| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| | |
|
| |\
| |
| | |
DEP: issue deprecation warning when creating ragged array (NEP 34)
|
| | | |
|
| |/
|
|
|
| |
This argument isn't used, and is confusing. (Small test added by
seberg)
|
| |\
| |
| | |
BUG: Remove the broken clip wrapper
|
| | |
| |
| |
| |
| |
| | |
ndarray.clip is already just a wrapper for the ufunc, so there is no need to do type-specific wrapping here any more
Fixes gh-14140
|
| |/ |
|
| |\
| |
| | |
API: Make MaskedArray.mask return a view, rather than the underlying mask
|
| | |
| |
| |
| |
| |
| |
| |
| | |
This prevents consumers from reshaping the mask in place, which breaks things
As a result, `x.mask is x.mask` returns `False`, but this was already true of `x.data is x.data`.
May also be related to gh-10270
|
| |/ |
|
| | |
|
| |
|
|
|
|
|
| |
The pickle module was being imported from numpy.core.numeric. It was
defined there in order to use pickle5 when available in Python3 and
cpickle in Python2. The numpy.compat module seems a better place for
that.
|
| | |
|
| |\
| |
| | |
BUG: Make `arr.ctypes.data` hold onto a reference to the underlying array
|
| | |
| |
| |
| |
| |
| |
| |
| | |
We do this by going through ctypes.from_buffer, which keeps the object the buffer came from alive.
A bunch of hacks are required to produce a sufficiently simple buffer for ctypes to accept.
Fixes #9647
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
(#12448)
* Review F401,F841,F842 flake8 errors (unused variables, imports)
* Review comments
* More tests in test_installed_npymath_ini
* Review comments
|
| | |
| |
| |
| |
| | |
This should give an empty result, not an error. The problem
was that the empty list was interpreted as a list of strings.
|
| |/
|
|
| |
Fixes #10483
|
| |
|
|
|
| |
Cover the string and numeric types. Structured types were
already covered.
|
| |
|
|
|
|
|
| |
Fixes the regression introduced by #10211 where the masked array fill
value type was not properly converted by astype.
Closes #12070.
|
| |
|
|
|
|
|
|
| |
All imports of pickle from numpy modules are now done this way:
>>> from numpy.core.numeric import pickle
Also, some loops on protocol numbers are added over pickle tests that
were not caught from #12090
|
| |
|
|
|
|
|
|
| |
After the pytest migration, test classes no longer inherit
from unittest.TestCase and and the fail method does not
exist anymore.
In all these cases, we can use assert_raises and assert_raises_regex instead
|
| | |
|
| |
|
|
|
|
|
|
|
| |
The TestMaskedArrayMethods.test_sort_flexible test was failing for for
the subarray checks on big endian systems. That was probably due to
subarrays being sorted as bytestrings. This fixes the test by simply
checking that the sort does not error rather than checking the result.
Closes #11716.
|
| |\
| |
| | |
BUG: Prevent crashes on 0-length structured void scalars
|
| | | |
|
| |/
|
|
|
|
| |
This change was introduced as part of #10232
Fixes #11112
|
| |\
| |
| | |
BUG: Ensure __array_finalize__ cannot back-mangle shape
|
| | | |
|
| | |
| |
| | |
Fixes #11021
|
| |/ |
|
| |
|
|
|
|
| |
Further progress in isolating matrix in preparation of its
deprecation. There is one place left with an explicit reference
to matrix (in MaskedArray.count), which is to be solved later.
|
| |
|
|
|
| |
That function is nose specific and has not worked since `__init__` files
were added to the tests directories.
|
| |
|
|
|
|
|
|
|
|
|
| |
Use standard pytest markers everywhere in the numpy tests. At this point
there should be no nose dependency. However, nose is required to test
the legacy decorators if so desired.
At this point, numpy test cannot be run in the way with runtests, rather
installed numpy can be tested with `pytest --pyargs numpy` as long as
that is not run from the repo. Run it from the tools directory or some
such.
|
| |\
| |
| | |
ENH: Add "stablesort" option to inp.sort as an alias for "mergesort".
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Depending on the input, this would return:
* A single slice, if mask=nomask
* A list of slices, if mask is an array
* None, if mask is fully masked
The documented return value is a list, and all downstream callers of this
function end up having to correct for it not being one.
This affects the result of np.ma.notmasked_contiguous, which also did not
document these unusual return values.
|
| |/ |
|
| |\
| |
| | |
BUG: Calling ufuncs with a positional output argument causes the result to inherit the output's mask
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
inherit the output's mask
This brings `np.add(a, b, out)` in line with `np.add(a, b, out=out)`.
These previously differed because gh-10459 causes them to call __array_wrap__ in different ways (with and without the output argument in the context tuple, respectively).
Since the data in the `out` argument is never used by ufuncs, it seems consistent that the mask should not be either.
|
| | |
| |
| | |
[ci skip]
|
| | | |
|
| |/ |
|
| |
|
|
| |
It's possible this is the cause of gh-10270, where it seems something is wrong with the shape
|
| |\
| |
| | |
BUG: Don't reimplement isclose in np.ma
|
| | |
| |
| |
| | |
This fixes gh-10227, as well as some other problems that weren't previously known.
|
| |\ \
| |/
|/| |
MAINT,ENH: remove MaskedArray.astype, as the base type does everything.
|
| | |
| |
| |
| |
| | |
With a slight update to `__array_finalize__`, this also fixes the
mask ordering to be the same as that of the data.
|
| | |
| |
| |
| |
| | |
Uses '_fromnxfunction_seq' just like 'vstack', 'hstack', etc.
Added release note to 1.15.0.
|
| |/
|
|
| |
Fixes gh-9233
|
