Update numpy to 2.0.1

[pyup-bot]

This PR updates numpy from 1.8.0 to 2.0.1.

Changelog

2.0

-   `npy_interrupt.h` and the corresponding macros like `NPY_SIGINT_ON`
 have been removed. We recommend querying `PyErr_CheckSignals()` or
 `PyOS_InterruptOccurred()` periodically (these do currently require
 holding the GIL though).

-   The `noprefix.h` header has been removed. Replace missing symbols
 with their prefixed counterparts (usually an added `NPY_` or
 `npy_`).

 ([gh-23919](https://github.com/numpy/numpy/pull/23919))

-   `PyUFunc_GetPyVals`, `PyUFunc_handlefperr`, and `PyUFunc_checkfperr`
 have been removed. If needed, a new backwards compatible function to
 raise floating point errors could be restored. Reason for removal:
 there are no known users and the functions would have made
 `with np.errstate()` fixes much more difficult).

 ([gh-23922](https://github.com/numpy/numpy/pull/23922))

-   The `numpy/old_defines.h` which was part of the API deprecated since
 NumPy 1.7 has been removed. This removes macros of the form
 `PyArray_CONSTANT`. The
 [replace_old_macros.sed](https://github.com/numpy/numpy/blob/main/tools/replace_old_macros.sed)
 script may be useful to convert them to the `NPY_CONSTANT` version.

 ([gh-24011](https://github.com/numpy/numpy/pull/24011))

-   The `legacy_inner_loop_selector` member of the ufunc struct is
 removed to simplify improvements to the dispatching system. There
 are no known users overriding or directly accessing this member.

 ([gh-24271](https://github.com/numpy/numpy/pull/24271))

-   `NPY_INTPLTR` has been removed to avoid confusion (see `intp`
 redefinition).

 ([gh-24888](https://github.com/numpy/numpy/pull/24888))

-   The advanced indexing `MapIter` and related API has been removed.
 The (truly) public part of it was not well tested and had only one
 known user (Theano). Making it private will simplify improvements to
 speed up `ufunc.at`, make advanced indexing more maintainable, and
 was important for increasing the maximum number of dimensions of
 arrays to 64. Please let us know if this API is important to you so
 we can find a solution together.

 ([gh-25138](https://github.com/numpy/numpy/pull/25138))

-   The `NPY_MAX_ELSIZE` macro has been removed, as it only ever
 reflected builtin numeric types and served no internal purpose.

 ([gh-25149](https://github.com/numpy/numpy/pull/25149))

-   `PyArray_REFCNT` and `NPY_REFCOUNT` are removed. Use `Py_REFCNT`
 instead.

 ([gh-25156](https://github.com/numpy/numpy/pull/25156))

-   `PyArrayFlags_Type` and `PyArray_NewFlagsObject` as well as
 `PyArrayFlagsObject` are private now. There is no known use-case;
 use the Python API if needed.

-   `PyArray_MoveInto`, `PyArray_CastTo`, `PyArray_CastAnyTo` are
 removed use `PyArray_CopyInto` and if absolutely needed
 `PyArray_CopyAnyInto` (the latter does a flat copy).

-   `PyArray_FillObjectArray` is removed, its only true use was for
 implementing `np.empty`. Create a new empty array or use
 `PyArray_FillWithScalar()` (decrefs existing objects).

-   `PyArray_CompareUCS4` and `PyArray_CompareString` are removed. Use
 the standard C string comparison functions.

-   `PyArray_ISPYTHON` is removed as it is misleading, has no known
 use-cases, and is easy to replace.

-   `PyArray_FieldNames` is removed, as it is unclear what it would be
 useful for. It also has incorrect semantics in some possible
 use-cases.

-   `PyArray_TypestrConvert` is removed, since it seems a misnomer and
 unlikely to be used by anyone. If you know the size or are limited
 to few types, just use it explicitly, otherwise go via Python
 strings.

 ([gh-25292](https://github.com/numpy/numpy/pull/25292))

-   `PyDataType_GetDatetimeMetaData` is removed, it did not actually do
 anything since at least NumPy 1.7.

 ([gh-25802](https://github.com/numpy/numpy/pull/25802))

-   `PyArray_GetCastFunc` is removed. Note that custom legacy user
 dtypes can still provide a castfunc as their implementation, but any
 access to them is now removed. The reason for this is that NumPy
 never used these internally for many years. If you use simple
 numeric types, please just use C casts directly. In case you require
 an alternative, please let us know so we can create new API such as
 `PyArray_CastBuffer()` which could use old or new cast functions
 depending on the NumPy version.

 ([gh-25161](https://github.com/numpy/numpy/pull/25161))

New Features

`np.add` was extended to work with `unicode` and `bytes` dtypes.

> ([gh-24858](https://github.com/numpy/numpy/pull/24858))

A new `bitwise_count` function

This new function counts the number of 1-bits in a number.
`numpy.bitwise_count` works on all the numpy integer types
and integer-like objects.

python
>>> a = np.array([2**i - 1 for i in range(16)])
>>> np.bitwise_count(a)
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15],
   dtype=uint8)


([gh-19355](https://github.com/numpy/numpy/pull/19355))

macOS Accelerate support, including the ILP64

Support for the updated Accelerate BLAS/LAPACK library, including ILP64
(64-bit integer) support, in macOS 13.3 has been added. This brings
arm64 support, and significant performance improvements of up to 10x for
commonly used linear algebra operations. When Accelerate is selected at
build time, or if no explicit BLAS library selection is done, the 13.3+
version will automatically be used if available.

([gh-24053](https://github.com/numpy/numpy/pull/24053))

Binary wheels are also available. On macOS \>=14.0, users who install
NumPy from PyPI will get wheels built against Accelerate rather than
OpenBLAS.

([gh-25255](https://github.com/numpy/numpy/pull/25255))

Option to use weights for quantile and percentile functions

A `weights` keyword is now available for `numpy.quantile`, `numpy.percentile`,
`numpy.nanquantile` and `numpy.nanpercentile`. Only `method="inverted_cdf"`
supports weights.

([gh-24254](https://github.com/numpy/numpy/pull/24254))

Improved CPU optimization tracking

A new tracer mechanism is available which enables tracking of the
enabled targets for each optimized function (i.e., that uses
hardware-specific SIMD instructions) in the NumPy library. With this
enhancement, it becomes possible to precisely monitor the enabled CPU
dispatch targets for the dispatched functions.

A new function named `opt_func_info` has been added to the new namespace
`numpy.lib.introspect`, offering this tracing capability.  This function allows
you to retrieve information about the enabled targets based on function names
and data type signatures.

([gh-24420](https://github.com/numpy/numpy/pull/24420))

A new Meson backend for `f2py`

`f2py` in compile mode (i.e. `f2py -c`) now accepts the
`--backend meson` option. This is the default option for Python \>=3.12.
For older Python versions, `f2py` will still default to
`--backend distutils`.

To support this in realistic use-cases, in compile mode `f2py` takes a
`--dep` flag one or many times which maps to `dependency()` calls in the
`meson` backend, and does nothing in the `distutils` backend.

There are no changes for users of `f2py` only as a code generator, i.e.
without `-c`.

([gh-24532](https://github.com/numpy/numpy/pull/24532))

`bind(c)` support for `f2py`

Both functions and subroutines can be annotated with `bind(c)`. `f2py`
will handle both the correct type mapping, and preserve the unique label
for other C interfaces.

**Note:** `bind(c, name = 'routine_name_other_than_fortran_routine')` is
not honored by the `f2py` bindings by design, since `bind(c)` with the
`name` is meant to guarantee only the same name in C and Fortran, not in
Python and Fortran.

([gh-24555](https://github.com/numpy/numpy/pull/24555))

A new `strict` option for several testing functions

The `strict` keyword is now available for `numpy.testing.assert_allclose`,
`numpy.testing.assert_equal`, and `numpy.testing.assert_array_less`. Setting
`strict=True` will disable the broadcasting behaviour for scalars and ensure
that input arrays have the same data type.

([gh-24680](https://github.com/numpy/numpy/pull/24680),
[gh-24770](https://github.com/numpy/numpy/pull/24770),
[gh-24775](https://github.com/numpy/numpy/pull/24775))

Add `np.core.umath.find` and `np.core.umath.rfind` UFuncs

Add two `find` and `rfind` UFuncs that operate on unicode or byte
strings and are used in `np.char`. They operate similar to `str.find`
and `str.rfind`.

