| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |\
| |
| |
| |
| | |
MAINT: replace `NPY_INLINE` with `inline`
[ci skip]
|
| | |
| |
| |
| | |
Closes gh-22100
|
| |\ \
| | |
| | | |
MAINT: remove 'six' dependency from pyinstaller
|
| | | | |
|
| |\ \ \
| | | |
| | | | |
MAINT: pin ubuntu and python for emscripten
|
| | |/ / |
|
| |\ \ \
| |/ /
|/| | |
DOC: lib: Use keepdims in a couple docstrings.
|
| | | | |
|
| |/ / |
|
| |/
|
|
|
|
| |
The `shape` parameter must be specified when opened in appending mode. Docstring
and exception message wording are updated to reflect this.
Co-authored-by: Ross Barnowski <rossbar@berkeley.edu>
|
| |\
| |
| | |
MAINT: Remove the aarch64 python 3.8 wheel builds
|
| | |
| |
| |
| |
| | |
This is a quick fix for the failing tests on merge. There will be
a followup PR updating other ci infrastucture to Python 3.9
|
| |\ \
| | |
| | | |
MAINT: check user-defined dtype has an ensure_canonical implementation
|
| | | | |
|
| |\ \ \
| |/ /
|/| | |
MAINT: Rename symbols in textreading/ that may clash when statically linked
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
While it is not the standard way NumPy is built, some users build and
link NumPy statically together with other modules. In this case generic
names like `tokenize` or `to_float` can clash with similar symbols from
other modules. We can defend against this either by using a C++
namespace or by prefixing symbols likely to clash with a prefix like
npy_.
|
| |\ \ \
| | | |
| | | | |
DEP: Finalize MachAr and machar deprecations
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This removes the attributes on finfo and the "public" module. It also
deprecates `np.core.MachAr`. We should be able to get away with just
deleting it, but there seems little reason to not just deprecate it for now.
|
| |\ \ \ \
| |_|_|/
|/| | | |
MAINT: Prepare main for NumPy 1.25.0 development
|
| |/ / / |
|
| | |/
|/| |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* ENH: raise TypeError when arange() is called with string dtype
* Add release note for dtype=str change to arange()
* DOC: Minor wording/formatting touchups to release note.
* Update numpy/core/tests/test_multiarray.py
Co-authored-by: Ross Barnowski <rossbar@berkeley.edu>
* Move check to PyArray_ArangeObj
* remove old code
* BUG,MAINT: Clean out arange string error and other paths
* BUGS: Fixup and cleanup arange code a bit
* DOC: Update release note to new message
* BUG: Fix refcounting and simplify arange
* MAINT: Use SETREF to make arange dtype discovery more compact
* MAINT: Update numpy/core/src/multiarray/ctors.c
Co-authored-by: Ross Barnowski <rossbar@berkeley.edu>
Co-authored-by: Sebastian Berg <sebastianb@nvidia.com>
Co-authored-by: Charles Harris <charlesr.harris@gmail.com>
|
| |\ \
| | |
| | | |
ENH: Allow creating structured void scalars by passing dtype
|
| | | |
| | |
| | |
| | | |
Co-authored-by: Matti Picus <matti.picus@gmail.com>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Adds an optional `dtype=` kwarg to `np.void`. If given (and not None),
this kwarg effectively turns it into:
res = np.array(data, dtype=dtype)[()]
Thanks for Marten's review and Bas' help with the typing.
Reviewed-by: Marten van Kerkwijk <mhvk@astro.utoronto.ca>
Co-authored-by: Bas van Beek <43369155+BvB93@users.noreply.github.com>
|
| |\ \ \
| | | |
| | | | |
DOC: Clarify docstring of `masked_equal` and `masked_values`
|
| | | | | |
|
| | | | | |
|
| | | | | |
|
| |\ \ \ \
| |_|_|/
|/| | | |
DOC: Remove traces of interrupt handling utilities
|
| |/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We do not use these in NumPy anymore, and at this point the whole
`npy_interrupt.h` header only exists in case someone is using it
out there.
We may wish to just remove it at some point, vendoring the header is
simple enough after all (and no known downstream usage exists).
See also gh-7545, gh-12541
|
| |\ \ \
| | | |
| | | | |
DEP: Next step in scalar type alias deprecations/futurewarnings
|
| | | | | |
|
| | | | | |
|
| | | | | |
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Finalizes the scalar type alias deprecations making them an error.
However, at the same time adds a `FutureWarning` that new aliases
will be introduced in the future.
(They would eventually be preferred over the `str_`, etc. version.)
It may make sense, that this FutureWarning is already propelled soon
since it interacts with things such as changing the representation of
strings to `np.str_("")` if the preferred alias becomes `np.str`.
It also introduces a new deprecation to remove the 0 sized bit-aliases
and the bitsize `bool8` alias. (Unfortunately, these are here still allowed
as part of the `np.sctypeDict`).
|
| |\ \ \ \
| | | | |
| | | | | |
CI: Increase travis timeout to 30 minutes (2).
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This is an attempted fix for the Python 3.8 aarch64 builds that are
timing out when merged to master.
|
| |\ \ \ \ \
| |/ / / /
| | | | | |
CI: Increase travis timeout to 20 minutes.
|
| |/ / / /
| | | |
| | | |
| | | |
| | | | |
This is an attempted fix for the Python 3.8 aarch64 builds that are
timing out when merged to master.
|
| |\ \ \ \
| | | | |
| | | | | |
MAINT: Update main after 1.23.5 release.
|
| | | | | | |
|
| |\ \ \ \ \
| |/ / / /
|/| | | | |
BLD: Use cibuildwheel 2.9.0 for Python 3.8 aarch64 builds
|
| |/ / / /
| | | |
| | | |
| | | |
| | | |
| | | | |
The Python 3.8 aarch65 builds have been failing in main for some
time, but work on maintenance/1.23.x. This PR uses the version of
cibuildwheel used in 1.23.x to see if that fixes the problem.
|
| |\ \ \ \
| | | | |
| | | | | |
DOC: Add example for np.ma.power
|
| |/ / / / |
|
| |\ \ \ \
| | | | |
| | | | | |
BLD: update wheel builds on macos to macos-12 image
|
| | | | | |
| | | | |
| | | | |
| | | | | |
preserve state
|
