Add dask to array-api-compat

.github/workflows/array-api-tests-dask.yml

@@ -0,0 +1,12 @@
+name: Array API Tests (Dask)
+
+on: [push, pull_request]
+
+jobs:
+  array-api-tests-dask:
+    uses: ./.github/workflows/array-api-tests.yml
+    with:
+      package-name: dask
+      module-name: dask.array
+      extra-requires: numpy
+      pytest-extra-args: --disable-deadline --max-examples=5

.github/workflows/tests.yml

@@ -15,7 +15,7 @@ jobs:
       - name: Install Dependencies
         run: |
           python -m pip install --upgrade pip
-          python -m pip install pytest numpy torch
+          python -m pip install pytest numpy torch dask[array]
 
       - name: Run Tests
         run: |

array_api_compat/common/_aliases.py

@@ -303,6 +303,8 @@ def _asarray(
         import numpy as xp
     elif namespace == 'cupy':
         import cupy as xp
+    elif namespace == 'dask.array':
+        import dask.array as xp
     else:
         raise ValueError("Unrecognized namespace argument to asarray()")
 
@@ -322,11 +324,19 @@ def _asarray(
     if copy in COPY_FALSE:
         # copy=False is not yet implemented in xp.asarray
         raise NotImplementedError("copy=False is not yet implemented")
-    if isinstance(obj, xp.ndarray):
+    if (hasattr(xp, "ndarray") and isinstance(obj, xp.ndarray)) or hasattr(obj, "__array__"):
         if dtype is not None and obj.dtype != dtype:
             copy = True
         if copy in COPY_TRUE:
-            return xp.array(obj, copy=True, dtype=dtype)
+            copy_kwargs = {}
+            if namespace != "dask.array":
+                copy_kwargs["copy"] = True
+            else:
+                # No copy kw in dask.asarray so we go thorugh np.asarray first
+                # (like dask also does) but copy after
+                import numpy as np
+                obj = np.asarray(obj).copy()
+            return xp.array(obj, dtype=dtype, **copy_kwargs)
         return obj
 
     return xp.asarray(obj, dtype=dtype, **kwargs)

array_api_compat/common/_helpers.py

@@ -40,13 +40,23 @@ def _is_torch_array(x):
     # TODO: Should we reject ndarray subclasses?
     return isinstance(x, torch.Tensor)
 
+def _is_dask_array(x):
+    # Avoid importing dask if it isn't already
+    if 'dask.array' not in sys.modules:
+        return False
+
+    import dask.array
+
+    return isinstance(x, dask.array.Array)
+
 def is_array_api_obj(x):
     """
     Check if x is an array API compatible array object.
     """
     return _is_numpy_array(x) \
         or _is_cupy_array(x) \
         or _is_torch_array(x) \
+        or _is_dask_array(x) \
         or hasattr(x, '__array_namespace__')
 
 def _check_api_version(api_version):
@@ -95,6 +105,13 @@ def your_function(x, y):
             else:
                 import torch
                 namespaces.add(torch)
+        elif _is_dask_array(x):
+            _check_api_version(api_version)
+            if _use_compat:
+                from ..dask import array as dask_namespace
+                namespaces.add(dask_namespace)
+            else:
+                raise TypeError("_use_compat cannot be False if input array is a dask array!")
         elif hasattr(x, '__array_namespace__'):
             namespaces.add(x.__array_namespace__(api_version=api_version))
         else:
@@ -219,6 +236,13 @@ def to_device(x: "Array", device: "Device", /, *, stream: "Optional[Union[int, A
         return _cupy_to_device(x, device, stream=stream)
     elif _is_torch_array(x):
         return _torch_to_device(x, device, stream=stream)
+    elif _is_dask_array(x):
+        if stream is not None:
+            raise ValueError("The stream argument to to_device() is not supported")
+        # TODO: What if our array is on the GPU already?
+        if device == 'cpu':
+            return x
+        raise ValueError(f"Unsupported device {device!r}")
     return x.to_device(device, stream=stream)
 
 def size(x):

array_api_compat/common/_linalg.py

@@ -77,7 +77,8 @@ def matrix_rank(x: ndarray,
     # dimensional arrays.
     if x.ndim < 2:
         raise xp.linalg.LinAlgError("1-dimensional array given. Array must be at least two-dimensional")
-    S = xp.linalg.svd(x, compute_uv=False, **kwargs)
+    S = xp.linalg.svdvals(x, **kwargs)
+    #S = xp.linalg.svd(x, compute_uv=False, **kwargs)
     if rtol is None:
         tol = S.max(axis=-1, keepdims=True) * max(x.shape[-2:]) * xp.finfo(S.dtype).eps
     else:

array_api_compat/dask/array/__init__.py

@@ -0,0 +1,8 @@
+from dask.array import *
+
+# These imports may overwrite names from the import * above.
+from ._aliases import *
+
+__array_api_version__ = '2022.12'
+
+__import__(__package__ + '.linalg')

