Numpy Versions Save

The fundamental package for scientific computing with Python.

v1.23.3

3 weeks ago

NumPy 1.23.3 Release Notes

NumPy 1.23.3 is a maintenance release that fixes bugs discovered after the 1.23.2 release. There is no major theme for this release, the main improvements are for some downstream builds and some annotation corner cases. The Python versions supported for this release are 3.8-3.11.

Note that we will move to MacOS 11 for the NumPy 1.23.4 release, the 10.15 version currently used will no longer be supported by our build infrastructure at that point.

Contributors

A total of 16 people contributed to this release. People with a "+" by their names contributed a patch for the first time.

  • Aaron Meurer
  • Bas van Beek
  • Charles Harris
  • Ganesh Kathiresan
  • Gavin Zhang +
  • Iantra Solari+
  • Jyn Spring 琴春 +
  • Matti Picus
  • Rafael Cardoso Fernandes Sousa
  • Rafael Sousa +
  • Ralf Gommers
  • Rin Cat (鈴猫) +
  • Saransh Chopra +
  • Sayed Adel
  • Sebastian Berg
  • Serge Guelton

Pull requests merged

A total of 14 pull requests were merged for this release.

  • #22136: BLD: Add Python 3.11 wheels to aarch64 build
  • #22148: MAINT: Update setup.py for Python 3.11.
  • #22155: CI: Test NumPy build against old versions of GCC(6, 7, 8)
  • #22156: MAINT: support IBM i system
  • #22195: BUG: Fix circleci build
  • #22214: BUG: Expose heapsort algorithms in a shared header
  • #22215: BUG: Support using libunwind for backtrack
  • #22216: MAINT: fix an incorrect pointer type usage in f2py
  • #22220: BUG: change overloads to play nice with pyright.
  • #22221: TST,BUG: Use fork context to fix MacOS savez test
  • #22222: TYP,BUG: Reduce argument validation in C-based __class_getitem__
  • #22223: TST: ensure np.equal.reduce raises a TypeError
  • #22224: BUG: Fix the implementation of numpy.array_api.vecdot
  • #22230: BUG: Better report integer division overflow (backport)

Checksums

MD5

a60bf0b1d440bf18d87c49409036d05a  numpy-1.23.3-cp310-cp310-macosx_10_9_x86_64.whl
59b43423a692f5351c6a43b852b210d7  numpy-1.23.3-cp310-cp310-macosx_11_0_arm64.whl
f482a4be6954b1b606320f0ffc1995dd  numpy-1.23.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a82e2ecc4060a37dae5424e624eabfe3  numpy-1.23.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
84916178e5f4d073d0008754cba7f300  numpy-1.23.3-cp310-cp310-win32.whl
605da65b9b66dfce8b62d847cb3841f7  numpy-1.23.3-cp310-cp310-win_amd64.whl
57cf29f781be955a9cd0de8d07fbce56  numpy-1.23.3-cp311-cp311-macosx_10_9_x86_64.whl
f395dcf622dff0ba44777cbae0442189  numpy-1.23.3-cp311-cp311-macosx_11_0_arm64.whl
55d6a6439913ba84ad89268e0ad59fa0  numpy-1.23.3-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
202bc3a8617f479ebe60ca0dec29964b  numpy-1.23.3-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
a42c3d058bcef47b26841bf9472a89bf  numpy-1.23.3-cp311-cp311-win32.whl
237dbd94e5529065c0c5cc4e47ceeb7e  numpy-1.23.3-cp311-cp311-win_amd64.whl
d0587d5b28d3fa7e0ec8fd3df76e4bd4  numpy-1.23.3-cp38-cp38-macosx_10_9_x86_64.whl
054234695ed3d955fb01f661db2c14fc  numpy-1.23.3-cp38-cp38-macosx_11_0_arm64.whl
4e75ac61e34f1bf23e7cbd6e2bfc7a32  numpy-1.23.3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
29ccb3a732027ee1abe23a9562c32d0c  numpy-1.23.3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
12817838edc1e1bea27df79f3a83da5d  numpy-1.23.3-cp38-cp38-win32.whl
ef430e830a9fea7d8db0218b901671f6  numpy-1.23.3-cp38-cp38-win_amd64.whl
b001f7e17df798f9b949bbe259924c77  numpy-1.23.3-cp39-cp39-macosx_10_9_x86_64.whl
bc1782f5d79187d63d14ed69a6a411e9  numpy-1.23.3-cp39-cp39-macosx_11_0_arm64.whl
f8fb0178bc34a198d5ce4e166076e1fc  numpy-1.23.3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
fb80d38c37aae1e4d416cd4de068ff0a  numpy-1.23.3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
318d0a2a27b7e361295c0382a0ff4a94  numpy-1.23.3-cp39-cp39-win32.whl
880dc73de09fccda0650e9404fa83608  numpy-1.23.3-cp39-cp39-win_amd64.whl
3b5a51f78718a1a82d2750ec159f9acf  numpy-1.23.3-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
56a0c90a303979d5bf8fc57e86e57ccb  numpy-1.23.3-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
5338d997a3178750834e742a257dfa4a  numpy-1.23.3-pp38-pypy38_pp73-win_amd64.whl
6efc60a3f6c1b74c849d53fbcc07807b  numpy-1.23.3.tar.gz

