ensure sorted indices tested

theano/sparse/sandbox/sp.py

@@ -273,7 +273,9 @@ class EnsureSortedIndices(Op):
     def grad(self, (x,), (gz,)):
         return [gz]
 
-
+    def infer_shape(self, node, i0_shapes):
+        return i0_shapes
+            
     def __str__(self):
         if self.inplace:
             return self.__class__.__name__ + "{inplace}"

theano/sparse/sandbox/test_sp.py

@@ -16,7 +16,7 @@ from theano import function, tensor
 import theano
 from theano.sparse.sandbox import sp
 from theano.tests import unittest_tools as utt
-
+from theano.sparse.tests.test_basic import random_lil
 
 class TestSP(unittest.TestCase):
     def test_convolution(self):
@@ -414,32 +414,25 @@ def test_diagonal():
         assert numpy.all(n == f(range(K)).toarray())
 
 def test_EnsureSortedIndices():
-    
-    input_tensor = theano.sparse.csc_dmatrix()
-    sort_op = sp.ensure_sorted_indices(input_tensor)
-    f = theano.function([input_tensor], sort_op)
-    
-    # test_sample = scipy.sparse.rand(1000, 1000, density=0.01)
-    # the function above is not working since scipy has lower version
-    # specify the number of rows and columns in the test case
-    r = 2000
-    c = 2000
-    # construct a test sample
-    test_sample = numpy.random.rand(r,c)
-    indices_zero_x = numpy.random.permutation(range(r))
-    indices_zero_y = numpy.random.permutation(range(c))
-    # take 70% of the test_sample and make them 0
-    sparse_ratio_x = int(r*0.9)
-    sparse_ratio_y = int(c*0.9)
-    for i in indices_zero_x[:sparse_ratio_x]:
-        for j in indices_zero_y[:sparse_ratio_y]:
-            test_sample[i,j] = 0
-    # sort test sample by scipy
-    test_sample = scipy.sparse.csc_matrix(test_sample)
-    sorted_for_real = test_sample.sorted_indices()
-    
-    sorted_theano = f(test_sample)
-    assert numpy.all(sorted_theano.todense() == sorted_for_real.todense())
+    x = 2000
+    y = 2000
+    sparsity = 1000
+    for i in range(2):
+        # testing both csc and csr
+        if i is 0:
+            # csc
+            input_tensor = theano.sparse.csc_dmatrix()
+            sample = scipy.sparse.csc_matrix(random_lil((x,y),'float64',sparsity))
+        else:
+            # csr
+            input_tensor = theano.sparse.csr_dmatrix()
+            sample = scipy.sparse.csr_matrix(random_lil((x,y),'float64',sparsity))
+            
+        sort_op = sp.ensure_sorted_indices(input_tensor)
+        f = theano.function([input_tensor], sort_op)
+        sorted_scipy = sample.sorted_indices()
+        sorted_theano = f(sample)
+        assert numpy.all(sorted_theano.todense() == sorted_scipy.todense())
     
 def test_diagonal_grad():
     def d(x):