array_api_compat/dask/array/_aliases.py

@@ -0,0 +1,123 @@
+from __future__ import annotations
+
+from ...common import _aliases
+from ...common._helpers import _check_device
+
+from ..._internal import get_xp
+
+import numpy as np
+from numpy import (
+    # Constants
+    e,
+    inf,
+    nan,
+    pi,
+    newaxis,
+    # Dtypes
+    bool_ as bool,
+    float32,
+    float64,
+    int8,
+    int16,
+    int32,
+    int64,
+    uint8,
+    uint16,
+    uint32,
+    uint64,
+    complex64,
+    complex128,
+    iinfo,
+    finfo,
+    can_cast,
+    result_type,
+)
+
+from typing import TYPE_CHECKING
+if TYPE_CHECKING:
+    from typing import Optional, Union
+    from ...common._typing import ndarray, Device, Dtype
+
+import dask.array as da
+
+isdtype = get_xp(np)(_aliases.isdtype)
+astype = _aliases.astype
+
+# Common aliases
+
+# This arange func is modified from the common one to
+# not pass stop/step as keyword arguments, which will cause
+# an error with dask
+
+# TODO: delete the xp stuff, it shouldn't be necessary
+def dask_arange(
+    start: Union[int, float],
+    /,
+    stop: Optional[Union[int, float]] = None,
+    step: Union[int, float] = 1,
+    *,
+    xp,
+    dtype: Optional[Dtype] = None,
+    device: Optional[Device] = None,
+    **kwargs
+) -> ndarray:
+    _check_device(xp, device)
+    args = [start]
+    if stop is not None:
+        args.append(stop)
+    else:
+        # stop is None, so start is actually stop
+        # prepend the default value for start which is 0
+        args.insert(0, 0)
+    args.append(step)
+    return xp.arange(*args, dtype=dtype, **kwargs)
+
+arange = get_xp(da)(dask_arange)
+eye = get_xp(da)(_aliases.eye)
+
+from functools import partial
+asarray = partial(_aliases._asarray, namespace='dask.array')
+asarray.__doc__ = _aliases._asarray.__doc__
+
+linspace = get_xp(da)(_aliases.linspace)
+eye = get_xp(da)(_aliases.eye)
+UniqueAllResult = get_xp(da)(_aliases.UniqueAllResult)
+UniqueCountsResult = get_xp(da)(_aliases.UniqueCountsResult)
+UniqueInverseResult = get_xp(da)(_aliases.UniqueInverseResult)
+unique_all = get_xp(da)(_aliases.unique_all)
+unique_counts = get_xp(da)(_aliases.unique_counts)
+unique_inverse = get_xp(da)(_aliases.unique_inverse)
+unique_values = get_xp(da)(_aliases.unique_values)
+permute_dims = get_xp(da)(_aliases.permute_dims)
+std = get_xp(da)(_aliases.std)
+var = get_xp(da)(_aliases.var)
+empty = get_xp(da)(_aliases.empty)
+empty_like = get_xp(da)(_aliases.empty_like)
+full = get_xp(da)(_aliases.full)
+full_like = get_xp(da)(_aliases.full_like)
+ones = get_xp(da)(_aliases.ones)
+ones_like = get_xp(da)(_aliases.ones_like)
+zeros = get_xp(da)(_aliases.zeros)
+zeros_like = get_xp(da)(_aliases.zeros_like)
+reshape = get_xp(da)(_aliases.reshape)
+matrix_transpose = get_xp(da)(_aliases.matrix_transpose)
+vecdot = get_xp(da)(_aliases.vecdot)
+
+from dask.array import (
+    # Element wise aliases
+    arccos as acos,
+    arccosh as acosh,
+    arcsin as asin,
+    arcsinh as asinh,
+    arctan as atan,
+    arctan2 as atan2,
+    arctanh as atanh,
+    left_shift as bitwise_left_shift,
+    right_shift as bitwise_right_shift,
+    invert as bitwise_invert,
+    power as pow,
+    # Other
+    concatenate as concat,
+)
+
+del da

array_api_compat/dask/array/linalg.py

@@ -0,0 +1,42 @@
+from __future__ import annotations
+
+from dask.array.linalg import *
+from ...common import _linalg
+from ..._internal import get_xp
+from dask.array import matmul, tensordot, trace, outer
+from ._aliases import matrix_transpose, vecdot
+
+import dask.array as da
+
+from typing import TYPE_CHECKING
+if TYPE_CHECKING:
+    from typing import Optional, Union, Tuple
+    from ...common._typing import ndarray, Device, Dtype
+
+#cross = get_xp(da)(_linalg.cross)
+#outer = get_xp(da)(_linalg.outer)
+EighResult = _linalg.EighResult
+QRResult = _linalg.QRResult
+SlogdetResult = _linalg.SlogdetResult
+SVDResult = _linalg.SVDResult
+qr = get_xp(da)(_linalg.qr)
+#svd = get_xp(da)(_linalg.svd)
+cholesky = get_xp(da)(_linalg.cholesky)
+matrix_rank = get_xp(da)(_linalg.matrix_rank)
+#pinv = get_xp(da)(_linalg.pinv)
+matrix_norm = get_xp(da)(_linalg.matrix_norm)
+
+def svdvals(x: ndarray) -> Union[ndarray, Tuple[ndarray, ...]]:
+    # TODO: can't avoid computing U or V for dask
+    _, s, _ =  svd(x)
+    return s
+
+vector_norm = get_xp(da)(_linalg.vector_norm)
+diagonal = get_xp(da)(_linalg.diagonal)
+
+#__all__ = linalg_all + _linalg.__all__
+
+del get_xp
+del da
+#del linalg_all
+del _linalg

dask-skips.txt

@@ -0,0 +1,5 @@
+# FFT isn't conformant
+array_api_tests/test_fft.py
+
+# slow and not implemented in dask
+array_api_tests/test_linalg.py::test_matrix_power