SHA256

c9f707b5bb73bf277d812ded9896f9512a43edff72712f31667d0a8c2f8e71ee  numpy-1.23.3-cp310-cp310-macosx_10_9_x86_64.whl
ffcf105ecdd9396e05a8e58e81faaaf34d3f9875f137c7372450baa5d77c9a54  numpy-1.23.3-cp310-cp310-macosx_11_0_arm64.whl
0ea3f98a0ffce3f8f57675eb9119f3f4edb81888b6874bc1953f91e0b1d4f440  numpy-1.23.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
004f0efcb2fe1c0bd6ae1fcfc69cc8b6bf2407e0f18be308612007a0762b4089  numpy-1.23.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
98dcbc02e39b1658dc4b4508442a560fe3ca5ca0d989f0df062534e5ca3a5c1a  numpy-1.23.3-cp310-cp310-win32.whl
39a664e3d26ea854211867d20ebcc8023257c1800ae89773cbba9f9e97bae036  numpy-1.23.3-cp310-cp310-win_amd64.whl
1f27b5322ac4067e67c8f9378b41c746d8feac8bdd0e0ffede5324667b8a075c  numpy-1.23.3-cp311-cp311-macosx_10_9_x86_64.whl
2ad3ec9a748a8943e6eb4358201f7e1c12ede35f510b1a2221b70af4bb64295c  numpy-1.23.3-cp311-cp311-macosx_11_0_arm64.whl
bdc9febce3e68b697d931941b263c59e0c74e8f18861f4064c1f712562903411  numpy-1.23.3-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
301c00cf5e60e08e04d842fc47df641d4a181e651c7135c50dc2762ffe293dbd  numpy-1.23.3-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
7cd1328e5bdf0dee621912f5833648e2daca72e3839ec1d6695e91089625f0b4  numpy-1.23.3-cp311-cp311-win32.whl
8355fc10fd33a5a70981a5b8a0de51d10af3688d7a9e4a34fcc8fa0d7467bb7f  numpy-1.23.3-cp311-cp311-win_amd64.whl
bc6e8da415f359b578b00bcfb1d08411c96e9a97f9e6c7adada554a0812a6cc6  numpy-1.23.3-cp38-cp38-macosx_10_9_x86_64.whl
22d43376ee0acd547f3149b9ec12eec2f0ca4a6ab2f61753c5b29bb3e795ac4d  numpy-1.23.3-cp38-cp38-macosx_11_0_arm64.whl
a64403f634e5ffdcd85e0b12c08f04b3080d3e840aef118721021f9b48fc1460  numpy-1.23.3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
efd9d3abe5774404becdb0748178b48a218f1d8c44e0375475732211ea47c67e  numpy-1.23.3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
f8c02ec3c4c4fcb718fdf89a6c6f709b14949408e8cf2a2be5bfa9c49548fd85  numpy-1.23.3-cp38-cp38-win32.whl
e868b0389c5ccfc092031a861d4e158ea164d8b7fdbb10e3b5689b4fc6498df6  numpy-1.23.3-cp38-cp38-win_amd64.whl
09f6b7bdffe57fc61d869a22f506049825d707b288039d30f26a0d0d8ea05164  numpy-1.23.3-cp39-cp39-macosx_10_9_x86_64.whl
8c79d7cf86d049d0c5089231a5bcd31edb03555bd93d81a16870aa98c6cfb79d  numpy-1.23.3-cp39-cp39-macosx_11_0_arm64.whl
e5d5420053bbb3dd64c30e58f9363d7a9c27444c3648e61460c1237f9ec3fa14  numpy-1.23.3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
d5422d6a1ea9b15577a9432e26608c73a78faf0b9039437b075cf322c92e98e7  numpy-1.23.3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
c1ba66c48b19cc9c2975c0d354f24058888cdc674bebadceb3cdc9ec403fb5d1  numpy-1.23.3-cp39-cp39-win32.whl
78a63d2df1d947bd9d1b11d35564c2f9e4b57898aae4626638056ec1a231c40c  numpy-1.23.3-cp39-cp39-win_amd64.whl
17c0e467ade9bda685d5ac7f5fa729d8d3e76b23195471adae2d6a6941bd2c18  numpy-1.23.3-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
91b8d6768a75247026e951dce3b2aac79dc7e78622fc148329135ba189813584  numpy-1.23.3-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
94c15ca4e52671a59219146ff584488907b1f9b3fc232622b47e2cf832e94fb8  numpy-1.23.3-pp38-pypy38_pp73-win_amd64.whl
51bf49c0cd1d52be0a240aa66f3458afc4b95d8993d2d04f0d91fa60c10af6cd  numpy-1.23.3.tar.gz

v1.23.2

1 month ago

NumPy 1.23.2 Release Notes

NumPy 1.23.2 is a maintenance release that fixes bugs discovered after the 1.23.1 release. Notable features are:

  • Typing changes needed for Python 3.11
  • Wheels for Python 3.11.0rc1

The Python versions supported for this release are 3.8-3.11.

Contributors

A total of 9 people contributed to this release. People with a "+" by their names contributed a patch for the first time.

  • Alexander Grund +
  • Bas van Beek
  • Charles Harris
  • Jon Cusick +
  • Matti Picus
  • Michael Osthege +
  • Pal Barta +
  • Ross Barnowski
  • Sebastian Berg

Pull requests merged

A total of 15 pull requests were merged for this release.

  • #22030: ENH: Add __array_ufunc__ typing support to the nin=1 ufuncs
  • #22031: MAINT, TYP: Fix np.angle dtype-overloads
  • #22032: MAINT: Do not let _GenericAlias wrap the underlying classes'...
  • #22033: TYP,MAINT: Allow einsum subscripts to be passed via integer...
  • #22034: MAINT,TYP: Add object-overloads for the np.generic rich comparisons
  • #22035: MAINT,TYP: Allow the squeeze and transpose method to...
  • #22036: BUG: Fix subarray to object cast ownership details
  • #22037: BUG: Use Popen to silently invoke f77 -v
  • #22038: BUG: Avoid errors on NULL during deepcopy
  • #22039: DOC: Add versionchanged for converter callable behavior.
  • #22057: MAINT: Quiet the anaconda uploads.
  • #22078: ENH: reorder includes for testing on top of system installations...
  • #22106: TST: fix test_linear_interpolation_formula_symmetric
  • #22107: BUG: Fix skip condition for test_loss_of_precision[complex256]
  • #22115: BLD: Build python3.11.0rc1 wheels.

Checksums

MD5

fe1e3480ea8c417c8f7b05f543c1448d  numpy-1.23.2-cp310-cp310-macosx_10_9_x86_64.whl
0ab14b1afd0a55a374ca69b3b39cab3c  numpy-1.23.2-cp310-cp310-macosx_11_0_arm64.whl
df059e5405bfe75c0ac77b01abbdb237  numpy-1.23.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
4ed412c4c078e96edf11ca3b11eef76b  numpy-1.23.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
0caad53d9a5e3c5e8cd29f19a9f0c014  numpy-1.23.2-cp310-cp310-win32.whl
01e508b8b4f591daff128da1cfde8e1f  numpy-1.23.2-cp310-cp310-win_amd64.whl
8ecdb7e2a87255878b748550d91cfbe0  numpy-1.23.2-cp311-cp311-macosx_10_9_x86_64.whl
e3004aae46cec9e234f78eaf473272e0  numpy-1.23.2-cp311-cp311-macosx_11_0_arm64.whl
ec23c73caf581867d5ca9255b802f144  numpy-1.23.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
9b8389f528fe113247954248f0b78ce1  numpy-1.23.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
a54b136daa2fbb483909f08eecbfa3c5  numpy-1.23.2-cp311-cp311-win32.whl
ead32e141857c5ef33b1a6cd88aefc0f  numpy-1.23.2-cp311-cp311-win_amd64.whl
df1f18e52d0a2840d101fdc9c2c6af84  numpy-1.23.2-cp38-cp38-macosx_10_9_x86_64.whl
04c986880bb24fac2f44face75eab914  numpy-1.23.2-cp38-cp38-macosx_11_0_arm64.whl
edeba58edb214390112810f7ead903a8  numpy-1.23.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
c26ea699d94d7f1009c976c66cc4def3  numpy-1.23.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
c246a78b09f8893d998d449dcab0fac3  numpy-1.23.2-cp38-cp38-win32.whl
b5c5a2f961402259e301c49b8b05de55  numpy-1.23.2-cp38-cp38-win_amd64.whl
d156dfae94d33eeff7fb9c6e5187e049  numpy-1.23.2-cp39-cp39-macosx_10_9_x86_64.whl
7f2ad7867c577eab925a31de76486765  numpy-1.23.2-cp39-cp39-macosx_11_0_arm64.whl
76262a8e5d7a4d945446467467300a10  numpy-1.23.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
8ee105f4574d61a2d494418b55f63fcb  numpy-1.23.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
2b7c79cae66023f8e716150223201981  numpy-1.23.2-cp39-cp39-win32.whl
d7af57dd070ccb165f3893412eb602e3  numpy-1.23.2-cp39-cp39-win_amd64.whl
355a231dbd87a0f2125cc23eb8f97075  numpy-1.23.2-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
4ab13c35056f67981d03f9ceec41db42  numpy-1.23.2-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
3a6f1e1256ee9be10d8cdf6be578fe52  numpy-1.23.2-pp38-pypy38_pp73-win_amd64.whl
9bf2a361509797de14ceee607387fe0f  numpy-1.23.2.tar.gz

