Merge pull request #375 from unnonouno/coo-shape-init

Coo shape init

cupy/sparse/coo.py

@@ -5,10 +5,12 @@
 except ImportError:
     _scipy_available = False
 
+import cupy
 from cupy import cusparse
 from cupy.sparse import base
 from cupy.sparse import csr
 from cupy.sparse import data as sparse_data
+from cupy.sparse import util
 
 
 class coo_matrix(sparse_data._data_matrix):
@@ -17,6 +19,10 @@ class coo_matrix(sparse_data._data_matrix):
 
     Now it has only one initializer format below:
 
+    ``coo_matrix((M, N), [dtype])``
+        It constructs an empty matrix whose shape is ``(M, N)``. Default dtype
+        is float64.
+
     ``coo_matrix((data, (row, col))``
         All ``data``, ``row`` and ``col`` are one-dimenaional
         :class:`cupy.ndarray`.
@@ -39,7 +45,17 @@ def __init__(self, arg1, shape=None, dtype=None, copy=False):
             raise ValueError(
                 'Only two-dimensional sparse arrays are supported.')
 
-        if isinstance(arg1, tuple) and len(arg1) == 2:
+        if util.isshape(arg1):
+            m, n = arg1
+            m, n = int(m), int(n)
+            data = cupy.zeros(0, dtype if dtype else 'd')
+            row = cupy.zeros(0, dtype='i')
+            col = cupy.zeros(0, dtype='i')
+            # shape and copy argument is ignored
+            shape = (m, n)
+            copy = False
+
+        elif isinstance(arg1, tuple) and len(arg1) == 2:
             try:
                 data, (row, col) = arg1
             except (TypeError, ValueError):

cupy/sparse/util.py

@@ -16,4 +16,4 @@ def isshape(x):
     if not isinstance(x, tuple) or len(x) != 2:
         return False
     m, n = x
-    return int(m) == m and int(n) == n
+    return isintlike(m) and isintlike(n)

tests/cupy_tests/sparse_tests/test_coo.py

@@ -39,6 +39,10 @@ def _make_empty(xp, sp, dtype):
     return sp.coo_matrix((data, (row, col)), shape=(3, 4))
 
 
+def _make_shape(xp, sp, dtype):
+    return sp.coo_matrix((3, 4))
+
+
 @testing.parameterize(*testing.product({
     'dtype': [numpy.float32, numpy.float64],
 }))
@@ -247,7 +251,7 @@ def test_unsupported_dtype(self):
 
 
 @testing.parameterize(*testing.product({
-    'make_method': ['_make', '_make_unordered', '_make_empty'],
+    'make_method': ['_make', '_make_unordered', '_make_empty', '_make_shape'],
     'dtype': [numpy.float32, numpy.float64],
 }))
 @unittest.skipUnless(scipy_available, 'requires scipy')
@@ -257,6 +261,11 @@ class TestCooMatrixScipyComparison(unittest.TestCase):
     def make(self):
         return globals()[self.make_method]
 
+    @testing.numpy_cupy_equal(sp_name='sp')
+    def test_dtype(self, xp, sp):
+        m = self.make(xp, sp, self.dtype)
+        return m.dtype
+
     @testing.numpy_cupy_equal(sp_name='sp')
     def test_nnz(self, xp, sp):
         m = self.make(xp, sp, self.dtype)