Merge pull request #95 from Quansight-Labs/filter_windows

Add filter windows, tensordot

Author: ev-br

autogen/numpy_api_dump.py

@@ -134,10 +134,6 @@ def binary_repr(num, width=None):
     raise NotImplementedError
 
 
-def blackman(M):
-    raise NotImplementedError
-
-
 def block(arrays):
     raise NotImplementedError
 
@@ -337,14 +333,6 @@ def gradient(f, *varargs, axis=None, edge_order=1):
     raise NotImplementedError
 
 
-def hamming(M):
-    raise NotImplementedError
-
-
-def hanning(M):
-    raise NotImplementedError
-
-
 def histogram(a, bins=10, range=None, normed=None, weights=None, density=None):
     raise NotImplementedError
 
@@ -409,10 +397,6 @@ def ix_(*args):
     raise NotImplementedError
 
 
-def kaiser(M, beta):
-    raise NotImplementedError
-
-
 def lexsort(keys, axis=-1):
     raise NotImplementedError
 
@@ -759,10 +743,6 @@ def take(a, indices, axis=None, out=None, mode="raise"):
     raise NotImplementedError
 
 
-def tensordot(a, b, axes=2):
-    raise NotImplementedError
-
-
 def trapz(y, x=None, dx=1.0, axis=-1):
     raise NotImplementedError
 

torch_np/__init__.py

@@ -14,7 +14,7 @@
 
 inf = float("inf")
 nan = float("nan")
-from math import pi  # isort: skip
+from math import pi, e  # isort: skip
 
 False_ = asarray(False, bool_)
 True_ = asarray(True, bool_)

torch_np/_funcs.py

@@ -1158,6 +1158,13 @@ def vdot(a: ArrayLike, b: ArrayLike, /):
     return result.item()
 
 
+@normalizer
+def tensordot(a: ArrayLike, b: ArrayLike, axes=2):
+    if isinstance(axes, (list, tuple)):
+        axes = [[ax] if isinstance(ax, int) else ax for ax in axes]
+    return torch.tensordot(a, b, dims=axes)
+
+
 @normalizer
 def dot(a: ArrayLike, b: ArrayLike, out: Optional[NDArray] = None):
     dtype = _dtypes_impl.result_type_impl((a.dtype, b.dtype))
@@ -1850,3 +1857,36 @@ def __getitem__(self, item):
 
 index_exp = IndexExpression(maketuple=True)
 s_ = IndexExpression(maketuple=False)
+
+
+# ### Filter windows ###
+
+
+@normalizer
+def hamming(M):
+    dtype = _dtypes_impl.default_float_dtype
+    return torch.hamming_window(M, periodic=False, dtype=dtype)
+
+
+@normalizer
+def hanning(M):
+    dtype = _dtypes_impl.default_float_dtype
+    return torch.hann_window(M, periodic=False, dtype=dtype)
+
+
+@normalizer
+def kaiser(M, beta):
+    dtype = _dtypes_impl.default_float_dtype
+    return torch.kaiser_window(M, beta=beta, periodic=False, dtype=dtype)
+
+
+@normalizer
+def blackman(M):
+    dtype = _dtypes_impl.default_float_dtype
+    return torch.blackman_window(M, periodic=False, dtype=dtype)
+
+
+@normalizer
+def bartlett(M):
+    dtype = _dtypes_impl.default_float_dtype
+    return torch.bartlett_window(M, periodic=False, dtype=dtype)

torch_np/_ndarray.py

@@ -64,8 +64,13 @@ class ndarray:
     def __init__(self, t=None):
         if t is None:
             self.tensor = torch.Tensor()
+        elif isinstance(t, torch.Tensor):
+            self.tensor = t
         else:
-            self.tensor = torch.as_tensor(t)
+            raise ValueError(
+                "ndarray constructor is not recommended; prefer"
+                "either array(...) or zeros/empty(...)"
+            )
 
     @property
     def shape(self):

torch_np/tests/numpy_tests/core/test_numeric.py

@@ -2988,15 +2988,21 @@ def test_shape_mismatch_error_message(self):
             np.broadcast([[1, 2, 3]], [[4], [5]], [6, 7])
 