SHA256

e603ca1fb47b913942f3e660a15e55a9ebca906857edfea476ae5f0fe9b457d5  numpy-1.23.2-cp310-cp310-macosx_10_9_x86_64.whl
633679a472934b1c20a12ed0c9a6c9eb167fbb4cb89031939bfd03dd9dbc62b8  numpy-1.23.2-cp310-cp310-macosx_11_0_arm64.whl
17e5226674f6ea79e14e3b91bfbc153fdf3ac13f5cc54ee7bc8fdbe820a32da0  numpy-1.23.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
bdc02c0235b261925102b1bd586579b7158e9d0d07ecb61148a1799214a4afd5  numpy-1.23.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
df28dda02c9328e122661f399f7655cdcbcf22ea42daa3650a26bce08a187450  numpy-1.23.2-cp310-cp310-win32.whl
8ebf7e194b89bc66b78475bd3624d92980fca4e5bb86dda08d677d786fefc414  numpy-1.23.2-cp310-cp310-win_amd64.whl
dc76bca1ca98f4b122114435f83f1fcf3c0fe48e4e6f660e07996abf2f53903c  numpy-1.23.2-cp311-cp311-macosx_10_9_x86_64.whl
ecfdd68d334a6b97472ed032b5b37a30d8217c097acfff15e8452c710e775524  numpy-1.23.2-cp311-cp311-macosx_11_0_arm64.whl
5593f67e66dea4e237f5af998d31a43e447786b2154ba1ad833676c788f37cde  numpy-1.23.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
ac987b35df8c2a2eab495ee206658117e9ce867acf3ccb376a19e83070e69418  numpy-1.23.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
d98addfd3c8728ee8b2c49126f3c44c703e2b005d4a95998e2167af176a9e722  numpy-1.23.2-cp311-cp311-win32.whl
8ecb818231afe5f0f568c81f12ce50f2b828ff2b27487520d85eb44c71313b9e  numpy-1.23.2-cp311-cp311-win_amd64.whl
909c56c4d4341ec8315291a105169d8aae732cfb4c250fbc375a1efb7a844f8f  numpy-1.23.2-cp38-cp38-macosx_10_9_x86_64.whl
8247f01c4721479e482cc2f9f7d973f3f47810cbc8c65e38fd1bbd3141cc9842  numpy-1.23.2-cp38-cp38-macosx_11_0_arm64.whl
b8b97a8a87cadcd3f94659b4ef6ec056261fa1e1c3317f4193ac231d4df70215  numpy-1.23.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
bd5b7ccae24e3d8501ee5563e82febc1771e73bd268eef82a1e8d2b4d556ae66  numpy-1.23.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
9b83d48e464f393d46e8dd8171687394d39bc5abfe2978896b77dc2604e8635d  numpy-1.23.2-cp38-cp38-win32.whl
dec198619b7dbd6db58603cd256e092bcadef22a796f778bf87f8592b468441d  numpy-1.23.2-cp38-cp38-win_amd64.whl
4f41f5bf20d9a521f8cab3a34557cd77b6f205ab2116651f12959714494268b0  numpy-1.23.2-cp39-cp39-macosx_10_9_x86_64.whl
806cc25d5c43e240db709875e947076b2826f47c2c340a5a2f36da5bb10c58d6  numpy-1.23.2-cp39-cp39-macosx_11_0_arm64.whl
8f9d84a24889ebb4c641a9b99e54adb8cab50972f0166a3abc14c3b93163f074  numpy-1.23.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
c403c81bb8ffb1c993d0165a11493fd4bf1353d258f6997b3ee288b0a48fce77  numpy-1.23.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
cf8c6aed12a935abf2e290860af8e77b26a042eb7f2582ff83dc7ed5f963340c  numpy-1.23.2-cp39-cp39-win32.whl
5e28cd64624dc2354a349152599e55308eb6ca95a13ce6a7d5679ebff2962913  numpy-1.23.2-cp39-cp39-win_amd64.whl
806970e69106556d1dd200e26647e9bee5e2b3f1814f9da104a943e8d548ca38  numpy-1.23.2-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
2bd879d3ca4b6f39b7770829f73278b7c5e248c91d538aab1e506c628353e47f  numpy-1.23.2-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
be6b350dfbc7f708d9d853663772a9310783ea58f6035eec649fb9c4371b5389  numpy-1.23.2-pp38-pypy38_pp73-win_amd64.whl
b78d00e48261fbbd04aa0d7427cf78d18401ee0abd89c7559bbf422e5b1c7d01  numpy-1.23.2.tar.gz

v1.23.1

2 months ago

NumPy 1.23.1 Release Notes

The NumPy 1.23.1 is a maintenance release that fixes bugs discovered after the 1.23.0 release. Notable fixes are:

  • Fix searchsorted for float16 NaNs
  • Fix compilation on Apple M1
  • Fix KeyError in crackfortran operator support (Slycot)

The Python version supported for this release are 3.8-3.10.

Contributors

A total of 7 people contributed to this release. People with a "+" by their names contributed a patch for the first time.

  • Charles Harris
  • Matthias Koeppe +
  • Pranab Das +
  • Rohit Goswami
  • Sebastian Berg
  • Serge Guelton
  • Srimukh Sripada +

Pull requests merged

A total of 8 pull requests were merged for this release.

