allow mixed types in array func protocol

Author: szha

python/mxnet/numpy/multiarray.py

@@ -48,7 +48,7 @@
 from ..context import current_context
 from ..ndarray import numpy as _mx_nd_np
 from ..ndarray.numpy import _internal as _npi
-from ..ndarray.ndarray import _storage_type, from_numpy
+from ..ndarray.ndarray import _storage_type
 from .utils import _get_np_op
 from .fallback import *  # pylint: disable=wildcard-import,unused-wildcard-import
 from . import fallback
@@ -148,10 +148,12 @@ def _reshape_view(a, *shape):  # pylint: disable=redefined-outer-name
 
 def _as_mx_np_array(object, ctx=None):
     """Convert object to mxnet.numpy.ndarray."""
-    if isinstance(object, _np.ndarray):
+    if isinstance(object, ndarray):
+        return object
+    elif isinstance(object, _np.ndarray):
         if not object.flags['C_CONTIGUOUS']:
             object = _np.ascontiguousarray(object, dtype=object.dtype)
-        ret = from_numpy(object)
+        ret = array(object, ctx=ctx)
         return ret if ctx is None else ret.as_in_ctx(ctx=ctx)
     elif isinstance(object, (integer_types, numeric_types)):
         return object
@@ -358,11 +360,17 @@ def __array_function__(self, func, types, args, kwargs):  # pylint: disable=bad-
             out = func(*new_args, **new_kwargs)
             return _as_mx_np_array(out, ctx=cur_ctx)
         else:
-            # Note: this allows subclasses that don't override
-            # __array_function__ to handle mxnet.numpy.ndarray objects
-            if not py_all(issubclass(t, ndarray) for t in types):
-                return NotImplemented
-            return mx_np_func(*args, **kwargs)
+            if py_all(issubclass(t, ndarray) for t in types):
+                return mx_np_func(*args, **kwargs)
+            else:
+                try:
+                    cur_ctx = next(a.ctx for a in args if hasattr(a, 'ctx'))
+                except StopIteration:
+                    cur_ctx = next(a.ctx for a in kwargs.values() if hasattr(a, 'ctx'))
+                new_args = _as_mx_np_array(args, ctx=cur_ctx)
+                new_kwargs = {k: _as_mx_np_array(v, cur_ctx) for k, v in kwargs.items()}
+                return mx_np_func(*new_args, **new_kwargs)
+
 
     def _get_np_basic_indexing(self, key):
         """

tests/python/unittest/test_numpy_ndarray.py

@@ -1385,6 +1385,7 @@ def test_from_numpy(np_array, zero_copy):
     mx_array = mx.npx.from_numpy(np_array, zero_copy=zero_copy)
     mx.test_utils.assert_almost_equal(np_array, mx_array.asnumpy())
 
+@use_np
 def test_from_numpy_exception():
     np_array = _np.array([[1, 2], [3, 4], [5, 6]], dtype="float32")
     mx_array = mx.npx.from_numpy(np_array)
@@ -1397,3 +1398,9 @@ def test_from_numpy_exception():
     assert not np_array.flags["C_CONTIGUOUS"]
     with pytest.raises(ValueError):
         mx_array = mx.nd.from_numpy(np_array)
+
+@use_np
+def test_mixed_array_types():
+    np_array = _np.array([[1, 2], [3, 4], [5, 6]], dtype="float32")
+    mx_array = mx.np.ones((3, 1))
+    assert_almost_equal(mx_array + np_array, 1+np_array)