([gh-24868](https://github.com/numpy/numpy/pull/24868))

`diagonal` and `trace` for `numpy.linalg`

`numpy.linalg.diagonal` and `numpy.linalg.trace` have been added, which are
array API standard-compatible variants of `numpy.diagonal` and `numpy.trace`.
They differ in the default axis selection which define 2-D sub-arrays.

([gh-24887](https://github.com/numpy/numpy/pull/24887))

New `long` and `ulong` dtypes

`numpy.long` and `numpy.ulong` have been added as NumPy integers mapping to
C\'s `long` and `unsigned long`. Prior to NumPy 1.24, `numpy.long` was an alias
to Python\'s `int`.

([gh-24922](https://github.com/numpy/numpy/pull/24922))

`svdvals` for `numpy.linalg`

`numpy.linalg.svdvals` has been added. It computes singular values for (a stack
of) matrices. Executing `np.svdvals(x)` is the same as calling `np.svd(x,
compute_uv=False, hermitian=False)`. This function is compatible with the array
API standard.

([gh-24940](https://github.com/numpy/numpy/pull/24940))

A new `isdtype` function

`numpy.isdtype` was added to provide a canonical way to classify NumPy\'s
dtypes in compliance with the array API standard.

([gh-25054](https://github.com/numpy/numpy/pull/25054))

A new `astype` function

`numpy.astype` was added to provide an array API standard-compatible
alternative to the `numpy.ndarray.astype` method.

([gh-25079](https://github.com/numpy/numpy/pull/25079))

Array API compatible functions\' aliases

13 aliases for existing functions were added to improve compatibility
with the array API standard:

-   Trigonometry: `acos`, `acosh`, `asin`, `asinh`, `atan`, `atanh`,
 `atan2`.
-   Bitwise: `bitwise_left_shift`, `bitwise_invert`,
 `bitwise_right_shift`.
-   Misc: `concat`, `permute_dims`, `pow`.
-   In `numpy.linalg`: `tensordot`, `matmul`.

([gh-25086](https://github.com/numpy/numpy/pull/25086))

New `unique_*` functions

The `numpy.unique_all`, `numpy.unique_counts`, `numpy.unique_inverse`, and
`numpy.unique_values` functions have been added. They provide functionality of
`numpy.unique` with different sets of flags. They are array API
standard-compatible, and because the number of arrays they return does not
depend on the values of input arguments, they are easier to target for JIT
compilation.

([gh-25088](https://github.com/numpy/numpy/pull/25088))

Matrix transpose support for ndarrays

NumPy now offers support for calculating the matrix transpose of an
array (or stack of arrays). The matrix transpose is equivalent to
swapping the last two axes of an array. Both `np.ndarray` and
`np.ma.MaskedArray` now expose a `.mT` attribute, and there is a
matching new `numpy.matrix_transpose` function.

([gh-23762](https://github.com/numpy/numpy/pull/23762))

Array API compatible functions for `numpy.linalg`

Six new functions and two aliases were added to improve compatibility
with the Array API standard for \`numpy.linalg\`:

-   `numpy.linalg.matrix_norm` - Computes the matrix norm of
 a matrix (or a stack of matrices).

-   `numpy.linalg.vector_norm` - Computes the vector norm of
 a vector (or batch of vectors).

-   `numpy.vecdot` - Computes the (vector) dot product of
 two arrays.

-   `numpy.linalg.vecdot` - An alias for
 `numpy.vecdot`.

-   `numpy.linalg.matrix_transpose` - An alias for
 `numpy.matrix_transpose`.

 ([gh-25155](https://github.com/numpy/numpy/pull/25155))

-   `numpy.linalg.outer` has been added. It computes the
 outer product of two vectors. It differs from
 `numpy.outer` by accepting one-dimensional arrays only.
 This function is compatible with the array API standard.

 ([gh-25101](https://github.com/numpy/numpy/pull/25101))

-   `numpy.linalg.cross` has been added. It computes the
 cross product of two (arrays of) 3-dimensional vectors. It differs
 from `numpy.cross` by accepting three-dimensional
 vectors only. This function is compatible with the array API
 standard.

 ([gh-25145](https://github.com/numpy/numpy/pull/25145))

A `correction` argument for `var` and `std`

A `correction` argument was added to `numpy.var` and `numpy.std`, which is an
array API standard compatible alternative to `ddof`. As both arguments serve a
similar purpose, only one of them can be provided at the same time.

([gh-25169](https://github.com/numpy/numpy/pull/25169))

`ndarray.device` and `ndarray.to_device`

An `ndarray.device` attribute and `ndarray.to_device` method were added
to `numpy.ndarray` for array API standard compatibility.

Additionally, `device` keyword-only arguments were added to:
`numpy.asarray`, `numpy.arange`, `numpy.empty`, `numpy.empty_like`,
`numpy.eye`, `numpy.full`, `numpy.full_like`, `numpy.linspace`, `numpy.ones`,
`numpy.ones_like`, `numpy.zeros`, and `numpy.zeros_like`.

For all these new arguments, only `device="cpu"` is supported.

([gh-25233](https://github.com/numpy/numpy/pull/25233))

StringDType has been added to NumPy

We have added a new variable-width UTF-8 encoded string data type, implementing
a \"NumPy array of Python strings\", including support for a user-provided
missing data sentinel. It is intended as a drop-in replacement for arrays of
Python strings and missing data sentinels using the object dtype. See 
[NEP 55](https://numpy.org/neps/nep-0055-string_dtype.html) and the documentation
of stringdtype for more details.

([gh-25347](https://github.com/numpy/numpy/pull/25347))

New keywords for `cholesky` and `pinv`

The `upper` and `rtol` keywords were added to
`numpy.linalg.cholesky` and `numpy.linalg.pinv`,
respectively, to improve array API standard compatibility.

For `numpy.linalg.pinv`, if neither `rcond` nor `rtol` is
specified, the `rcond`\'s default is used. We plan to deprecate and
remove `rcond` in the future.

([gh-25388](https://github.com/numpy/numpy/pull/25388))

New keywords for `sort`, `argsort` and `linalg.matrix_rank`

New keyword parameters were added to improve array API standard
compatibility:

-   `rtol` was added to `numpy.linalg.matrix_rank`.
-   `stable` was added to `numpy.sort` and
 `numpy.argsort`.

([gh-25437](https://github.com/numpy/numpy/pull/25437))

New `numpy.strings` namespace for string ufuncs

NumPy now implements some string operations as ufuncs. The old `np.char`
namespace is still available, and where possible the string manipulation
functions in that namespace have been updated to use the new ufuncs,
substantially improving their performance.

Where possible, we suggest updating code to use functions in
`np.strings` instead of `np.char`. In the future we may deprecate
`np.char` in favor of `np.strings`.

([gh-25463](https://github.com/numpy/numpy/pull/25463))

`numpy.fft` support for different precisions and in-place calculations

The various FFT routines in `numpy.fft` now do their
calculations natively in float, double, or long double precision,
depending on the input precision, instead of always calculating in
double precision. Hence, the calculation will now be less precise for
single and more precise for long double precision. The data type of the
output array will now be adjusted accordingly.

Furthermore, all FFT routines have gained an `out` argument that can be
used for in-place calculations.

([gh-25536](https://github.com/numpy/numpy/pull/25536))

configtool and pkg-config support

A new `numpy-config` CLI script is available that can be queried for the
NumPy version and for compile flags needed to use the NumPy C API. This
will allow build systems to better support the use of NumPy as a
dependency. Also, a `numpy.pc` pkg-config file is now included with
Numpy. In order to find its location for use with `PKG_CONFIG_PATH`, use
`numpy-config --pkgconfigdir`.

([gh-25730](https://github.com/numpy/numpy/pull/25730))

Array API standard support in the main namespace

The main `numpy` namespace now supports the array API standard. See
`array-api-standard-compatibility` for
details.

([gh-25911](https://github.com/numpy/numpy/pull/25911))

Improvements

Strings are now supported by `any`, `all`, and the logical ufuncs.

> ([gh-25651](https://github.com/numpy/numpy/pull/25651))

Integer sequences as the shape argument for `memmap`

`numpy.memmap` can now be created with any integer sequence
as the `shape` argument, such as a list or numpy array of integers.
Previously, only the types of tuple and int could be used without
raising an error.