  • #21866: BUG: Fix discovered MachAr (still used within valgrind)
  • #21867: BUG: Handle NaNs correctly for float16 during sorting
  • #21868: BUG: Use keepdims during normalization in np.average and...
  • #21869: DOC: mention changes to max_rows behaviour in np.loadtxt
  • #21870: BUG: Reject non integer array-likes with size 1 in delete
  • #21949: BLD: Make can_link_svml return False for 32bit builds on x86_64
  • #21951: BUG: Reorder extern "C" to only apply to function declarations...
  • #21952: BUG: Fix KeyError in crackfortran operator support

Checksums

MD5

79f0d8c114f282b834b49209d6955f98  numpy-1.23.1-cp310-cp310-macosx_10_9_x86_64.whl
42a89a88ef26b768e8933ce46b1cc2bd  numpy-1.23.1-cp310-cp310-macosx_11_0_arm64.whl
1c1d68b3483eaf99b9a3583c8ac8bf47  numpy-1.23.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
9d3e9f7f9b3dce6cf15209e4f25f346e  numpy-1.23.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
a9afb7c34b48d08fc50427ae6516b42d  numpy-1.23.1-cp310-cp310-win32.whl
a0e02823883bdfcec49309e108f65e13  numpy-1.23.1-cp310-cp310-win_amd64.whl
f40cdf4ec7bb0cf31a90a4fa294323c2  numpy-1.23.1-cp38-cp38-macosx_10_9_x86_64.whl
80115a959f0fe30d6c401b2650a61c70  numpy-1.23.1-cp38-cp38-macosx_11_0_arm64.whl
1cf199b3a93960c4f269853a56a8d8eb  numpy-1.23.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
aa6f0f192312c79cd770c2c395e9982a  numpy-1.23.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
d07bee0ea3142a96cb5e4e16aca273ca  numpy-1.23.1-cp38-cp38-win32.whl
02d0734ae8ad5e18a40c6c6de18486a0  numpy-1.23.1-cp38-cp38-win_amd64.whl
e1ca14acd7d83bc74bdf6ab0bb4bd195  numpy-1.23.1-cp39-cp39-macosx_10_9_x86_64.whl
c9152c62b2f31e742e24bfdc97b28666  numpy-1.23.1-cp39-cp39-macosx_11_0_arm64.whl
05b0b37c92f7a7e7c01afac0a5322b40  numpy-1.23.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
d9810bb71a0ef9837e87ea5c44fcab5e  numpy-1.23.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
4255577f857e838f7a94e3a614ddc5eb  numpy-1.23.1-cp39-cp39-win32.whl
787486e3cd87b98024ffe1c969c4db7a  numpy-1.23.1-cp39-cp39-win_amd64.whl
5c7b2d1471b1b9ec6ff1cb3fe1f8ac14  numpy-1.23.1-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
40d5b2ff869707b0d97325ce44631135  numpy-1.23.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
44ce1e07927cc09415df9898857792da  numpy-1.23.1-pp38-pypy38_pp73-win_amd64.whl
4f8636a9c1a77ca0fb923ba55378891f  numpy-1.23.1.tar.gz

SHA256

b15c3f1ed08df4980e02cc79ee058b788a3d0bef2fb3c9ca90bb8cbd5b8a3a04  numpy-1.23.1-cp310-cp310-macosx_10_9_x86_64.whl
9ce242162015b7e88092dccd0e854548c0926b75c7924a3495e02c6067aba1f5  numpy-1.23.1-cp310-cp310-macosx_11_0_arm64.whl
e0d7447679ae9a7124385ccf0ea990bb85bb869cef217e2ea6c844b6a6855073  numpy-1.23.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
3119daed207e9410eaf57dcf9591fdc68045f60483d94956bee0bfdcba790953  numpy-1.23.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
3ab67966c8d45d55a2bdf40701536af6443763907086c0a6d1232688e27e5447  numpy-1.23.1-cp310-cp310-win32.whl
1865fdf51446839ca3fffaab172461f2b781163f6f395f1aed256b1ddc253622  numpy-1.23.1-cp310-cp310-win_amd64.whl
aeba539285dcf0a1ba755945865ec61240ede5432df41d6e29fab305f4384db2  numpy-1.23.1-cp38-cp38-macosx_10_9_x86_64.whl
7e8229f3687cdadba2c4faef39204feb51ef7c1a9b669247d49a24f3e2e1617c  numpy-1.23.1-cp38-cp38-macosx_11_0_arm64.whl
68b69f52e6545af010b76516f5daaef6173e73353e3295c5cb9f96c35d755641  numpy-1.23.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
1408c3527a74a0209c781ac82bde2182b0f0bf54dea6e6a363fe0cc4488a7ce7  numpy-1.23.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
47f10ab202fe4d8495ff484b5561c65dd59177949ca07975663f4494f7269e3e  numpy-1.23.1-cp38-cp38-win32.whl
37e5ebebb0eb54c5b4a9b04e6f3018e16b8ef257d26c8945925ba8105008e645  numpy-1.23.1-cp38-cp38-win_amd64.whl
173f28921b15d341afadf6c3898a34f20a0569e4ad5435297ba262ee8941e77b  numpy-1.23.1-cp39-cp39-macosx_10_9_x86_64.whl
876f60de09734fbcb4e27a97c9a286b51284df1326b1ac5f1bf0ad3678236b22  numpy-1.23.1-cp39-cp39-macosx_11_0_arm64.whl
35590b9c33c0f1c9732b3231bb6a72d1e4f77872390c47d50a615686ae7ed3fd  numpy-1.23.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a35c4e64dfca659fe4d0f1421fc0f05b8ed1ca8c46fb73d9e5a7f175f85696bb  numpy-1.23.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
c2f91f88230042a130ceb1b496932aa717dcbd665350beb821534c5c7e15881c  numpy-1.23.1-cp39-cp39-win32.whl
37ece2bd095e9781a7156852e43d18044fd0d742934833335599c583618181b9  numpy-1.23.1-cp39-cp39-win_amd64.whl
8002574a6b46ac3b5739a003b5233376aeac5163e5dcd43dd7ad062f3e186129  numpy-1.23.1-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
5d732d17b8a9061540a10fda5bfeabca5785700ab5469a5e9b93aca5e2d3a5fb  numpy-1.23.1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
55df0f7483b822855af67e38fb3a526e787adf189383b4934305565d71c4b148  numpy-1.23.1-pp38-pypy38_pp73-win_amd64.whl
d748ef349bfef2e1194b59da37ed5a29c19ea8d7e6342019921ba2ba4fd8b624  numpy-1.23.1.tar.gz

v1.23.0

3 months ago

NumPy 1.23.0 Release Notes

The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. The highlights are:

  • Implementation of loadtxt in C, greatly improving its performance.
  • Exposing DLPack at the Python level for easy data exchange.
  • Changes to the promotion and comparisons of structured dtypes.
  • Improvements to f2py.

