Merge pull request #44 from Quansight-Labs/op_rop

Fix  __rop__ dunders

Author: ev-br

torch_np/_helpers.py

@@ -94,6 +94,8 @@ def to_tensors_or_none(*inputs):
 
 
 def _outer(x, y):
+    from ._ndarray import asarray
+
     x_tensor, y_tensor = to_tensors(x, y)
     result = torch.outer(x_tensor, y_tensor)
     return asarray(result)

torch_np/_ndarray.py

@@ -182,7 +182,11 @@ def __iadd__(self, other):
         return _binary_ufuncs.add(self, other, out=self)
 
     # sub, self - other
-    __sub__ = __rsub__ = _binary_ufuncs.subtract
+    __sub__ = _binary_ufuncs.subtract
+
+    # XXX: generate a function just for this? AND other non-commutative ops.
+    def __rsub__(self, other):
+        return _binary_ufuncs.subtract(other, self)
 
     def __isub__(self, other):
         return _binary_ufuncs.subtract(self, other, out=self)
@@ -194,20 +198,29 @@ def __imul__(self, other):
         return _binary_ufuncs.multiply(self, other, out=self)
 
     # div, self / other
-    __truediv__ = __rtruediv__ = _binary_ufuncs.divide
+    __truediv__ = _binary_ufuncs.divide
+
+    def __rtruediv__(self, other):
+        return _binary_ufuncs.divide(other, self)
 
     def __itruediv__(self, other):
         return _binary_ufuncs.divide(self, other, out=self)
 
     # floordiv, self // other
-    __floordiv__ = __rfloordiv__ = _binary_ufuncs.floor_divide
+    __floordiv__ = _binary_ufuncs.floor_divide
+
+    def __rfloordiv__(self, other):
+        return _binary_ufuncs.floor_divide(other, self)
 
     def __ifloordiv__(self, other):
         return _binary_ufuncs.floor_divide(self, other, out=self)
 
     # power, self**exponent
     __pow__ = __rpow__ = _binary_ufuncs.float_power
 
+    def __rpow__(self, exponent):
+        return _binary_ufuncs.float_power(exponent, self)
+
     def __ipow__(self, exponent):
         return _binary_ufuncs.float_power(self, exponent, out=self)
 

torch_np/_wrapper.py

@@ -21,6 +21,7 @@
 
 from . import _dtypes, _helpers, _decorators  # isort: skip  # XXX
 
+
 # Things to decide on (punt for now)
 #
 # 1. Q: What are the return types of wrapper functions: plain torch.Tensors or
@@ -273,12 +274,15 @@ def ones_like(a, dtype=None, order="K", subok=False, shape=None):
     return result
 
 
-# XXX: dtype=float
 @_decorators.dtype_to_torch
-def zeros(shape, dtype=float, order="C", *, like=None):
+def zeros(shape, dtype=None, order="C", *, like=None):
     _util.subok_not_ok(like)
     if order != "C":
         raise NotImplementedError
+    if dtype is None:
+        from ._detail._scalar_types import default_float_type
+
+        dtype = default_float_type.torch_dtype
     return asarray(torch.zeros(shape, dtype=dtype))
 
 

torch_np/tests/test_ufuncs_basic.py

@@ -233,7 +233,7 @@ def test_basic(self, ufunc, op, iop):
 
         # __radd__
         a = np.array([1, 2, 3])
-        assert_equal(op(1, a), ufunc(a, 1))
+        assert_equal(op(1, a), ufunc(1, a))
         assert_equal(op(a.tolist(), a), ufunc(a, a.tolist()))
 
         # __iadd__