([gh-23729](https://github.com/numpy/numpy/pull/23729))

`errstate` is now faster and context safe

The `numpy.errstate` context manager/decorator is now faster
and safer. Previously, it was not context safe and had (rare) issues
with thread-safety.

([gh-23936](https://github.com/numpy/numpy/pull/23936))

AArch64 quicksort speed improved by using Highway\'s VQSort

The first introduction of the Google Highway library, using VQSort on
AArch64. Execution time is improved by up to 16x in some cases, see the
PR for benchmark results. Extensions to other platforms will be done in
the future.

([gh-24018](https://github.com/numpy/numpy/pull/24018))

Complex types - underlying C type changes

-   The underlying C types for all of NumPy\'s complex types have been
 changed to use C99 complex types.

-   While this change does not affect the memory layout of complex
 types, it changes the API to be used to directly retrieve or write
 the real or complex part of the complex number, since direct field
 access (as in `c.real` or `c.imag`) is no longer an option. You can
 now use utilities provided in `numpy/npy_math.h` to do these
 operations, like this:

  c
 npy_cdouble c;
 npy_csetreal(&c, 1.0);
 npy_csetimag(&c, 0.0);
 printf("%d + %di\n", npy_creal(c), npy_cimag(c));
 

-   To ease cross-version compatibility, equivalent macros and a
 compatibility layer have been added which can be used by downstream
 packages to continue to support both NumPy 1.x and 2.x. See
 `complex-numbers` for more info.

-   `numpy/npy_common.h` now includes `complex.h`, which means that
 `complex` is now a reserved keyword.

([gh-24085](https://github.com/numpy/numpy/pull/24085))

`iso_c_binding` support and improved common blocks for `f2py`

Previously, users would have to define their own custom `f2cmap` file to
use type mappings defined by the Fortran2003 `iso_c_binding` intrinsic
module. These type maps are now natively supported by `f2py`

([gh-24555](https://github.com/numpy/numpy/pull/24555))

`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`.

([gh-25186](https://github.com/numpy/numpy/pull/25186))

Call `str` automatically on third argument to functions like `assert_equal`

The third argument to functions like
`numpy.testing.assert_equal` now has `str` called on it
automatically. This way it mimics the built-in `assert` statement, where
`assert_equal(a, b, obj)` works like `assert a == b, obj`.

([gh-24877](https://github.com/numpy/numpy/pull/24877))

Support for array-like `atol`/`rtol` in `isclose`, `allclose`

The keywords `atol` and `rtol` in `numpy.isclose` and
`numpy.allclose` now accept both scalars and arrays. An
array, if given, must broadcast to the shapes of the first two array
arguments.

([gh-24878](https://github.com/numpy/numpy/pull/24878))

Consistent failure messages in test functions

Previously, some `numpy.testing` assertions printed messages
that referred to the actual and desired results as `x` and `y`. Now,
these values are consistently referred to as `ACTUAL` and `DESIRED`.

([gh-24931](https://github.com/numpy/numpy/pull/24931))

n-D FFT transforms allow `s[i] == -1`

The `numpy.fft.fftn`, `numpy.fft.ifftn`,
`numpy.fft.rfftn`, `numpy.fft.irfftn`,
`numpy.fft.fft2`, `numpy.fft.ifft2`,
`numpy.fft.rfft2` and `numpy.fft.irfft2`
functions now use the whole input array along the axis `i` if
`s[i] == -1`, in line with the array API standard.

([gh-25495](https://github.com/numpy/numpy/pull/25495))

Guard PyArrayScalar_VAL and PyUnicodeScalarObject for the limited API

`PyUnicodeScalarObject` holds a `PyUnicodeObject`, which is not
available when using `Py_LIMITED_API`. Add guards to hide it and
consequently also make the `PyArrayScalar_VAL` macro hidden.

([gh-25531](https://github.com/numpy/numpy/pull/25531))

Changes

-   `np.gradient()` now returns a tuple rather than a list making the
 return value immutable.

 ([gh-23861](https://github.com/numpy/numpy/pull/23861))

-   Being fully context and thread-safe, `np.errstate` can only be
 entered once now.

-   `np.setbufsize` is now tied to `np.errstate()`: leaving an
 `np.errstate` context will also reset the `bufsize`.

 ([gh-23936](https://github.com/numpy/numpy/pull/23936))

-   A new public `np.lib.array_utils` submodule has been introduced and
 it currently contains three functions: `byte_bounds` (moved from
 `np.lib.utils`), `normalize_axis_tuple` and `normalize_axis_index`.

 ([gh-24540](https://github.com/numpy/numpy/pull/24540))

-   Introduce `numpy.bool` as the new canonical name for
 NumPy\'s boolean dtype, and make `numpy.bool\_` an alias
 to it. Note that until NumPy 1.24, `np.bool` was an alias to
 Python\'s builtin `bool`. The new name helps with array API standard
 compatibility and is a more intuitive name.

 ([gh-25080](https://github.com/numpy/numpy/pull/25080))

-   The `dtype.flags` value was previously stored as a signed integer.
 This means that the aligned dtype struct flag lead to negative flags
 being set (-128 rather than 128). This flag is now stored unsigned
 (positive). Code which checks flags manually may need to adapt. This
 may include code compiled with Cython 0.29.x.

 ([gh-25816](https://github.com/numpy/numpy/pull/25816))

Representation of NumPy scalars changed

As per NEP 51, the scalar representation has been updated to include the type
information to avoid confusion with Python scalars.

Scalars are now printed as `np.float64(3.0)` rather than just `3.0`.
This may disrupt workflows that store representations of numbers (e.g.,
to files) making it harder to read them. They should be stored as
explicit strings, for example by using `str()` or `f"{scalar!s}"`. For
the time being, affected users can use
`np.set_printoptions(legacy="1.25")` to get the old behavior (with
possibly a few exceptions). Documentation of downstream projects may
require larger updates, if code snippets are tested. We are working on
tooling for
[doctest-plus](https://github.com/scientific-python/pytest-doctestplus/issues/107)
to facilitate updates.

([gh-22449](https://github.com/numpy/numpy/pull/22449))

Truthiness of NumPy strings changed

NumPy strings previously were inconsistent about how they defined if the
string is `True` or `False` and the definition did not match the one
used by Python. Strings are now considered `True` when they are
non-empty and `False` when they are empty. This changes the following
distinct cases:

-   Casts from string to boolean were previously roughly equivalent to
 `string_array.astype(np.int64).astype(bool)`, meaning that only
 valid integers could be cast. Now a string of `"0"` will be
 considered `True` since it is not empty. If you need the old
 behavior, you may use the above step (casting to integer first) or
 `string_array == "0"` (if the input is only ever `0` or `1`). To get
 the new result on old NumPy versions use `string_array != ""`.
-   `np.nonzero(string_array)` previously ignored whitespace so that a
 string only containing whitespace was considered `False`. Whitespace
 is now considered `True`.

This change does not affect `np.loadtxt`, `np.fromstring`, or
`np.genfromtxt`. The first two still use the integer definition, while
`genfromtxt` continues to match for `"true"` (ignoring case). However,
if `np.bool_` is used as a converter the result will change.

The change does affect `np.fromregex` as it uses direct assignments.

([gh-23871](https://github.com/numpy/numpy/pull/23871))

A `mean` keyword was added to var and std function

Often when the standard deviation is needed the mean is also needed. The
same holds for the variance and the mean. Until now the mean is then
calculated twice, the change introduced here for the `numpy.var` and
`numpy.std` functions allows for passing in a precalculated mean as an keyword
argument. See the docstrings for details and an example illustrating the
speed-up.

([gh-24126](https://github.com/numpy/numpy/pull/24126))

Remove datetime64 deprecation warning when constructing with timezone

The `numpy.datetime64` method now issues a UserWarning rather than a
DeprecationWarning whenever a timezone is included in the datetime string that
is provided.

([gh-24193](https://github.com/numpy/numpy/pull/24193))

Default integer dtype is now 64-bit on 64-bit Windows

The default NumPy integer is now 64-bit on all 64-bit systems as the
historic 32-bit default on Windows was a common source of issues. Most
users should not notice this. The main issues may occur with code
interfacing with libraries written in a compiled language like C. For
more information see `migration_windows_int64`.