See below for the details,

New functions

  • A masked array specialization of ndenumerate is now available as numpy.ma.ndenumerate. It provides an alternative to numpy.ndenumerate and skips masked values by default.

    (gh-20020)

  • numpy.from_dlpack has been added to allow easy exchange of data using the DLPack protocol. It accepts Python objects that implement the __dlpack__ and __dlpack_device__ methods and returns a ndarray object which is generally the view of the data of the input object.

    (gh-21145)

Deprecations

  • Setting __array_finalize__ to None is deprecated. It must now be a method and may wish to call super().__array_finalize__(obj) after checking for None or if the NumPy version is sufficiently new.

    (gh-20766)

  • Using axis=32 (axis=np.MAXDIMS) in many cases had the same meaning as axis=None. This is deprecated and axis=None must be used instead.

    (gh-20920)

  • The hook function PyDataMem_SetEventHook has been deprecated and the demonstration of its use in tool/allocation_tracking has been removed. The ability to track allocations is now built-in to python via tracemalloc.

    (gh-20394)

  • numpy.distutils has been deprecated, as a result of distutils itself being deprecated. It will not be present in NumPy for Python >= 3.12, and will be removed completely 2 years after the release of Python 3.12 For more details, see distutils-status-migration{.interpreted-text role="ref"}.

    (gh-20875)

  • numpy.loadtxt will now give a DeprecationWarning when an integer dtype is requested but the value is formatted as a floating point number.

    (gh-21663)

Expired deprecations

  • The NpzFile.iteritems() and NpzFile.iterkeys() methods have been removed as part of the continued removal of Python 2 compatibility. This concludes the deprecation from 1.15.

    (gh-16830)

  • The alen and asscalar functions have been removed.

    (gh-20414)

  • The UPDATEIFCOPY array flag has been removed together with the enum NPY_ARRAY_UPDATEIFCOPY. The associated (and deprecated) PyArray_XDECREF_ERR was also removed. These were all deprecated in 1.14. They are replaced by WRITEBACKIFCOPY, that requires calling PyArray_ResoveWritebackIfCopy before the array is deallocated.

    (gh-20589)

  • Exceptions will be raised during array-like creation. When an object raised an exception during access of the special attributes __array__ or __array_interface__, this exception was usually ignored. This behaviour was deprecated in 1.21, and the exception will now be raised.

    (gh-20835)

  • Multidimensional indexing with non-tuple values is not allowed. Previously, code such as arr[ind] where ind = [[0, 1], [0, 1]] produced a FutureWarning and was interpreted as a multidimensional index (i.e., arr[tuple(ind)]). Now this example is treated like an array index over a single dimension (arr[array(ind)]). Multidimensional indexing with anything but a tuple was deprecated in NumPy 1.15.

    (gh-21029)

  • Changing to a dtype of different size in F-contiguous arrays is no longer permitted. Deprecated since Numpy 1.11.0. See below for an extended explanation of the effects of this change.

    (gh-20722)

New Features

crackfortran has support for operator and assignment overloading

crackfortran parser now understands operator and assignment definitions in a module. They are added in the body list of the module which contains a new key implementedby listing the names of the subroutines or functions implementing the operator or assignment.

(gh-15006)

f2py supports reading access type attributes from derived type statements

As a result, one does not need to use public or private statements to specify derived type access properties.

(gh-15844)

New parameter ndmin added to genfromtxt

This parameter behaves the same as ndmin from numpy.loadtxt.

(gh-20500)

np.loadtxt now supports quote character and single converter function

numpy.loadtxt now supports an additional quotechar keyword argument which is not set by default. Using quotechar='"' will read quoted fields as used by the Excel CSV dialect.

Further, it is now possible to pass a single callable rather than a dictionary for the converters argument.

(gh-20580)

Changing to dtype of a different size now requires contiguity of only the last axis

Previously, viewing an array with a dtype of a different item size required that the entire array be C-contiguous. This limitation would unnecessarily force the user to make contiguous copies of non-contiguous arrays before being able to change the dtype.

This change affects not only ndarray.view, but other construction mechanisms, including the discouraged direct assignment to ndarray.dtype.

This change expires the deprecation regarding the viewing of F-contiguous arrays, described elsewhere in the release notes.

(gh-20722)

Deterministic output files for F2PY

For F77 inputs, f2py will generate modname-f2pywrappers.f unconditionally, though these may be empty. For free-form inputs, modname-f2pywrappers.f, modname-f2pywrappers2.f90 will both be generated unconditionally, and may be empty. This allows writing generic output rules in cmake or meson and other build systems. Older behavior can be restored by passing --skip-empty-wrappers to f2py. f2py-meson{.interpreted-text role="ref"} details usage.

(gh-21187)

keepdims parameter for average

The parameter keepdims was added to the functions numpy.average and numpy.ma.average. The parameter has the same meaning as it does in reduction functions such as numpy.sum or numpy.mean.

(gh-21485)

New parameter equal_nan added to np.unique

np.unique was changed in 1.21 to treat all NaN values as equal and return a single NaN. Setting equal_nan=False will restore pre-1.21 behavior to treat NaNs as unique. Defaults to True.

(gh-21623)

Compatibility notes

1D np.linalg.norm preserves float input types, even for scalar results

Previously, this would promote to float64 when the ord argument was not one of the explicitly listed values, e.g. ord=3:

>>> f32 = np.float32([1, 2])
>>> np.linalg.norm(f32, 2).dtype
dtype('float32')
>>> np.linalg.norm(f32, 3)
dtype('float64')  # numpy 1.22
dtype('float32')  # numpy 1.23

This change affects only float32 and float16 vectors with ord other than -Inf, 0, 1, 2, and Inf.

(gh-17709)

Changes to structured (void) dtype promotion and comparisons

In general, NumPy now defines correct, but slightly limited, promotion for structured dtypes by promoting the subtypes of each field instead of raising an exception:

>>> np.result_type(np.dtype("i,i"), np.dtype("i,d"))
dtype([('f0', '<i4'), ('f1', '<f8')])

For promotion matching field names, order, and titles are enforced, however padding is ignored. Promotion involving structured dtypes now always ensures native byte-order for all fields (which may change the result of np.concatenate) and ensures that the result will be "packed", i.e. all fields are ordered contiguously and padding is removed. See structured_dtype_comparison_and_promotion{.interpreted-text role="ref"} for further details.

The repr of aligned structures will now never print the long form including offsets and itemsize unless the structure includes padding not guaranteed by align=True.

In alignment with the above changes to the promotion logic, the casting safety has been updated:

  • "equiv" enforces matching names and titles. The itemsize is allowed to differ due to padding.
  • "safe" allows mismatching field names and titles
  • The cast safety is limited by the cast safety of each included field.
  • The order of fields is used to decide cast safety of each individual field. Previously, the field names were used and only unsafe casts were possible when names mismatched.

