobs(n_jobs)))
 
-    code_views = Parallel(n_jobs=n_jobs, verbose=verbose)(
+    code_views = Parallel(n_jobs=n_jobs, verbose=verbose, mmap_mode='c')(
         delayed(_sparse_encode)(
             X[this_slice], dictionary, gram,
             cov[:, this_slice] if cov is not None else None,
diff --git a/sklearn/decomposition/tests/test_dict_learning.py b/sklearn/decomposition/tests/test_dict_learning.py
@@ -238,3 +238,13 @@ def test_sparse_coder_estimator():
                        transform_alpha=0.001).transform(X)
     assert_true(not np.all(code == 0))
     assert_less(np.sqrt(np.sum((np.dot(code, V) - X) ** 2)), 0.1)
+
+
+def test_sparse_coder_mmap():
+    # Test that SparseCoder does not error by passing reading only
+    # arrays to child processes
+
+    init_dict = np.random.rand(500, 64)
+    data = np.random.rand(8096, 64)
+    sc = SparseCoder(init_dict, transform_algorithm='omp', n_jobs=2)
+    sc.fit_transform(data)