([gh-24224](https://github.com/numpy/numpy/pull/24224))

Renamed `numpy.core` to `numpy._core`

Accessing `numpy.core` now emits a DeprecationWarning. In practice we
have found that most downstream usage of `numpy.core` was to access
functionality that is available in the main `numpy` namespace. If for
some reason you are using functionality in `numpy.core` that is not
available in the main `numpy` namespace, this means you are likely using
private NumPy internals. You can still access these internals via
`numpy._core` without a deprecation warning but we do not provide any
backward compatibility guarantees for NumPy internals. Please open an
issue if you think a mistake was made and something needs to be made
public.

([gh-24634](https://github.com/numpy/numpy/pull/24634))

The \"relaxed strides\" debug build option, which was previously enabled
through the `NPY_RELAXED_STRIDES_DEBUG` environment variable or the
`-Drelaxed-strides-debug` config-settings flag has been removed.

([gh-24717](https://github.com/numpy/numpy/pull/24717))

Redefinition of `np.intp`/`np.uintp` (almost never a change)

Due to the actual use of these types almost always matching the use of
`size_t`/`Py_ssize_t` this is now the definition in C. Previously, it
matched `intptr_t` and `uintptr_t` which would often have been subtly
incorrect. This has no effect on the vast majority of machines since the
size of these types only differ on extremely niche platforms.

However, it means that:

-   Pointers may not necessarily fit into an `intp` typed array anymore.
 The `p` and `P` character codes can still be used, however.
-   Creating `intptr_t` or `uintptr_t` typed arrays in C remains
 possible in a cross-platform way via `PyArray_DescrFromType('p')`.
-   The new character codes `nN` were introduced.
-   It is now correct to use the Python C-API functions when parsing to
 `npy_intp` typed arguments.

([gh-24888](https://github.com/numpy/numpy/pull/24888))

`numpy.fft.helper` made private

`numpy.fft.helper` was renamed to `numpy.fft._helper` to indicate that
it is a private submodule. All public functions exported by it should be
accessed from `numpy.fft`.

([gh-24945](https://github.com/numpy/numpy/pull/24945))

`numpy.linalg.linalg` made private

`numpy.linalg.linalg` was renamed to `numpy.linalg._linalg` to indicate
that it is a private submodule. All public functions exported by it
should be accessed from `numpy.linalg`.

([gh-24946](https://github.com/numpy/numpy/pull/24946))

Out-of-bound axis not the same as `axis=None`

In some cases `axis=32` or for concatenate any large value was the same
as `axis=None`. Except for `concatenate` this was deprecate. Any out of
bound axis value will now error, make sure to use `axis=None`.

([gh-25149](https://github.com/numpy/numpy/pull/25149))

New `copy` keyword meaning for `array` and `asarray` constructors

Now `numpy.array` and `numpy.asarray` support
three values for `copy` parameter:

-   `None` - A copy will only be made if it is necessary.
-   `True` - Always make a copy.
-   `False` - Never make a copy. If a copy is required a `ValueError` is
 raised.

The meaning of `False` changed as it now raises an exception if a copy
is needed.

([gh-25168](https://github.com/numpy/numpy/pull/25168))

The `__array__` special method now takes a `copy` keyword argument.

NumPy will pass `copy` to the `__array__` special method in situations
where it would be set to a non-default value (e.g. in a call to
`np.asarray(some_object, copy=False)`). Currently, if an unexpected
keyword argument error is raised after this, NumPy will print a warning
and re-try without the `copy` keyword argument. Implementations of
objects implementing the `__array__` protocol should accept a `copy`
keyword argument with the same meaning as when passed to
`numpy.array` or `numpy.asarray`.

([gh-25168](https://github.com/numpy/numpy/pull/25168))

Cleanup of initialization of `numpy.dtype` with strings with commas

The interpretation of strings with commas is changed slightly, in that a
trailing comma will now always create a structured dtype. E.g., where
previously `np.dtype("i")` and `np.dtype("i,")` were treated as
identical, now `np.dtype("i,")` will create a structured dtype, with a
single field. This is analogous to `np.dtype("i,i")` creating a
structured dtype with two fields, and makes the behaviour consistent
with that expected of tuples.

At the same time, the use of single number surrounded by parenthesis to
indicate a sub-array shape, like in `np.dtype("(2)i,")`, is deprecated.
Instead; one should use `np.dtype("(2,)i")` or `np.dtype("2i")`.
Eventually, using a number in parentheses will raise an exception, like
is the case for initializations without a comma, like
`np.dtype("(2)i")`.

([gh-25434](https://github.com/numpy/numpy/pull/25434))

Change in how complex sign is calculated

Following the array API standard, the complex sign is now calculated as
`z / |z|` (instead of the rather less logical case where the sign of the
real part was taken, unless the real part was zero, in which case the
sign of the imaginary part was returned). Like for real numbers, zero is
returned if `z==0`.

([gh-25441](https://github.com/numpy/numpy/pull/25441))

Return types of functions that returned a list of arrays

Functions that returned a list of ndarrays have been changed to return a
tuple of ndarrays instead. Returning tuples consistently whenever a
sequence of arrays is returned makes it easier for JIT compilers like
Numba, as well as for static type checkers in some cases, to support
these functions. Changed functions are: `numpy.atleast_1d`, `numpy.atleast_2d`,
`numpy.atleast_3d`, `numpy.broadcast_arrays`, `numpy.meshgrid`,
`numpy.ogrid`, `numpy.histogramdd`.

`np.unique` `return_inverse` shape for multi-dimensional inputs

When multi-dimensional inputs are passed to `np.unique` with
`return_inverse=True`, the `unique_inverse` output is now shaped such
that the input can be reconstructed directly using
`np.take(unique, unique_inverse)` when `axis=None`, and
`np.take_along_axis(unique, unique_inverse, axis=axis)` otherwise.

([gh-25553](https://github.com/numpy/numpy/pull/24126),
[gh-25570](https://github.com/numpy/numpy/pull/25570))

`any` and `all` return booleans for object arrays

The `any` and `all` functions and methods now return booleans also for
object arrays. Previously, they did a reduction which behaved like the
Python `or` and `and` operators which evaluates to one of the arguments.
You can use `np.logical_or.reduce` and `np.logical_and.reduce` to
achieve the previous behavior.

([gh-25712](https://github.com/numpy/numpy/pull/25712))