The main important change here is that name mismatches are now considered "safe" casts.

(gh-19226)

NPY_RELAXED_STRIDES_CHECKING has been removed

NumPy cannot be compiled with NPY_RELAXED_STRIDES_CHECKING=0 anymore. Relaxed strides have been the default for many years and the option was initially introduced to allow a smoother transition.

(gh-20220)

np.loadtxt has recieved several changes

The row counting of numpy.loadtxt was fixed. loadtxt ignores fully empty lines in the file, but counted them towards max_rows. When max_rows is used and the file contains empty lines, these will now not be counted. Previously, it was possible that the result contained fewer than max_rows rows even though more data was available to be read. If the old behaviour is required, itertools.islice may be used:

import itertools
lines = itertools.islice(open("file"), 0, max_rows)
result = np.loadtxt(lines, ...)

While generally much faster and improved, numpy.loadtxt may now fail to converter certain strings to numbers that were previously successfully read. The most important cases for this are:

  • Parsing floating point values such as 1.0 into integers is now deprecated.
  • Parsing hexadecimal floats such as 0x3p3 will fail
  • An _ was previously accepted as a thousands delimiter 100_000. This will now result in an error.

If you experience these limitations, they can all be worked around by passing appropriate converters=. NumPy now supports passing a single converter to be used for all columns to make this more convenient. For example, converters=float.fromhex can read hexadecimal float numbers and converters=int will be able to read 100_000.

Further, the error messages have been generally improved. However, this means that error types may differ. In particularly, a ValueError is now always raised when parsing of a single entry fails.

(gh-20580)

Improvements

ndarray.__array_finalize__ is now callable

This means subclasses can now use super().__array_finalize__(obj) without worrying whether ndarray is their superclass or not. The actual call remains a no-op.

(gh-20766)

Add support for VSX4/Power10

With VSX4/Power10 enablement, the new instructions available in Power ISA 3.1 can be used to accelerate some NumPy operations, e.g., floor_divide, modulo, etc.

(gh-20821)

np.fromiter now accepts objects and subarrays

The numpy.fromiter function now supports object and subarray dtypes. Please see he function documentation for examples.

(gh-20993)

Math C library feature detection now uses correct signatures

Compiling is preceded by a detection phase to determine whether the underlying libc supports certain math operations. Previously this code did not respect the proper signatures. Fixing this enables compilation for the wasm-ld backend (compilation for web assembly) and reduces the number of warnings.

(gh-21154)

np.kron now maintains subclass information

np.kron maintains subclass information now such as masked arrays while computing the Kronecker product of the inputs

>>> x = ma.array([[1, 2], [3, 4]], mask=[[0, 1], [1, 0]])
>>> np.kron(x,x)
masked_array(
  data=[[1, --, --, --],
        [--, 4, --, --],
        [--, --, 4, --],
        [--, --, --, 16]],
  mask=[[False,  True,  True,  True],
        [ True, False,  True,  True],
        [ True,  True, False,  True],
        [ True,  True,  True, False]],
  fill_value=999999)

:warning: Warning, np.kron output now follows ufunc ordering (multiply) to determine the output class type

>>> class myarr(np.ndarray):
>>>    __array_priority__ = -1
>>> a = np.ones([2, 2])
>>> ma = myarray(a.shape, a.dtype, a.data)
>>> type(np.kron(a, ma)) == np.ndarray
False # Before it was True
>>> type(np.kron(a, ma)) == myarr
True

(gh-21262)

Performance improvements and changes

Faster np.loadtxt

numpy.loadtxt is now generally much faster than previously as most of it is now implemented in C.

(gh-20580)

Faster reduction operators

Reduction operations like numpy.sum, numpy.prod, numpy.add.reduce, numpy.logical_and.reduce on contiguous integer-based arrays are now much faster.

(gh-21001)

Faster np.where

numpy.where is now much faster than previously on unpredictable/random input data.

(gh-21130)

Faster operations on NumPy scalars

Many operations on NumPy scalars are now significantly faster, although rare operations (e.g. with 0-D arrays rather than scalars) may be slower in some cases. However, even with these improvements users who want the best performance for their scalars, may want to convert a known NumPy scalar into a Python one using scalar.item().

(gh-21188)

Faster np.kron

numpy.kron is about 80% faster as the product is now computed using broadcasting.

(gh-21354)

Checksums

MD5

21839aaeab3088e685d7c8d0e1856a23  numpy-1.23.0-cp310-cp310-macosx_10_9_x86_64.whl
e657684ea521c50de0197aabfb44e78d  numpy-1.23.0-cp310-cp310-macosx_11_0_arm64.whl
219017660861fdec59b852630e3fef2a  numpy-1.23.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
03c3df83b8327910482a7d24ebe9213b  numpy-1.23.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
b8f06ce4054acc147845a9643bd36082  numpy-1.23.0-cp310-cp310-win32.whl
877322db5a62634eef4e351db99a070d  numpy-1.23.0-cp310-cp310-win_amd64.whl
7bb54f95e74306eff733466b6343695f  numpy-1.23.0-cp38-cp38-macosx_10_9_x86_64.whl
5514a0030e5cf065e916950737d6d129  numpy-1.23.0-cp38-cp38-macosx_11_0_arm64.whl
22d43465791814fe50e03ded430bd80c  numpy-1.23.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
771a1f7e488327645bac5b54dd2f6286  numpy-1.23.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
449bfa2d55aff3e722d2fc85a7549620  numpy-1.23.0-cp38-cp38-win32.whl
60c7d27cf92dadb6d206df6e65b1032f  numpy-1.23.0-cp38-cp38-win_amd64.whl
dc2a5c5d2223f7b45a45f7f760d0f2db  numpy-1.23.0-cp39-cp39-macosx_10_9_x86_64.whl
ba5729353c3521ed7ee72c796e77a546  numpy-1.23.0-cp39-cp39-macosx_11_0_arm64.whl
06d5cd49de096482944dead2eb92d783  numpy-1.23.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
6ff50a994f6006349b5f1415e4da6f45  numpy-1.23.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
49185f219512403ef23d43d6f2adbefd  numpy-1.23.0-cp39-cp39-win32.whl
ff126a84dcf91700f9ca13ff606d109f  numpy-1.23.0-cp39-cp39-win_amd64.whl
e1462428487dc599cdffb723dec642c4  numpy-1.23.0-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
fef1d20265135737fbc0f91ca4441990  numpy-1.23.0-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
4f8142288202a32c682d01921d6c2c78  numpy-1.23.0-pp38-pypy38_pp73-win_amd64.whl
513e4241d06b8fae5732cd049cdf3b57  numpy-1.23.0.tar.gz

