* This release supports Python versions 3.11-3.14
# 2.3.0
* Interactive examples in the NumPy documentation.
* Building NumPy with OpenMP Parallelization.
* Improved support for free threaded Python.
* Improved annotations.
* New function numpy.strings.slice
## Deprecations
* The numpy.typing.mypy_plugin has been deprecated in favor of platform-
agnostic static type inference.
* The numpy.typing.NBitBase type has been deprecated and will be removed
in a future version.
## Expired deprecations
* Removed the np.compat package source code
* Remove numpy/npy_1_7_deprecated_api.h and C macros like NPY_OWNDATA
in favor of NPY_ARRAY_OWNDATA.
* Remove alias generate_divbyzero_error to npy_set_floatstatus_divbyzero
and generate_overflow_error to npy_set_floatstatus_overflow.
* Remove np.tostring
* Raise on np.conjugate of non-numeric types
* Raise when using np.bincount(...minlength=None), use 0 instead
* Passing shape=None to functions with a non-optional shape argument errors,
use () instead
* Inexact matches for mode and searchside raise
* Setting __array_finalize__ = None errors
* np.fromfile and np.fromstring error on bad data, previously they would guess
* datetime64 and timedelta64 construction with a tuple no longer accepts an
event value, either use a two-tuple of (unit, num) or a 4-tuple of (unit, num, den, 1)
* Parsing an integer via a float string is no longer supported. To avoid this
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=188
- Update to 2.2.2
* New functions matvec and vecmat
* Many improved annotations
* Improved support for the new StringDType
* Improved support for free threaded Python
* Fixes for f2py
## Deprecations
* _add_newdoc_ufunc is now deprecated. ufunc.__doc__ = newdoc should
be used instead.
## Expired deprecations
* bool(np.array([])) and other empty arrays will now raise an error.
Use arr.size > 0 instead to check whether an array has no elements.
## Compatibility notes
* numpy.cov now properly transposes single-row (2d array) design
matrices when rowvar=False. Previously, single-row design matrices
would return a scalar in this scenario, which is not correct, so
this is a behavior change and an array of the appropriate shape
will now be returned.
## New Features
* New functions for matrix-vector and vector-matrix products
* np.complexfloating[T, T] can now also be written as np.complexfloating[T]
* UFuncs now support __dict__ attribute and allow overriding __doc__
(either directly or via ufunc.__dict__["__doc__"]). __dict__ can be
used to also override other properties, such as __module__ or
__qualname__.
* The “nbit” type parameter of np.number and its subtypes now defaults
to typing.Any. This way, type-checkers will infer annotations such
as x: np.floating as x: np.floating[Any], even in strict mode.
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=182
- update to 2.1.3:
* #27512: MAINT: prepare 2.1.x for further development
* #27537: MAINT: Bump actions/cache from 4.0.2 to 4.1.1
* #27538: MAINT: Bump pypa/cibuildwheel from 2.21.2 to 2.21.3
* #27539: MAINT: MSVC does not support #warning directive
* #27543: BUG: Fix user dtype can-cast with python scalar during promotion
* #27561: DEV: bump python to 3.12 in environment.yml
* #27562: BLD: update vendored Meson to 1.5.2
* #27563: BUG: weighted quantile for some zero weights (#27549)
* #27565: MAINT: Use miniforge for macos conda test.
* #27566: BUILD: satisfy gcc-13 pendantic errors
* #27569: BUG: handle possible error for PyTraceMallocTrack
* #27570: BLD: start building Windows free-threaded wheels [wheel build]
* #27571: BUILD: vendor tempita from Cython
* #27574: BUG: Fix warning "differs in levels of indirection" in npy_atomic.h...
* #27592: MAINT: Update Highway to latest
* #27593: BUG: Adjust numpy.i for SWIG 4.3 compatibility
* #27616: BUG: Fix Linux QEMU CI workflow
* #27668: BLD: Do not set __STDC_VERSION__ to zero during build
* #27669: ENH: fix wasm32 runtime type error in numpy._core
* #27672: BUG: Fix a reference count leak in npy_find_descr_for_scalar.