**Content from release note snippets in doc/release/upcoming_changes:**

Checksums

MD5

 b2f97f907cc640f5f619ea4ebd1231d3  numpy-2.0.0b1-cp310-cp310-macosx_10_9_x86_64.whl
 db158043b6fad6e523e23b3eb2de5d88  numpy-2.0.0b1-cp310-cp310-macosx_11_0_arm64.whl
 39086961c062d97c5b42da057b9b1947  numpy-2.0.0b1-cp310-cp310-macosx_14_0_arm64.whl
 3362d35bf69b852b98b41b8373253a0f  numpy-2.0.0b1-cp310-cp310-macosx_14_0_x86_64.whl
 66e907969e32ec43e887cabcc1884763  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 b8d1bece144e3b6aae641d44821f815f  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 96ab156ec312bb451e8c5e19de4a28b7  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_aarch64.whl
 c04819a4f3395b81d124ffc6330925e9  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_x86_64.whl
 6af68b8eb8fe583ffabab9bd7da1c620  numpy-2.0.0b1-cp310-cp310-win32.whl
 3b8a9514e5795985bcba20e213d55b54  numpy-2.0.0b1-cp310-cp310-win_amd64.whl
 0128ad9249f70d97a057a23e0cef1515  numpy-2.0.0b1-cp311-cp311-macosx_10_9_x86_64.whl
 612c018a7676ce3747cb863762750e1d  numpy-2.0.0b1-cp311-cp311-macosx_11_0_arm64.whl
 6b1480446aff53c71c903fc1248bca94  numpy-2.0.0b1-cp311-cp311-macosx_14_0_arm64.whl
 8d66a0af99edf30dc9de487b3f8c1639  numpy-2.0.0b1-cp311-cp311-macosx_14_0_x86_64.whl
 f9154a0885b2647d7e81f32900390ebb  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 9dd14e2b594a2d47eb25ecc759d5adaa  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 8434d07fc4eb80c5df9ae5ebf95546eb  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_aarch64.whl
 a0402697c93a9d6bc8d979fabd6bf179  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_x86_64.whl
 2ba67ffb4b92b54394b6929b3a899cb2  numpy-2.0.0b1-cp311-cp311-win32.whl
 d75e2f02c698e492b7b07f0659f9bbe4  numpy-2.0.0b1-cp311-cp311-win_amd64.whl
 558fefd135de6fcebe2b94d857a84c32  numpy-2.0.0b1-cp312-cp312-macosx_10_9_x86_64.whl
 d684790e4509e7daa99a1aef1d0be536  numpy-2.0.0b1-cp312-cp312-macosx_11_0_arm64.whl
 fd5d4f1d1da0cc685c54e9abd2f9dceb  numpy-2.0.0b1-cp312-cp312-macosx_14_0_arm64.whl
 65183c1302348d3db60eaf3b62c1e577  numpy-2.0.0b1-cp312-cp312-macosx_14_0_x86_64.whl
 305eaf68e214011557303988f4635271  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 e3a84e27effd888cf93eb2c1aad759e7  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 88eef47ecdd11ac0939291abe0c74b6f  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_aarch64.whl
 fd390078c0046c20a659035c1826185f  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_x86_64.whl
 23db11989d2d0086ff12655355245a2a  numpy-2.0.0b1-cp312-cp312-win32.whl
 323d05ef29a9c8166d865ab221faf7dc  numpy-2.0.0b1-cp312-cp312-win_amd64.whl
 f5ad7adf599b65050ccd116802f0265d  numpy-2.0.0b1-cp39-cp39-macosx_10_9_x86_64.whl
 89a94dddb18e4210e01ee6ca24012fcb  numpy-2.0.0b1-cp39-cp39-macosx_11_0_arm64.whl
 409a537dc5ea249b3e6868dd37932342  numpy-2.0.0b1-cp39-cp39-macosx_14_0_arm64.whl
 0db893de846425d58b90f05c1db3d191  numpy-2.0.0b1-cp39-cp39-macosx_14_0_x86_64.whl
 c73ba41d166a5f2e72cdc48b8554c6e6  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 786236fc9099283255133273535b8de0  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 f2e8717957a6b3b37f881e8939a2af37  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_aarch64.whl
 dad671b45f6e13c28ead06064b03eaee  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_x86_64.whl
 76f8f89ff91d06df684cf47d7ea6d8ab  numpy-2.0.0b1-cp39-cp39-win32.whl
 d4dcbd6157783aa0e78710549f13876f  numpy-2.0.0b1-cp39-cp39-win_amd64.whl
 41a13de3afff77390b0d1ea3c7e407db  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl
 fc2ff82233376f853161c7f9bc6d44b7  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_arm64.whl
 860609ee9f1f24d4f28fbbcf3d31cdc9  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_x86_64.whl
 2a97175cec7a5b1280ed2a991fea23ff  numpy-2.0.0b1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 1656013175e650e053c15fd886be58f1  numpy-2.0.0b1-pp39-pypy39_pp73-win_amd64.whl
 c06e95d7cadfa33a1f4549c9a5dcba05  numpy-2.0.0b1.tar.gz

SHA256

 411ed8eb48eb679fc732f22e90c9adb994ec6ad2d9c2f53593325a975f9fa501  numpy-2.0.0b1-cp310-cp310-macosx_10_9_x86_64.whl
 f8aca0561166702070ea9abcafd70da44df48be70d16f0a886e359127436fdcc  numpy-2.0.0b1-cp310-cp310-macosx_11_0_arm64.whl
 0d217dae0f20a3400c1d80aa8401af9de93b9bb4ea7518b8ba200ff8ff62529e  numpy-2.0.0b1-cp310-cp310-macosx_14_0_arm64.whl
 824351cb4cce66c1f8e16c1698c01de8d5e4197461f78197c327281f107fc1b2  numpy-2.0.0b1-cp310-cp310-macosx_14_0_x86_64.whl
 cae0959a4f5a9c16896a87a43c9e81384f48b69f835f55050948071488820486  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 3d47a42c1e48e46dbbe32e0395f8aa6e8ddd251771ed9ec47fc07aa89b8aac89  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 909024f923c019d1b9dca16871844f1c313c422bd430a0b7e4a24a3acb766483  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_aarch64.whl
 fc6e82bea99727aeed964808f26bed95323825a75e94c015eb913fb6ec3dbdf8  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_x86_64.whl
 36862cad55650afbcb3f0e3a5edc07ba4c1090eb649208a41fadcf82cf1b2966  numpy-2.0.0b1-cp310-cp310-win32.whl
 0e6a63c725143a6be0e48effcf01b8361b80ab20e2444704356f9d9db48ba429  numpy-2.0.0b1-cp310-cp310-win_amd64.whl
 e6c3ba4bcb6cf3fd4ace244075fa214b4f0c090f12437378200a2de68144c166  numpy-2.0.0b1-cp311-cp311-macosx_10_9_x86_64.whl
 89bbb14534e53c6175aabc8449a8bdf83f02da62f13d1b5facbb2fd1fecae2e2  numpy-2.0.0b1-cp311-cp311-macosx_11_0_arm64.whl
 b14b6e6ca51afdcfc589cb9d6fb73aedf38009a1a0ecab15f77e3d0e0754cac0  numpy-2.0.0b1-cp311-cp311-macosx_14_0_arm64.whl
 ffef68423c1edc5d10321f9787fb9d8c20a36fc08ffdba863d103924d02dadce  numpy-2.0.0b1-cp311-cp311-macosx_14_0_x86_64.whl
 7e8725313b8a8aaa9cfac450713b1a74a8d79ae010ee0d0dd97505abf54d247b  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 d83e18f1c4164dbcaa01adc8f4a3aebc3c5fa635d2009d8dc1bf53dd7eab0063  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 91e37a5bb38c11bde547aefeb79dd382b5d9d1d140931927bca46c9d198e08f3  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_aarch64.whl
 d39f1005a627c5960f67b02c1f76f265e0d4219b6d7948a7809dc14443fcbeb6  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_x86_64.whl
 3ed4afbcdb8db622b90ef33bf0c0d080f287ec590032f9033be5cbc51e005b66  numpy-2.0.0b1-cp311-cp311-win32.whl
 941382abe21d26222310275a91f053386450b5364f1307641d03babfec5b1931  numpy-2.0.0b1-cp311-cp311-win_amd64.whl
 a78a38ff86aa651534979d597fdb178c7ae2c9934d95bcc921971ceea14ef54a  numpy-2.0.0b1-cp312-cp312-macosx_10_9_x86_64.whl
 e5222fb05011c310d294c40e2b8640c9351aaf3238c0605486a3f041a7befabd  numpy-2.0.0b1-cp312-cp312-macosx_11_0_arm64.whl
 0f69c008a8533879ea0480fe11b28154c0dc12567522406f2c887bc549a98865  numpy-2.0.0b1-cp312-cp312-macosx_14_0_arm64.whl
 a5b47099876fceefb5ac4d2cfe4ee7337de22253aafe6f2e545b84d100bf9e22  numpy-2.0.0b1-cp312-cp312-macosx_14_0_x86_64.whl
 a81816e4dc75351dd1ce2d84f381856b8962eef1757ddfe13007d2a8bb966fda  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 3ecb219af16b0dbf58bbe1fdb4d074582f9a99567d85c630cf82c3b40168a15d  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 7479d8f43dd78a3bd1c8a3c7c9b06e71639c480a0223c31a4aeb2c7e8fd62151  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_aarch64.whl
 1665b832541449c7079ee9d41f334ab832a1d84511cc834c0bc8d98bf96d1df5  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_x86_64.whl
 585471edf1f205fb589632581cc7b30c6c0e78d79b3c754739bb62ff568fa587  numpy-2.0.0b1-cp312-cp312-win32.whl
 881df25d857873947d54dbed01d98c417f3feb5df86ece719eebf1edbbb2095c  numpy-2.0.0b1-cp312-cp312-win_amd64.whl
 797dc478feed31f78bca1c69d9a167c6294599927c184f4e9b569ad8895ca6e5  numpy-2.0.0b1-cp39-cp39-macosx_10_9_x86_64.whl
 49cb06682f4588c2553a63445b7e37aec731452fe380c3bd142377783a9ba014  numpy-2.0.0b1-cp39-cp39-macosx_11_0_arm64.whl
 5fd7ec50b9650ac0aa4fd318eceb9059ed3c0ab3aa79d5f260a10158521f9770  numpy-2.0.0b1-cp39-cp39-macosx_14_0_arm64.whl
 72526252a5d1da5067181bfd3df9cc6d7dcd024b757f5d35e8f1d0c08cb729c1  numpy-2.0.0b1-cp39-cp39-macosx_14_0_x86_64.whl
 7870b854823217f34e6258328f46e40f68784f61408deb37a29ca64762c60c10  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
 26a0978595ac2e8160d27f7537ff94402eaaf3ea7a768e7f99170ed91453d1bf  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 5801a93e424c12366d8b0b411dfeb7102f7429f0934059a39b1529f02ea2606b  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_aarch64.whl
 2f67038ecdf4b372d81fa00530547a5d04b77da5b1e4fc55f58021f3135331ea  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_x86_64.whl
 f32b6ec16518b3ba1a2d3a100d9b413cf24aaeeefdec19f1cddec55cb4a31dac  numpy-2.0.0b1-cp39-cp39-win32.whl
 70a22408ed088725fe44a6f55a077d1f704977b262e53d30ba485a01229028a3  numpy-2.0.0b1-cp39-cp39-win_amd64.whl
 a69f1624d036953f3f2795f22e6be452ee6d24937ae14f77c2e536589e20caa2  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl
 c27970540ee6b4b8325779cd22eee0283cb9dc6511130ff54e774fcd0a261d4b  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_arm64.whl
 393adcc241ff7010b43e4660710a43c322189ff67461afba18bbaf9f5581b221  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_x86_64.whl
 4205b3efa27b74cb096443bdda178f5032ffc6b41306a7d4a0b903b4b614b146  numpy-2.0.0b1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
 52f9cd632f9f5e179e98769d76702ce9a307439f36191607d5ee06cb8a986d01  numpy-2.0.0b1-pp39-pypy39_pp73-win_amd64.whl
 e0bb33a37d0d0b9a19cd41a093877f830e06bd4d989341b9792896cf08e629f7  numpy-2.0.0b1.tar.gz