SHA256

58bfd40eb478f54ff7a5710dd61c8097e169bc36cc68333d00a9bcd8def53b38  numpy-1.23.0-cp310-cp310-macosx_10_9_x86_64.whl
196cd074c3f97c4121601790955f915187736f9cf458d3ee1f1b46aff2b1ade0  numpy-1.23.0-cp310-cp310-macosx_11_0_arm64.whl
f1d88ef79e0a7fa631bb2c3dda1ea46b32b1fe614e10fedd611d3d5398447f2f  numpy-1.23.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
d54b3b828d618a19779a84c3ad952e96e2c2311b16384e973e671aa5be1f6187  numpy-1.23.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
2b2da66582f3a69c8ce25ed7921dcd8010d05e59ac8d89d126a299be60421171  numpy-1.23.0-cp310-cp310-win32.whl
97a76604d9b0e79f59baeca16593c711fddb44936e40310f78bfef79ee9a835f  numpy-1.23.0-cp310-cp310-win_amd64.whl
d8cc87bed09de55477dba9da370c1679bd534df9baa171dd01accbb09687dac3  numpy-1.23.0-cp38-cp38-macosx_10_9_x86_64.whl
f0f18804df7370571fb65db9b98bf1378172bd4e962482b857e612d1fec0f53e  numpy-1.23.0-cp38-cp38-macosx_11_0_arm64.whl
ac86f407873b952679f5f9e6c0612687e51547af0e14ddea1eedfcb22466babd  numpy-1.23.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
ae8adff4172692ce56233db04b7ce5792186f179c415c37d539c25de7298d25d  numpy-1.23.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
fe8b9683eb26d2c4d5db32cd29b38fdcf8381324ab48313b5b69088e0e355379  numpy-1.23.0-cp38-cp38-win32.whl
5043bcd71fcc458dfb8a0fc5509bbc979da0131b9d08e3d5f50fb0bbb36f169a  numpy-1.23.0-cp38-cp38-win_amd64.whl
1c29b44905af288b3919803aceb6ec7fec77406d8b08aaa2e8b9e63d0fe2f160  numpy-1.23.0-cp39-cp39-macosx_10_9_x86_64.whl
98e8e0d8d69ff4d3fa63e6c61e8cfe2d03c29b16b58dbef1f9baa175bbed7860  numpy-1.23.0-cp39-cp39-macosx_11_0_arm64.whl
79a506cacf2be3a74ead5467aee97b81fca00c9c4c8b3ba16dbab488cd99ba10  numpy-1.23.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
092f5e6025813e64ad6d1b52b519165d08c730d099c114a9247c9bb635a2a450  numpy-1.23.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
d6ca8dabe696c2785d0c8c9b0d8a9b6e5fdbe4f922bde70d57fa1a2848134f95  numpy-1.23.0-cp39-cp39-win32.whl
fc431493df245f3c627c0c05c2bd134535e7929dbe2e602b80e42bf52ff760bc  numpy-1.23.0-cp39-cp39-win_amd64.whl
f9c3fc2adf67762c9fe1849c859942d23f8d3e0bee7b5ed3d4a9c3eeb50a2f07  numpy-1.23.0-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
d0d2094e8f4d760500394d77b383a1b06d3663e8892cdf5df3c592f55f3bff66  numpy-1.23.0-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
94b170b4fa0168cd6be4becf37cb5b127bd12a795123984385b8cd4aca9857e5  numpy-1.23.0-pp38-pypy38_pp73-win_amd64.whl
bd3fa4fe2e38533d5336e1272fc4e765cabbbde144309ccee8675509d5cd7b05  numpy-1.23.0.tar.gz

v1.23.0rc3

3 months ago

NumPy 1.23.0 Release Notes

The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. The highlights are:

  • Implementation of loadtxt in C, greatly improving its performance.
  • Exposing DLPack at the Python level for easy data exchange.
  • Changes to the promotion and comparisons of structured dtypes.
  • Improvements to f2py.

See below for the details,

New functions

  • A masked array specialization of ndenumerate is now available as numpy.ma.ndenumerate. It provides an alternative to numpy.ndenumerate and skips masked values by default.

    (gh-20020)

  • numpy.from_dlpack has been added to allow easy exchange of data using the DLPack protocol. It accepts Python objects that implement the __dlpack__ and __dlpack_device__ methods and returns a ndarray object which is generally the view of the data of the input object.

    (gh-21145)

Deprecations

  • Setting __array_finalize__ to None is deprecated. It must now be a method and may wish to call super().__array_finalize__(obj) after checking for None or if the NumPy version is sufficiently new.

    (gh-20766)

  • Using axis=32 (axis=np.MAXDIMS) in many cases had the same meaning as axis=None. This is deprecated and axis=None must be used instead.

    (gh-20920)

  • The hook function PyDataMem_SetEventHook has been deprecated and the demonstration of its use in tool/allocation_tracking has been removed. The ability to track allocations is now built-in to python via tracemalloc.

    (gh-20394)

  • numpy.distutils has been deprecated, as a result of distutils itself being deprecated. It will not be present in NumPy for Python >= 3.12, and will be removed completely 2 years after the release of Python 3.12 For more details, see distutils-status-migration{.interpreted-text role="ref"}.

    (gh-20875)

  • numpy.loadtxt will now give a DeprecationWarning when an integer dtype is requested but the value is formatted as a floating point number.

    (gh-21663)

Expired deprecations

  • The NpzFile.iteritems() and NpzFile.iterkeys() methods have been removed as part of the continued removal of Python 2 compatibility. This concludes the deprecation from 1.15.

    (gh-16830)

  • The alen and asscalar functions have been removed.

    (gh-20414)

  • The UPDATEIFCOPY array flag has been removed together with the enum NPY_ARRAY_UPDATEIFCOPY. The associated (and deprecated) PyArray_XDECREF_ERR was also removed. These were all deprecated in 1.14. They are replaced by WRITEBACKIFCOPY, that requires calling PyArray_ResoveWritebackIfCopy before the array is deallocated.

    (gh-20589)

  • Exceptions will be raised during array-like creation. When an object raised an exception during access of the special attributes __array__ or __array_interface__, this exception was usually ignored. This behaviour was deprecated in 1.21, and the exception will now be raised.

    (gh-20835)

  • Multidimensional indexing with non-tuple values is not allowed. Previously, code such as arr[ind] where ind = [[0, 1], [0, 1]] produced a FutureWarning and was interpreted as a multidimensional index (i.e., arr[tuple(ind)]). Now this example is treated like an array index over a single dimension (arr[array(ind)]). Multidimensional indexing with anything but a tuple was deprecated in NumPy 1.15.

    (gh-21029)

  • Changing to a dtype of different size in F-contiguous arrays is no longer permitted. Deprecated since Numpy 1.11.0. See below for an extended explanation of the effects of this change.

