changelog.md

Changelog

1.9.1 (2024-10-29)

Major Changes

• __array_api_version__ for the wrapped APIs is now set to 2023.12.

Minor Changes

• Wrap sign so that it always uses the standard definition for complex numbers, and always propagates nans.

• Wrap dask.array.fft.

• Readd python_requires to the package metadata.

1.9 (2024-10-??)

Major Changes

• New helper functions to determine if a namespace is from a given library ({func}~.is_numpy_namespace, {func}~.is_torch_namespace, etc.).

• More support for the 2023.12 version of the standard. This includes

  • Wrappers for cumulative_sum().
  • Wrappers for unstack().
  • Update floating-point type promotion in sum(), prod(), and trace() to be inline with the 2023.12 specification (32-bit types no longer promote to 64-bit when dtype=None).
  • Add the inspection APIs to the wrapped namespaces. These can be accessed with xp.__array_namespace_info__().
  • Various fixes to the clip() wrappers.

• torch.conj now wrapps torch.conj_physical, which makes a copy rather than setting the conjugation bit, as arrays with the conjugation bit set do not support some APIs.

• torch.sign is now wrapped to support complex numbers and propogate nans properly.

Minor Changes

• NumPy 2.0 is now wrapped again. Previously it was unwrapped because it has full 2022.12 array API support but it now requires wrapping again for 2023.12 support.

• Support for JAX 0.4.32 and newer which implements the array API directly in jax.numpy.

• hypot, minimum, and maximum (new in 2023.12) are wrapped in PyTorch to support proper scalar type promotion.

1.8 (2024-07-24)

Major Changes

• Add support for ndonnx. Array API support itself lives in the ndonnx library, but this adds the {func}~.is_ndonnx_array helper function. (@adityagoel4512).

• Partial support for the 2023.12 version of the standard. This includes

  • Wrappers for clip().
  • torch wrapper for copysign() with correct type promotion.

  Note that many of the new functions in the 2023.12 version of the standard are already fully implemented in upstream libraries and will already work.

1.7.1 (2024-05-28)

Minor Changes

• Fix a typo in setup.py (@sunpoet).

1.7 (2024-05-24)

Major Changes

• Add support for sparse. Note that unlike other array libraries, array-api-compat does not contain any wrappers for sparse functions. All sparse array API support is in sparse itself. Thus, there is no array_api_compat.sparse submodule, and array_namespace(<pydata/sparse array>) returns the sparse module.

• Added the function is_pydata_sparse_array(x).

Minor Changes

• Fix JAX float0 arrays. See jax-ml/jax#20620. (@NeilGirdhar)

• Fix torch.linalg.vector_norm() when axis=().

• Fix torch.linalg.solve() to apply the array API standard rules for when x2 should be treated as a vector vs. a matrix.

• Fix PyTorch test failures on CI by skipping uint16, uint32, uint64 tests.

1.6 (2024-03-29)

Major Changes

• Drop support for Python 3.8.

• NumPy 2.0 is now left completely unwrapped.

• New flag use_compat to {func}~.array_namespace to force the use or non-use of the compat wrapper namespace. The default is to return a compat namespace when it is appropiate.

• Fix the copy flag to asarray for NumPy, CuPy, and Dask.

• Fix the device flag to asarray for CuPy.

• Fix various issues with asarray for Dask.

Minor Changes

• Test Python 3.12 on CI.

• Add more tests for {func}~.array_namespace.

• Add more tests for asarray.

• Add a test that there are no hard dependencies.

1.5.1 (2024-03-20)

Minor Changes

• Add HTML documentation. Includes new documentation on the scope of the package and new developer documentation.

• Fix array_api_compat.numpy.asarray(torch.Tensor) to return a NumPy array.

• Allow Python scalars in torch functions.

• Fix the torch.std wrapper when correction is an int.

• Fix issues with qr and svd in the Dask wrappers.

1.5 (2024-03-07)

Major Changes

• Add support for Dask (@lithomas1).