2.0.0

numpy.argsort` and `numpy.argpartition`.

Removed ambiguity when broadcasting in `np.solve`

The broadcasting rules for `np.solve(a, b)` were ambiguous when `b` had
1 fewer dimensions than `a`. This has been resolved in a
backward-incompatible way and is now compliant with the Array API. The
old behaviour can be reconstructed by using
`np.solve(a, b[..., None])[..., 0]`.

([gh-25914](https://github.com/numpy/numpy/pull/25914))

Modified representation for `Polynomial`

The representation method for
`numpy.polynomial.polynomial.Polynomial` was updated to
include the domain in the representation. The plain text and latex
representations are now consistent. For example the output of
`str(np.polynomial.Polynomial([1, 1], domain=[.1, .2]))` used to be
`1.0 + 1.0 x`, but now is `1.0 + 1.0 (-3.0000000000000004 + 20.0 x)`.

([gh-21760](https://github.com/numpy/numpy/pull/21760))

C API changes

-   The `PyArray_CGT`, `PyArray_CLT`, `PyArray_CGE`, `PyArray_CLE`,
 `PyArray_CEQ`, `PyArray_CNE` macros have been removed.

-   `PyArray_MIN` and `PyArray_MAX` have been moved from
 `ndarraytypes.h` to `npy_math.h`.

 ([gh-24258](https://github.com/numpy/numpy/pull/24258))

-   A C API for working with `numpy.dtypes.StringDType`
 arrays has been exposed. This includes functions for acquiring and
 releasing mutexes which lock access to the string data, as well as
 packing and unpacking UTF-8 bytestreams from array entries.

-   `NPY_NTYPES` has been renamed to `NPY_NTYPES_LEGACY` as it does not
 include new NumPy built-in DTypes. In particular the new string
 DType will likely not work correctly with code that handles legacy
 DTypes.

 ([gh-25347](https://github.com/numpy/numpy/pull/25347))

-   The C-API now only exports the static inline function versions of
 the array accessors (previously this depended on using \"deprecated
 API\"). While we discourage it, the struct fields can still be used
 directly.

 ([gh-25789](https://github.com/numpy/numpy/pull/25789))

-   NumPy now defines `PyArray_Pack` to set an individual memory address.
 Unlike `PyArray_SETITEM` this function is equivalent to setting an
 individual array item and does not require a NumPy array input.

 ([gh-25954](https://github.com/numpy/numpy/pull/25954))

-   The `->f` slot has been removed from `PyArray_Descr`. If you use this slot,
 replace accessing it with `PyDataType_GetArrFuncs` (see its documentation
 and the `numpy-2-migration-guide`). In some cases using other functions
 like `PyArray_GETITEM` may be an alternatives.

-   `PyArray_GETITEM` and `PyArray_SETITEM` now require the import of
 the NumPy API table to be used and are no longer defined in
 `ndarraytypes.h`.

 ([gh-25812](https://github.com/numpy/numpy/pull/25812))

-   Due to runtime dependencies, the definition for functionality
 accessing the dtype flags was moved from `numpy/ndarraytypes.h` and
 is only available after including `numpy/ndarrayobject.h` as it
 requires `import_array()`. This includes `PyDataType_FLAGCHK`,
 `PyDataType_REFCHK` and `NPY_BEGIN_THREADS_DESCR`.

-   The dtype flags on `PyArray_Descr` must now be accessed through the
 `PyDataType_FLAGS` inline function to be compatible with both 1.x
 and 2.x. This function is defined in `npy_2_compat.h` to allow
 backporting. Most or all users should use `PyDataType_FLAGCHK` which
 is available on 1.x and does not require backporting. Cython users
 should use Cython 3. Otherwise access will go through Python unless
 they use `PyDataType_FLAGCHK` instead.

 ([gh-25816](https://github.com/numpy/numpy/pull/25816))

Datetime functionality exposed in the C API and Cython bindings

The functions `NpyDatetime_ConvertDatetime64ToDatetimeStruct`,
`NpyDatetime_ConvertDatetimeStructToDatetime64`,
`NpyDatetime_ConvertPyDateTimeToDatetimeStruct`,
`NpyDatetime_GetDatetimeISO8601StrLen`,
`NpyDatetime_MakeISO8601Datetime`, and
`NpyDatetime_ParseISO8601Datetime` have been added to the C API to
facilitate converting between strings, Python datetimes, and NumPy
datetimes in external libraries.

([gh-21199](https://github.com/numpy/numpy/pull/21199))

Const correctness for the generalized ufunc C API

The NumPy C API\'s functions for constructing generalized ufuncs
(`PyUFunc_FromFuncAndData`, `PyUFunc_FromFuncAndDataAndSignature`,
`PyUFunc_FromFuncAndDataAndSignatureAndIdentity`) take `types` and
`data` arguments that are not modified by NumPy\'s internals. Like the
`name` and `doc` arguments, third-party Python extension modules are
likely to supply these arguments from static constants. The `types` and
`data` arguments are now const-correct: they are declared as
`const char *types` and `void *const *data`, respectively. C code should
not be affected, but C++ code may be.

([gh-23847](https://github.com/numpy/numpy/pull/23847))

Larger `NPY_MAXDIMS` and `NPY_MAXARGS`, `NPY_RAVEL_AXIS` introduced

`NPY_MAXDIMS` is now 64, you may want to review its use. This is usually
used in a stack allocation, where the increase should be safe. However,
we do encourage generally to remove any use of `NPY_MAXDIMS` and
`NPY_MAXARGS` to eventually allow removing the constraint completely.
For the conversion helper and C-API functions mirroring Python ones such as
`take`, `NPY_MAXDIMS` was used to mean `axis=None`. Such usage must be replaced
with `NPY_RAVEL_AXIS`. See also `migration_maxdims`.

([gh-25149](https://github.com/numpy/numpy/pull/25149))

`NPY_MAXARGS` not constant and `PyArrayMultiIterObject` size change

Since `NPY_MAXARGS` was increased, it is now a runtime constant and not
compile-time constant anymore. We expect almost no users to notice this.
But if used for stack allocations it now must be replaced with a custom
constant using `NPY_MAXARGS` as an additional runtime check.

The `sizeof(PyArrayMultiIterObject)` no longer includes the full size of
the object. We expect nobody to notice this change. It was necessary to
avoid issues with Cython.

([gh-25271](https://github.com/numpy/numpy/pull/25271))

Required changes for custom legacy user dtypes

In order to improve our DTypes it is unfortunately necessary to break
the ABI, which requires some changes for dtypes registered with
`PyArray_RegisterDataType`. Please see the documentation of
`PyArray_RegisterDataType` for how to adapt your code and achieve
compatibility with both 1.x and 2.x.

