Merge pull request #228 from crusaderky/is_lazy_array

ENH: is_lazy_array()

Author: ev-br

Diff for: array_api_compat/common/_helpers.py

@@ -824,6 +824,63 @@ def is_writeable_array(x) -> bool:
     return True
 
 
+def is_lazy_array(x) -> bool:
+    """Return True if x is potentially a future or it may be otherwise impossible or
+    expensive to eagerly read its contents, regardless of their size, e.g. by
+    calling ``bool(x)`` or ``float(x)``.
+
+    Return False otherwise; e.g. ``bool(x)`` etc. is guaranteed to succeed and to be
+    cheap as long as the array has the right dtype and size.
+
+    Note
+    ----
+    This function errs on the side of caution for array types that may or may not be
+    lazy, e.g. JAX arrays, by always returning True for them.
+    """
+    if (
+        is_numpy_array(x)
+        or is_cupy_array(x)
+        or is_torch_array(x)
+        or is_pydata_sparse_array(x)
+    ):
+        return False
+
+    # **JAX note:** while it is possible to determine if you're inside or outside
+    # jax.jit by testing the subclass of a jax.Array object, as well as testing bool()
+    # as we do below for unknown arrays, this is not recommended by JAX best practices.
+
+    # **Dask note:** Dask eagerly computes the graph on __bool__, __float__, and so on.
+    # This behaviour, while impossible to change without breaking backwards
+    # compatibility, is highly detrimental to performance as the whole graph will end
+    # up being computed multiple times.
+
+    if is_jax_array(x) or is_dask_array(x) or is_ndonnx_array(x):
+        return True
+
+    # Unknown Array API compatible object. Note that this test may have dire consequences
+    # in terms of performance, e.g. for a lazy object that eagerly computes the graph
+    # on __bool__ (dask is one such example, which however is special-cased above).
+
+    # Select a single point of the array
+    s = size(x)
+    if s is None:
+        return True
+    xp = array_namespace(x)
+    if s > 1:
+        x = xp.reshape(x, (-1,))[0]
+    # Cast to dtype=bool and deal with size 0 arrays
+    x = xp.any(x)
+
+    try:
+        bool(x)
+        return False
+    # The Array API standard dictactes that __bool__ should raise TypeError if the
+    # output cannot be defined.
+    # Here we allow for it to raise arbitrary exceptions, e.g. like Dask does.
+    except Exception:
+        return True
+
+
 __all__ = [
     "array_namespace",
     "device",
@@ -845,6 +902,7 @@ def is_writeable_array(x) -> bool:
     "is_pydata_sparse_array",
     "is_pydata_sparse_namespace",
     "is_writeable_array",
+    "is_lazy_array",
     "size",
     "to_device",
 ]

Diff for: docs/helper-functions.rst

@@ -52,6 +52,7 @@ yet.
 .. autofunction:: is_pydata_sparse_array
 .. autofunction:: is_ndonnx_array
 .. autofunction:: is_writeable_array
+.. autofunction:: is_lazy_array
 .. autofunction:: is_numpy_namespace
 .. autofunction:: is_cupy_namespace
 .. autofunction:: is_torch_namespace

Diff for: tests/test_common.py

@@ -1,3 +1,10 @@
+import math
+
+import pytest
+import numpy as np
+import array
+from numpy.testing import assert_allclose
+
 from array_api_compat import (  # noqa: F401
     is_numpy_array, is_cupy_array, is_torch_array,
     is_dask_array, is_jax_array, is_pydata_sparse_array,
@@ -6,15 +13,10 @@
 )
 
 from array_api_compat import (
-    device, is_array_api_obj, is_writeable_array, size, to_device
+    device, is_array_api_obj, is_lazy_array, is_writeable_array, size, to_device
 )
 from ._helpers import import_, wrapped_libraries, all_libraries
 
-import pytest
-import numpy as np
-import array
-from numpy.testing import assert_allclose
-
 is_array_functions = {
     'numpy': 'is_numpy_array',
     'cupy': 'is_cupy_array',
@@ -115,6 +117,45 @@ def test_size_none(library):
     assert size(x) in (None, 5)
 
 
+@pytest.mark.parametrize("library", all_libraries)
+def test_is_lazy_array(library):
+    lib = import_(library)
+    x = lib.asarray([1, 2, 3])
+    assert isinstance(is_lazy_array(x), bool)
+
+
+@pytest.mark.parametrize("shape", [(math.nan,), (1, math.nan), (None, ), (1, None)])
+def test_is_lazy_array_nan_size(shape, monkeypatch):
+    """Test is_lazy_array() on an unknown Array API compliant object
+    with NaN (like Dask) or None (like ndonnx) in its shape
+    """
+    xp = import_("array_api_strict")
+    x = xp.asarray(1)
+    assert not is_lazy_array(x)
+    monkeypatch.setattr(type(x), "shape", shape)
+    assert is_lazy_array(x)
+
+
+@pytest.mark.parametrize("exc", [TypeError, AssertionError])
+def test_is_lazy_array_bool_raises(exc, monkeypatch):
+    """Test is_lazy_array() on an unknown Array API compliant object
+    where calling bool() raises:
+    - TypeError: e.g. like jitted JAX. This is the proper exception which
+      lazy arrays should raise as per the Array API specification
+    - something else: e.g. like Dask, where bool() triggers compute()
+      which can result in any kind of exception to be raised
+    """
+    xp = import_("array_api_strict")
+    x = xp.asarray(1)
+    assert not is_lazy_array(x)
+
+    def __bool__(self):
+        raise exc("Hello world")
+
+    monkeypatch.setattr(type(x), "__bool__", __bool__)
+    assert is_lazy_array(x)
+
+
 @pytest.mark.parametrize("library", all_libraries)
 def test_device(library):
     xp = import_(library, wrapper=True)
@@ -172,6 +213,7 @@ def test_asarray_cross_library(source_library, target_library, request):
 
     assert is_tgt_type(b), f"Expected {b} to be a {tgt_lib.ndarray}, but was {type(b)}"
 
+
 @pytest.mark.parametrize("library", wrapped_libraries)
 def test_asarray_copy(library):
     # Note, we have this test here because the test suite currently doesn't