* #27673: BUG: fixes for StringDType/unicode promoters
OBS-URL: https://build.opensuse.org/request/show/1221641
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=173
* #27333: MAINT: prepare 2.1.x for further development
* #27400: BUG: apply critical sections around populating the dispatch cache
* #27406: BUG: Stub out get_build_msvc_version if distutils.msvccompiler...
* #27416: BUILD: fix missing include for std::ptrdiff_t for C++23 language...
* #27433: BLD: pin setuptools to avoid breaking numpy.distutils
* #27437: BUG: Allow unsigned shift argument for np.roll
* #27439: BUG: Disable SVE VQSort
* #27471: BUG: rfftn axis bug
* #27479: BUG: Fix extra decref of PyArray_UInt8DType.
* #27480: CI: use PyPI not scientific-python-nightly-wheels for CI doc...
* #27481: MAINT: Check for SVE support on demand
* #27484: BUG: initialize the promotion state to be weak
* #27501: MAINT: Bump pypa/cibuildwheel from 2.20.0 to 2.21.2
* #27506: BUG: avoid segfault on bad arguments in ndarray.__array_function__
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=168
- Update to 2.1.0
* Support for Python 3.13.
* Preliminary support for free threaded Python 3.13.
* Support for the array-api 2023.12 standard.
## New functions
* A new function np.unstack(array, axis=...) was added, which
splits an array into a tuple of arrays along an axis. It serves
as the inverse of numpy.stack. (gh-26579)
## Deprecations
* The fix_imports keyword argument in numpy.save is deprecated.
Since NumPy 1.17, numpy.save uses a pickle protocol that no
longer supports Python 2, and ignored fix_imports keyword. This
keyword is kept only for backward compatibility. It is now
deprecated. (gh-26452)
* Passing non-integer inputs as the first argument of bincount is
now deprecated, because such inputs are silently cast to
integers with no warning about loss of precision. (gh-27076)
## Expired deprecations
* Scalars and 0D arrays are disallowed for numpy.nonzero and
numpy.ndarray.nonzero. (gh-26268)
* set_string_function internal function was removed and
PyArray_SetStringFunction was stubbed out. (gh-26611)
## C API changes
* API symbols now hidden but customizable
* Many shims removed from npy_3kcompat.h
* New PyUFuncObject field process_core_dims_func
## New Features
* Preliminary Support for Free-Threaded CPython 3.13
* f2py can generate freethreading-compatible C extensions
## Improvements
* histogram auto-binning now returns bin sizes >=1 for integer
input data
* ndarray shape-type parameter is now covariant and bound to
tuple[int, ...]
* np.quantile with method closest_observation chooses nearest
even order statistic
* lapack_lite is now thread safe
* The numpy.printoptions context manager is now thread and
async-safe
* Type hinting numpy.polynomial
* Improved numpy.dtypes type hints
## Performance improvements and changes
* ma.cov and ma.corrcoef are now significantly faster
## Changes
* ma.corrcoef may return a slightly different result
* Cast-safety fixes in copyto and full
- Release 2.0.1
## Improvements
* np.quantile with method closest_observation chooses nearest
even order statistic
OBS-URL: https://build.opensuse.org/request/show/1198038
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=165
* Adds support for array parameter declaration in fortran...
* Added ``bitwise_count`` UFuncs
* Add binding for random pyx files
* Use AVX512-FP16 SVML content for float16 umath functions
* allow int sequences as shape arguments in numpy.memmap
* Add .mT attribute for arrays
* Create complex scalars from real and imaginary parts
* add copy parameter for api.reshape function
* make use of locals() in a comprehension fully compatible...
* Add array API standard v2022.12 support to numpy.array_api
* Change string to bool conversions to be consistent with...
* Allow np.info on non-hashable objects with a dtype
* let zeros, empty, and empty_like accept dtype classes
* Bump C-ABI to 2 but accept older NumPy if compiled against...
* Use high accuracy SVML for double precision umath functions
* expose PyUFunc_GiveFloatingpointErrors in the dtype API
* PyObject_IsTrue and PyObject_Not error handling in setflags
* array2string does not add signs for positive integers.
* Vectorize np.partition and np.argpartition using AVX-512
* Create helper for conversion to arrays
* Add size check for threaded array assignment
* Finalize ``fastCopyAndTranpose`` and other old C-funcs/members...