([gh-25792](https://github.com/numpy/numpy/pull/25792))

New Public DType API

The C implementation of the NEP 42 DType API is now public. While the
DType API has shipped in NumPy for a few versions, it was only usable in
sessions with a special environment variable set. It is now possible to
write custom DTypes outside of NumPy using the new DType API and the
normal `import_array()` mechanism for importing the numpy C API.

See `dtype-api` for more details about the API. As always with a new feature,
please report any bugs you run into implementing or using a new DType. It is
likely that downstream C code that works with dtypes will need to be updated to
work correctly with new DTypes.

([gh-25754](https://github.com/numpy/numpy/pull/25754))

New C-API import functions

We have now added `PyArray_ImportNumPyAPI` and `PyUFunc_ImportUFuncAPI`
as static inline functions to import the NumPy C-API tables. The new
functions have two advantages over `import_array` and `import_ufunc`:

-   They check whether the import was already performed and are
 light-weight if not, allowing to add them judiciously (although this
 is not preferable in most cases).
-   The old mechanisms were macros rather than functions which included
 a `return` statement.

The `PyArray_ImportNumPyAPI()` function is included in `npy_2_compat.h`
for simpler backporting.

([gh-25866](https://github.com/numpy/numpy/pull/25866))

Structured dtype information access through functions

The dtype structures fields `c_metadata`, `names`, `fields`, and
`subarray` must now be accessed through new functions following the same
names, such as `PyDataType_NAMES`. Direct access of the fields is not
valid as they do not exist for all `PyArray_Descr` instances. The
`metadata` field is kept, but the macro version should also be
preferred.

([gh-25802](https://github.com/numpy/numpy/pull/25802))

Descriptor `elsize` and `alignment` access

Unless compiling only with NumPy 2 support, the `elsize` and `aligment`
fields must now be accessed via `PyDataType_ELSIZE`,
`PyDataType_SET_ELSIZE`, and `PyDataType_ALIGNMENT`. In cases where the
descriptor is attached to an array, we advise using `PyArray_ITEMSIZE`
as it exists on all NumPy versions. Please see
`migration_c_descr` for more information.

([gh-25943](https://github.com/numpy/numpy/pull/25943))

2.0.0rc1

avoid problems for their users.**

The Python versions supported by this release are 3.9-3.12.

NumPy 2.0 Python API removals

-   `np.geterrobj`, `np.seterrobj` and the related ufunc keyword
 argument `extobj=` have been removed. The preferred replacement for
 all of these is using the context manager `with np.errstate():`.

 ([gh-23922](https://github.com/numpy/numpy/pull/23922))

-   `np.cast` has been removed. The literal replacement for
 `np.cast[dtype](arg)` is `np.asarray(arg, dtype=dtype)`.

-   `np.source` has been removed. The preferred replacement is
 `inspect.getsource`.

-   `np.lookfor` has been removed.

 ([gh-24144](https://github.com/numpy/numpy/pull/24144))

-   `numpy.who` has been removed. As an alternative for the removed
 functionality, one can use a variable explorer that is available in
 IDEs such as Spyder or Jupyter Notebook.

 ([gh-24321](https://github.com/numpy/numpy/pull/24321))

-   Multiple niche enums, expired members and functions have been
 removed from the main namespace, such as: `ERR_*`, `SHIFT_*`,
 `np.fastCopyAndTranspose`, `np.kernel_version`, `np.numarray`,
 `np.oldnumeric` and `np.set_numeric_ops`.

 ([gh-24316](https://github.com/numpy/numpy/pull/24316))

-   Replaced `from ... import *` in the `numpy/__init__.py` with
 explicit imports. As a result, these main namespace members got
 removed: `np.FLOATING_POINT_SUPPORT`, `np.FPE_*`, `np.NINF`,
 `np.PINF`, `np.NZERO`, `np.PZERO`, `np.CLIP`, `np.WRAP`, `np.WRAP`,
 `np.RAISE`, `np.BUFSIZE`, `np.UFUNC_BUFSIZE_DEFAULT`,
 `np.UFUNC_PYVALS_NAME`, `np.ALLOW_THREADS`, `np.MAXDIMS`,
 `np.MAY_SHARE_EXACT`, `np.MAY_SHARE_BOUNDS`, `add_newdoc`,
 `np.add_docstring` and `np.add_newdoc_ufunc`.

 ([gh-24357](https://github.com/numpy/numpy/pull/24357))

-   Alias `np.float_` has been removed. Use `np.float64` instead.

-   Alias `np.complex_` has been removed. Use `np.complex128` instead.

-   Alias `np.longfloat` has been removed. Use `np.longdouble` instead.

-   Alias `np.singlecomplex` has been removed. Use `np.complex64`
 instead.

-   Alias `np.cfloat` has been removed. Use `np.complex128` instead.

-   Alias `np.longcomplex` has been removed. Use `np.clongdouble`
 instead.

-   Alias `np.clongfloat` has been removed. Use `np.clongdouble`
 instead.

-   Alias `np.string_` has been removed. Use `np.bytes_` instead.

-   Alias `np.unicode_` has been removed. Use `np.str_` instead.

-   Alias `np.Inf` has been removed. Use `np.inf` instead.

-   Alias `np.Infinity` has been removed. Use `np.inf` instead.

-   Alias `np.NaN` has been removed. Use `np.nan` instead.

-   Alias `np.infty` has been removed. Use `np.inf` instead.

-   Alias `np.mat` has been removed. Use `np.asmatrix` instead.

-   `np.issubclass_` has been removed. Use the `issubclass` builtin
 instead.

-   `np.asfarray` has been removed. Use `np.asarray` with a proper dtype
 instead.

-   `np.set_string_function` has been removed. Use `np.set_printoptions`
 instead with a formatter for custom printing of NumPy objects.

-   `np.tracemalloc_domain` is now only available from `np.lib`.

-   `np.recfromcsv` and `recfromtxt` are now only available from
 `np.lib.npyio`.

-   `np.issctype`, `np.maximum_sctype`, `np.obj2sctype`,
 `np.sctype2char`, `np.sctypes`, `np.issubsctype` were all removed
 from the main namespace without replacement, as they where niche
 members.

-   Deprecated `np.deprecate` and `np.deprecate_with_doc` has been
 removed from the main namespace. Use `DeprecationWarning` instead.

-   Deprecated `np.safe_eval` has been removed from the main namespace.
 Use `ast.literal_eval` instead.

 ([gh-24376](https://github.com/numpy/numpy/pull/24376))

-   `np.find_common_type` has been removed. Use `numpy.promote_types` or
 `numpy.result_type` instead. To achieve semantics for the
 `scalar_types` argument, use `numpy.result_type` and pass `0`,
 `0.0`, or `0j` as a Python scalar instead.

-   `np.round_` has been removed. Use `np.round` instead.

-   `np.nbytes` has been removed. Use `np.dtype(<dtype>).itemsize`
 instead.

 ([gh-24477](https://github.com/numpy/numpy/pull/24477))

-   `np.compare_chararrays` has been removed from the main namespace.
 Use `np.char.compare_chararrays` instead.

-   The `charrarray` in the main namespace has been deprecated. It can
 be imported without a deprecation warning from `np.char.chararray`
 for now, but we are planning to fully deprecate and remove
 `chararray` in the future.

-   `np.format_parser` has been removed from the main namespace. Use
 `np.rec.format_parser` instead.

 ([gh-24587](https://github.com/numpy/numpy/pull/24587))

-   Support for seven data type string aliases has been removed from
 `np.dtype`: `int0`, `uint0`, `void0`, `object0`, `str0`, `bytes0`
 and `bool8`.

 ([gh-24807](https://github.com/numpy/numpy/pull/24807))

-   The experimental `numpy.array_api` submodule has been removed. Use
 the main `numpy` namespace for regular usage instead, or the
 separate `array-api-strict` package for the compliance testing use
 case for which `numpy.array_api` was mostly used.

 ([gh-25911](https://github.com/numpy/numpy/pull/25911))

`__array_prepare__` is removed

UFuncs called `__array_prepare__` before running computations for normal
ufunc calls (not generalized ufuncs, reductions, etc.). The function was
also called instead of `__array_wrap__` on the results of some linear
algebra functions.