• Add support for JAX. Note that unlike other array libraries, array-api-compat does not contain any wrappers for JAX functions. All JAX array API support is in JAX itself. Thus, there is no array_api_compat.jax submodule, and array_namespace(<JAX array>) returns the jax.experimental.array_api module.

• The functions is_numpy_array(x), is_cupy_array(x), is_torch_array(x), is_dask_array(x), is_jax_array(x) are now part of the public array_api_compat API.

• Add wrappers for the fft extension module for NumPy, CuPy, and PyTorch.

Minor Changes

• Allow '2022.12' as the api_version in {func}~.array_namespace(). '2021.12' is also supported but will issue a warning since the returned namespace will still be a 2022.12 compliant one.

• Add wrapper for numpy.linalg.solve, which broadcasts the inputs according to the standard.

• Add wrappers for various PyTorch linalg functions.

• Fix a bug with numpy.linalg.vector_norm(keepdims=True).

• BREAKING: Update vecdot wrappers to apply axes before broadcasting, not after. This matches the updated 2023.12 standard wording, and also the behavior of the new numpy.vecdot gufunc in NumPy 2.0.

• Fix some linalg functions which were supposed to be in both the main namespace and the linalg extension namespace.

• Add Ruff to CI. (@adonath)

• Test that internal definitions of __all__ are self-consistent, which should help to avoid issues where wrappers are accidentally not exported to the compat namespaces properly.

1.4.1 (2024-01-18)

Minor Changes

• Add support for the upcoming NumPy 2.0 release.

• Added a torch wrapper for trace (torch.trace doesn't support the offset argument or stacking)

• Wrap numpy, cupy, and torch nonzero to raise an error for zero-dimensional input arrays.

• Add torch wrapper for newaxis.

• Improve error message for array_namespace

• Fix linalg.cholesky returning the conjugate of the expected upper decomposition for numpy and cupy.

1.4 (2023-09-13)

Major Changes

• Releases are now made with GitHub Actions (thanks @matthewfeickert).

Minor Changes

• Fix torch.result_type() cross-kind promotion (@lucascolley).

• Fix the torch.take() wrapper to make axis optional for ndim = 1.

• Add requires-python metadata to the package (@matthewfeickert).

1.3 (2023-06-20)

Major Changes

• Add 2022.12 standard support. This includes things like adding complex dtype support, adding the new take function, and various minor changes in the specification.

Minor Changes

• Support "cpu" in CuPy to_device().

• Return a new array in NumPy/CuPy reshape(copy=False).

• Fix signatures for PyTorch broadcast_to and permute_dims.

1.2 (2023-04-03)

Major Changes

• Support the linalg extension in the array_api_compat.torch namespace.

• Add isdtype().

Minor Changes

• Fix the k keyword argument to tril and triu in torch.

1.1.1 (2023-03-10)

Major Changes

• Rename get_namespace() to array_namespace() (get_namespace() is maintained as a backwards compatible alias).

Minor Changes

• The minimum supported NumPy version is now 1.21. Fixed a few issues with NumPy 1.21 (with unique_* and asarray), although there are also a few known issues with this version (see the README).

• Add api_version to get_namespace().

• array_namespace() (née get_namespace()) now works correctly with torch tensors.

• array_namespace() (née get_namespace()) now works correctly with numpy.array_api arrays.

• array_namespace() (née get_namespace()) now raises TypeError instead of ValueError.

• Fix the torch.std wrapper.

• Add torch wrappers for ones, empty, and zeros so that shape can be passed as a keyword argument.

1.1 (2023-02-24)

Major Changes

• Added support for PyTorch.

• Add helper function size() (required if torch is used as torch.Tensor.size is a method that is incompatible with the array API .size).

• All wrapper functions that wrap existing library functions now pass through arbitrary **kwargs.

Minor Changes

• Added CI to run against the array API testsuite.

• Fix sort(stable=False) and argsort(stable=False) with CuPy.

1.0 (2022-12-05)

Major Changes

• Initial release. Includes support for NumPy and CuPy.