 
-@pytest.mark.xfail(reason="TODO")
 class TestTensordot:
 
     def test_zero_dimension(self):
         # Test resolution to issue #5663
-        a = np.ndarray((3,0))
-        b = np.ndarray((0,4))
+        a = np.zeros((3,0))
+        b = np.zeros((0,4))
         td = np.tensordot(a, b, (1, 0))
         assert_array_equal(td, np.dot(a, b))
+
+    @pytest.mark.xfail(reason="no einsum")
+    def test_zero_dimension_einsum(self):
+        # Test resolution to issue #5663
+        a = np.zeros((3,0))
+        b = np.zeros((0,4))
+        td = np.tensordot(a, b, (1, 0))
         assert_array_equal(td, np.einsum('ij,jk', a, b))
 
     def test_zero_dimensional(self):

torch_np/tests/numpy_tests/lib/test_function_base.py

@@ -26,14 +26,15 @@
 
 # FIXME: make from torch_np
 from numpy.lib import (
-    bartlett, blackman,
-    delete, digitize, extract, gradient, hamming, hanning,
-    insert, interp, kaiser, msort, piecewise, place,
+    delete, digitize, extract, gradient,
+    insert, interp, msort, piecewise, place,
     select, setxor1d, trapz, trim_zeros, unwrap, vectorize
     )
 from torch_np._detail._util import normalize_axis_tuple
 
 from torch_np import corrcoef, cov, i0, angle, sinc, diff, meshgrid, unique
+from torch_np import flipud, hamming, hanning, kaiser, blackman, bartlett
+
 
 def get_mat(n):
     data = np.arange(n)
@@ -1701,22 +1702,21 @@ def test_period(self):
         assert sm_discont.dtype == wrap_uneven.dtype
 
 
-@pytest.mark.xfail(reason='TODO: implement')
 @pytest.mark.parametrize(
     "dtype", np.typecodes["AllInteger"] + np.typecodes["Float"]
 )
 @pytest.mark.parametrize("M", [0, 1, 10])
 class TestFilterwindows:
 
     def test_hanning(self, dtype: str, M: int) -> None:
-        scalar = np.array(M, dtype=dtype)[()]
+        scalar = M
 
         w = hanning(scalar)
-        ref_dtype = np.result_type(scalar.dtype, np.float64)
+        ref_dtype = np.result_type(dtype, np.float64)
         assert w.dtype == ref_dtype
 
         # check symmetry
-        assert_equal(w, flipud(w))
+        assert_allclose(w, flipud(w), atol=1e-15)
 
         # check known value
         if scalar < 1:
@@ -1727,14 +1727,14 @@ def test_hanning(self, dtype: str, M: int) -> None:
             assert_almost_equal(np.sum(w, axis=0), 4.500, 4)
 
     def test_hamming(self, dtype: str, M: int) -> None:
-        scalar = np.array(M, dtype=dtype)[()]
+        scalar = M
 
         w = hamming(scalar)
-        ref_dtype = np.result_type(scalar.dtype, np.float64)
+        ref_dtype = np.result_type(dtype, np.float64)
         assert w.dtype == ref_dtype
 
         # check symmetry
-        assert_equal(w, flipud(w))
+        assert_allclose(w, flipud(w), atol=1e-15)
 
         # check known value
         if scalar < 1:
@@ -1745,14 +1745,14 @@ def test_hamming(self, dtype: str, M: int) -> None:
             assert_almost_equal(np.sum(w, axis=0), 4.9400, 4)
 
     def test_bartlett(self, dtype: str, M: int) -> None:
-        scalar = np.array(M, dtype=dtype)[()]
+        scalar = M
 
         w = bartlett(scalar)
-        ref_dtype = np.result_type(scalar.dtype, np.float64)
+        ref_dtype = np.result_type(dtype, np.float64)
         assert w.dtype == ref_dtype
 
         # check symmetry
-        assert_equal(w, flipud(w))
+        assert_allclose(w, flipud(w), atol=1e-15)
 
         # check known value
         if scalar < 1:
@@ -1763,14 +1763,14 @@ def test_bartlett(self, dtype: str, M: int) -> None:
             assert_almost_equal(np.sum(w, axis=0), 4.4444, 4)
 
     def test_blackman(self, dtype: str, M: int) -> None:
-        scalar = np.array(M, dtype=dtype)[()]
+        scalar = M
 
         w = blackman(scalar)
-        ref_dtype = np.result_type(scalar.dtype, np.float64)
+        ref_dtype = np.result_type(dtype, np.float64)
         assert w.dtype == ref_dtype
 
         # check symmetry
-        assert_equal(w, flipud(w))
+        assert_allclose(w, flipud(w), atol=1e-15)
 
         # check known value
         if scalar < 1:
@@ -1781,10 +1781,10 @@ def test_blackman(self, dtype: str, M: int) -> None:
             assert_almost_equal(np.sum(w, axis=0), 3.7800, 4)
 
     def test_kaiser(self, dtype: str, M: int) -> None:
-        scalar = np.array(M, dtype=dtype)[()]
+        scalar = M
 
         w = kaiser(scalar, 0)
-        ref_dtype = np.result_type(scalar.dtype, np.float64)
+        ref_dtype = np.result_type(dtype, np.float64)
         assert w.dtype == ref_dtype
 
         # check symmetry