* assert_array_less should report max violations instead of...
* Introduce tracer for enabled CPU targets on each optimized...
* Extend np.add ufunc to work with unicode and byte dtypes
* Add find/rfind ufuncs for unicode and byte dtypes
* Make ``intp`` ``ssize_t`` and introduce characters nN
* Add isdigit/isspace/isdecimal/isnumeric ufuncs for string...
* DType API slot for descriptor finalization before array...
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=161
- Update to 1.26.4
* NumPy 1.26.4 is a maintenance release that fixes bugs and
regressions discovered after the 1.26.3 release. The Python
versions supported by this release are 3.9-3.12. This is the
last planned release in the 1.26.x series.
* BUG: Restore missing asstr import
* MAINT: prepare 1.26.x for further development
* BUG: numpy.array_api: fix linalg.cholesky upper decomp...
* MAINT, BLD: Fix unused inline functions warnings on clang
* TST: Fix test_numeric on riscv64
* MAINT: add newaxis to __all__ in numpy.array_api
* BUG: Use large file fallocate on 32 bit linux platforms
* TST: Fix test_warning_calls on Python 3.12
* TST: Bump pytz to 2023.3.post1
* BUG: Fix AVX512 build flags on Intel Classic Compiler
* BLD: fix potential issue with escape sequences in __config__.py
* BLD: unvendor meson-python on 1.26.x and upgrade to
meson-python...
* MAINT: Include header defining backtrace
* BUG: Fix np.quantile([Fraction(2,1)], 0.5) (#24711)
- Release 1.26.3
## Compatibility
* f2py will no longer accept ambiguous -m and .pyf CLI
combinations. When more than one .pyf file is passed, an error
is raised. When both -m and a .pyf is passed, a warning is
emitted and the -m provided name is ignored.
## Improvements
* f2py now handles common blocks which have kind specifications
from modules. This further expands the usability of intrinsics
like iso_fortran_env and iso_c_binding.
## Pull requests merged
* MAINT: prepare 1.26.x for further development
* TYP: add None to __getitem__ in numpy.array_api
* BLD,BUG: quadmath required where available [f2py]
* BUG: alpha doesn't use REAL(10)
* BUG: Fix FP overflow error in division when the divisor is
scalar
* MAINT: Pin scipy-openblas version.
* BUG: Fix f2py to enable use of string optional inout argument
* BUG: Fix -fsanitize=alignment issue in
numpy/_core/src/multiarray/arraytypes.c.src
* TST: Explicitly pass NumPy path to cython during tests (also...
* BUG: fix issues with newaxis and linalg.solve in
numpy.array_api
* BUG: Disallow shadowed modulenames
* BUG: Handle common blocks with kind specifications from modules
* BUG: Fix moving compiled executable to root with f2py -c on
Windows
* BUG: Fix single to half-precision conversion on PPC64/VSX3
* TST: f2py: fix issue in test skip condition
* Revert "MAINT: Pin scipy-openblas version."
* MAINT: do not use long type
* TST: PyPy needs another gc.collect on latest versions
* MAINT: Bump conda-incubator/setup-miniconda from 2.2.0 to 3.0.1
* BLD: update vendored Meson for AIX shared library fix
* MAINT: Init base in cpu_avx512_kn
* BUG: Fix failing test_features on SapphireRapids
* BUG: Fix non-contiguous memory load when ARM/Neon is enabled
* MAINT,BUG: Never import distutils above 3.12 [f2py]
* MAINT: make the import-time check for old Accelerate more
specific
* MAINT: Bump actions/setup-node and
larsoner/circleci-artifacts-redirector-action
* BUG: avoid seg fault from OOB access in RandomState.set_state()
* BUG: Fix two errors related to not checking for failed
allocations
* BUG: Fix regression with f2py wrappers when modules and
subroutines...
* BUG: Fix build issues on SPR
* BLD: fix uninitialized variable warnings from
simd/neon/memory.h
* BUG: Handle iso_c_type mappings more consistently
* BUG: Fix module name bug in signature files [urgent] [f2py]
* BUG: Handle .pyf.src and fix SciPy [urgent]
* DOC: f2py rewrite with meson details
* BUG: Add external library handling for meson [f2py]
* MAINT: Run f2py's meson backend with the same python that
ran...