It is now removed. If you use it, migrate to `__array_ufunc__` or rely
on `__array_wrap__` which is called with a context in all cases,
although only after the result array is filled. In those code paths,
`__array_wrap__` will now be passed a base class, rather than a subclass
array.

([gh-25105](https://github.com/numpy/numpy/pull/25105))

Deprecations

-   `np.compat` has been deprecated, as Python 2 is no longer supported.

-   `np.safe_eval` has been deprecated. `ast.literal_eval` should be
 used instead.

 ([gh-23830](https://github.com/numpy/numpy/pull/23830))

-   `np.recfromcsv`, `np.recfromtxt`, `np.disp`, `np.get_array_wrap`,
 `np.maximum_sctype`, `np.deprecate` and `np.deprecate_with_doc` have
 been deprecated.

 ([gh-24154](https://github.com/numpy/numpy/pull/24154))

-   `np.trapz` has been deprecated. Use `np.trapezoid` or a
 `scipy.integrate` function instead.

-   `np.in1d` has been deprecated. Use `np.isin` instead.

-   Alias `np.row_stack` has been deprecated. Use `np.vstack` directly.

 ([gh-24445](https://github.com/numpy/numpy/pull/24445))

-   `__array_wrap__` is now passed `arr, context, return_scalar` and
 support for implementations not accepting all three are deprecated.
 Its signature should be
 `__array_wrap__(self, arr, context=None, return_scalar=False)`

 ([gh-25408](https://github.com/numpy/numpy/pull/25408))

-   Arrays of 2-dimensional vectors for `np.cross` have been deprecated.
 Use arrays of 3-dimensional vectors instead.

 ([gh-24818](https://github.com/numpy/numpy/pull/24818))

-   `np.dtype("a")` alias for `np.dtype(np.bytes_)` was deprecated. Use
 `np.dtype("S")` alias instead.

 ([gh-24854](https://github.com/numpy/numpy/pull/24854))

-   Use of keyword arguments `x` and `y` with functions
 `assert_array_equal` and `assert_array_almost_equal` has been
 deprecated. Pass the first two arguments as positional arguments
 instead.

 ([gh-24978](https://github.com/numpy/numpy/pull/24978))

`numpy.fft` deprecations for n-D transforms with None values in arguments

Using `fftn`, `ifftn`, `rfftn`, `irfftn`, `fft2`, `ifft2`, `rfft2` or
`irfft2` with the `s` parameter set to a value that is not `None` and
the `axes` parameter set to `None` has been deprecated, in line with the
array API standard. To retain current behaviour, pass a sequence \[0,
\..., k-1\] to `axes` for an array of dimension k.

Furthermore, passing an array to `s` which contains `None` values is
deprecated as the parameter is documented to accept a sequence of
integers in both the NumPy docs and the array API specification. To use
the default behaviour of the corresponding 1-D transform, pass the value
matching the default for its `n` parameter. To use the default behaviour
for every axis, the `s` argument can be omitted.

([gh-25495](https://github.com/numpy/numpy/pull/25495))

`np.linalg.lstsq` now defaults to a new `rcond` value

`numpy.linalg.lstsq` now uses the new rcond value of the
machine precision times `max(M, N)`. Previously, the machine precision
was used but a FutureWarning was given to notify that this change will
happen eventually. That old behavior can still be achieved by passing
`rcond=-1`.

([gh-25721](https://github.com/numpy/numpy/pull/25721))

Expired deprecations

-   The `np.core.umath_tests` submodule has been removed from the public
 API. (Deprecated in NumPy 1.15)

 ([gh-23809](https://github.com/numpy/numpy/pull/23809))

-   The `PyDataMem_SetEventHook` deprecation has expired and it is
 removed. Use `tracemalloc` and the `np.lib.tracemalloc_domain`
 domain. (Deprecated in NumPy 1.23)

 ([gh-23921](https://github.com/numpy/numpy/pull/23921))

-   The deprecation of `set_numeric_ops` and the C functions
 `PyArray_SetNumericOps` and `PyArray_GetNumericOps` has been expired
 and the functions removed. (Deprecated in NumPy 1.16)

 ([gh-23998](https://github.com/numpy/numpy/pull/23998))

-   The `fasttake`, `fastclip`, and `fastputmask` `ArrFuncs` deprecation
 is now finalized.

-   The deprecated function `fastCopyAndTranspose` and its C counterpart
 are now removed.

-   The deprecation of `PyArray_ScalarFromObject` is now finalized.

 ([gh-24312](https://github.com/numpy/numpy/pull/24312))

-   `np.msort` has been removed. For a replacement, `np.sort(a, axis=0)`
 should be used instead.

 ([gh-24494](https://github.com/numpy/numpy/pull/24494))

-   `np.dtype(("f8", 1)` will now return a shape 1 subarray dtype rather
 than a non-subarray one.

 ([gh-25761](https://github.com/numpy/numpy/pull/25761))

-   Assigning to the `.data` attribute of an ndarray is disallowed and
 will raise.

-   `np.binary_repr(a, width)` will raise if width is too small.

-   Using `NPY_CHAR` in `PyArray_DescrFromType()` will raise, use
 `NPY_STRING` `NPY_UNICODE`, or `NPY_VSTRING` instead.

 ([gh-25794](https://github.com/numpy/numpy/pull/25794))

Compatibility notes

`loadtxt` and `genfromtxt` default encoding changed

`loadtxt` and `genfromtxt` now both default to `encoding=None` which may
mainly modify how `converters` work. These will now be passed `str`
rather than `bytes`. Pass the encoding explicitly to always get the new
or old behavior. For `genfromtxt` the change also means that returned
values will now be unicode strings rather than bytes.

([gh-25158](https://github.com/numpy/numpy/pull/25158))

`f2py` compatibility notes

-   `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.

 ([gh-25181](https://github.com/numpy/numpy/pull/25181))

-   The `f2py.compile()` helper has been removed because it leaked
 memory, has been marked as experimental for several years now, and
 was implemented as a thin `subprocess.run` wrapper. It was also one
 of the test bottlenecks. See
 [gh-25122](https://github.com/numpy/numpy/issues/25122) for the full
 rationale. It also used several `np.distutils` features which are
 too fragile to be ported to work with `meson`.

-   Users are urged to replace calls to `f2py.compile` with calls to
 `subprocess.run("python", "-m", "numpy.f2py",...` instead, and to
 use environment variables to interact with `meson`. [Native
 files](https://mesonbuild.com/Machine-files.html) are also an
 option.

 ([gh-25193](https://github.com/numpy/numpy/pull/25193))

Minor changes in behavior of sorting functions

Due to algorithmic changes and use of SIMD code, sorting functions with
methods that aren\'t stable may return slightly different results in

1.26.4

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.

Contributors

A total of 13 people contributed to this release. People with a \"+\" by
their names contributed a patch for the first time.

-   Charles Harris
-   Elliott Sales de Andrade
-   Lucas Colley +
-   Mark Ryan +
-   Matti Picus
-   Nathan Goldbaum
-   Ola x Nilsson +
-   Pieter Eendebak
-   Ralf Gommers
-   Sayed Adel
-   Sebastian Berg
-   Stefan van der Walt
-   Stefano Rivera

Pull requests merged

A total of 19 pull requests were merged for this release.

-   [25323](https://github.com/numpy/numpy/pull/25323): BUG: Restore missing asstr import
-   [25523](https://github.com/numpy/numpy/pull/25523): MAINT: prepare 1.26.x for further development
-   [25539](https://github.com/numpy/numpy/pull/25539): BUG: `numpy.array_api`: fix `linalg.cholesky` upper decomp\...
-   [25584](https://github.com/numpy/numpy/pull/25584): CI: Bump azure pipeline timeout to 120 minutes
-   [25585](https://github.com/numpy/numpy/pull/25585): MAINT, BLD: Fix unused inline functions warnings on clang
-   [25599](https://github.com/numpy/numpy/pull/25599): BLD: include fix for MinGW platform detection
-   [25618](https://github.com/numpy/numpy/pull/25618): TST: Fix test_numeric on riscv64
-   [25619](https://github.com/numpy/numpy/pull/25619): BLD: fix building for windows ARM64
-   [25620](https://github.com/numpy/numpy/pull/25620): MAINT: add `newaxis` to `__all__` in `numpy.array_api`
-   [256

[pyup-bot]

Closing this in favor of #1161