* MAINT: Update numpy/f2py/_backends from main.
* MAINT: Easy updates of f2py/*.py from main.
* MAINT: Update crackfortran.py and f2py2e.py from main
OBS-URL: https://build.opensuse.org/request/show/1168946
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=155
- Update to 1.26.2:
* TYP: Trim down the ``_NestedSequence.__getitem__`` signature
* BUG: fix choose refcount leak
* TST: fix running the test suite in builds without BLAS/LAPACK
* BUG: random: Fix generation of nan by dirichlet.
* TST: fix distutils tests for deprecations in recent setuptools...
* MAINT: Remove versioneer
* MAINT: Pin upper version of sphinx.
* ENH: Add prefix to _ALIGN Macro
* BUG: cleanup warnings [skip azp][skip circle][skip travis][skip...
* BUG: ``asv dev`` has been removed, use ``asv run``.
* BUG: Fix meson build failure due to unchanged inplace auto-generated...
* BUG: fix issue with git-version script, needs a shebang to run
* BUG: Use a default assignment for git_hash [skip ci]
* BUG: fix NPY_cast_info error handling in choose
* BUG: Fix common block handling in f2py
* BUG: Fix assumed length f2py regression
* MAINT: Harmonize fortranobject
* TYP: add kind argument to numpy.isin type specification
* BUG: fix comparisons between masked and unmasked structured arrays
* ENH: Adopt new macOS Accelerate BLAS/LAPACK Interfaces, including...
* TYP: Add the missing ``casting`` keyword to ``np.clip``
* TST: convert cython test from setup.py to meson
* MAINT: Fixup ``fromnumeric.pyi``
* BUG, ENH: Fix ``iso_c_binding`` type maps
* TYP: Allow ``binary_repr`` to accept any object
* TYP: Explicitly declare ``dtype`` and ``generic`` hashable
* ENH: Refactor the typing "reveal" tests using `typing.assert_type`
* ENH: ``meson`` backend for ``f2py``
* MAINT: Refactor partial load Workaround for Clang
OBS-URL: https://build.opensuse.org/request/show/1127600
OBS-URL: https://build.opensuse.org/package/show/openSUSE:Factory/python-numpy?expand=0&rev=122
* TYP: Trim down the ``_NestedSequence.__getitem__`` signature
* BUG: fix choose refcount leak
* TST: fix running the test suite in builds without BLAS/LAPACK
* BUG: random: Fix generation of nan by dirichlet.
* TST: fix distutils tests for deprecations in recent setuptools...
* MAINT: Remove versioneer
* MAINT: Pin upper version of sphinx.
* ENH: Add prefix to _ALIGN Macro
* BUG: cleanup warnings [skip azp][skip circle][skip travis][skip...
* BUG: ``asv dev`` has been removed, use ``asv run``.
* BUG: Fix meson build failure due to unchanged inplace auto-generated...
* BUG: fix issue with git-version script, needs a shebang to run
* BUG: Use a default assignment for git_hash [skip ci]
* BUG: fix NPY_cast_info error handling in choose
* BUG: Fix common block handling in f2py
* BUG: Fix assumed length f2py regression
* MAINT: Harmonize fortranobject
* TYP: add kind argument to numpy.isin type specification
* BUG: fix comparisons between masked and unmasked structured arrays
* ENH: Adopt new macOS Accelerate BLAS/LAPACK Interfaces, including...
* TYP: Add the missing ``casting`` keyword to ``np.clip``
* TST: convert cython test from setup.py to meson
* MAINT: Fixup ``fromnumeric.pyi``
* BUG, ENH: Fix ``iso_c_binding`` type maps
* TYP: Allow ``binary_repr`` to accept any object
* TYP: Explicitly declare ``dtype`` and ``generic`` hashable
* ENH: Refactor the typing "reveal" tests using `typing.assert_type`
* ENH: ``meson`` backend for ``f2py``
* MAINT: Refactor partial load Workaround for Clang
OBS-URL: https://build.opensuse.org/package/show/devel:languages:python:numeric/python-numpy?expand=0&rev=143