    (gh-20722)

New Features

crackfortran has support for operator and assignment overloading

crackfortran parser now understands operator and assignment definitions in a module. They are added in the body list of the module which contains a new key implementedby listing the names of the subroutines or functions implementing the operator or assignment.

(gh-15006)

f2py supports reading access type attributes from derived type statements

As a result, one does not need to use public or private statements to specify derived type access properties.

(gh-15844)

New parameter ndmin added to genfromtxt

This parameter behaves the same as ndmin from numpy.loadtxt.

(gh-20500)

np.loadtxt now supports quote character and single converter function

numpy.loadtxt now supports an additional quotechar keyword argument which is not set by default. Using quotechar='"' will read quoted fields as used by the Excel CSV dialect.

Further, it is now possible to pass a single callable rather than a dictionary for the converters argument.

(gh-20580)

Changing to dtype of a different size now requires contiguity of only the last axis

Previously, viewing an array with a dtype of a different item size required that the entire array be C-contiguous. This limitation would unnecessarily force the user to make contiguous copies of non-contiguous arrays before being able to change the dtype.

This change affects not only ndarray.view, but other construction mechanisms, including the discouraged direct assignment to ndarray.dtype.

This change expires the deprecation regarding the viewing of F-contiguous arrays, described elsewhere in the release notes.

(gh-20722)

Deterministic output files for F2PY

For F77 inputs, f2py will generate modname-f2pywrappers.f unconditionally, though these may be empty. For free-form inputs, modname-f2pywrappers.f, modname-f2pywrappers2.f90 will both be generated unconditionally, and may be empty. This allows writing generic output rules in cmake or meson and other build systems. Older behavior can be restored by passing --skip-empty-wrappers to f2py. f2py-meson{.interpreted-text role="ref"} details usage.

(gh-21187)

keepdims parameter for average

The parameter keepdims was added to the functions numpy.average and numpy.ma.average. The parameter has the same meaning as it does in reduction functions such as numpy.sum or numpy.mean.

(gh-21485)

New parameter equal_nan added to np.unique

np.unique was changed in 1.21 to treat all NaN values as equal and return a single NaN. Setting equal_nan=False will restore pre-1.21 behavior to treat NaNs as unique. Defaults to True.

(gh-21623)

Compatibility notes

1D np.linalg.norm preserves float input types, even for scalar results

Previously, this would promote to float64 when the ord argument was not one of the explicitly listed values, e.g. ord=3:

>>> f32 = np.float32([1, 2])
>>> np.linalg.norm(f32, 2).dtype
dtype('float32')
>>> np.linalg.norm(f32, 3)
dtype('float64')  # numpy 1.22
dtype('float32')  # numpy 1.23

This change affects only float32 and float16 vectors with ord other than -Inf, 0, 1, 2, and Inf.

(gh-17709)

Changes to structured (void) dtype promotion and comparisons

In general, NumPy now defines correct, but slightly limited, promotion for structured dtypes by promoting the subtypes of each field instead of raising an exception:

>>> np.result_type(np.dtype("i,i"), np.dtype("i,d"))
dtype([('f0', '<i4'), ('f1', '<f8')])

For promotion matching field names, order, and titles are enforced, however padding is ignored. Promotion involving structured dtypes now always ensures native byte-order for all fields (which may change the result of np.concatenate) and ensures that the result will be "packed", i.e. all fields are ordered contiguously and padding is removed. See structured_dtype_comparison_and_promotion{.interpreted-text role="ref"} for further details.

The repr of aligned structures will now never print the long form including offsets and itemsize unless the structure includes padding not guaranteed by align=True.

In alignment with the above changes to the promotion logic, the casting safety has been updated:

  • "equiv" enforces matching names and titles. The itemsize is allowed to differ due to padding.
  • "safe" allows mismatching field names and titles
  • The cast safety is limited by the cast safety of each included field.
  • The order of fields is used to decide cast safety of each individual field. Previously, the field names were used and only unsafe casts were possible when names mismatched.

The main important change here is that name mismatches are now considered "safe" casts.

(gh-19226)

NPY_RELAXED_STRIDES_CHECKING has been removed

NumPy cannot be compiled with NPY_RELAXED_STRIDES_CHECKING=0 anymore. Relaxed strides have been the default for many years and the option was initially introduced to allow a smoother transition.

(gh-20220)

np.loadtxt has recieved several changes

The row counting of numpy.loadtxt was fixed. loadtxt ignores fully empty lines in the file, but counted them towards max_rows. When max_rows is used and the file contains empty lines, these will now not be counted. Previously, it was possible that the result contained fewer than max_rows rows even though more data was available to be read. If the old behaviour is required, itertools.islice may be used:

import itertools
lines = itertools.islice(open("file"), 0, max_rows)
result = np.loadtxt(lines, ...)

While generally much faster and improved, numpy.loadtxt may now fail to converter certain strings to numbers that were previously successfully read. The most important cases for this are:

  • Parsing floating point values such as 1.0 into integers is now deprecated.
  • Parsing hexadecimal floats such as 0x3p3 will fail
  • An _ was previously accepted as a thousands delimiter 100_000. This will now result in an error.

If you experience these limitations, they can all be worked around by passing appropriate converters=. NumPy now supports passing a single converter to be used for all columns to make this more convenient. For example, converters=float.fromhex can read hexadecimal float numbers and converters=int will be able to read 100_000.

Further, the error messages have been generally improved. However, this means that error types may differ. In particularly, a ValueError is now always raised when parsing of a single entry fails.

(gh-20580)

Improvements

ndarray.__array_finalize__ is now callable

This means subclasses can now use super().__array_finalize__(obj) without worrying whether ndarray is their superclass or not. The actual call remains a no-op.

(gh-20766)

Add support for VSX4/Power10

With VSX4/Power10 enablement, the new instructions available in Power ISA 3.1 can be used to accelerate some NumPy operations, e.g., floor_divide, modulo, etc.

(gh-20821)

np.fromiter now accepts objects and subarrays

The numpy.fromiter function now supports object and subarray dtypes. Please see he function documentation for examples.

(gh-20993)

Math C library feature detection now uses correct signatures

Compiling is preceded by a detection phase to determine whether the underlying libc supports certain math operations. Previously this code did not respect the proper signatures. Fixing this enables compilation for the wasm-ld backend (compilation for web assembly) and reduces the number of warnings.

(gh-21154)

np.kron now maintains subclass information

np.kron maintains subclass information now such as masked arrays while computing the Kronecker product of the inputs

>>> x = ma.array([[1, 2], [3, 4]], mask=[[0, 1], [1, 0]])
>>> np.kron(x,x)
masked_array(
  data=[[1, --, --, --],
        [--, 4, --, --],
        [--, --, 4, --],
        [--, --, --, 16]],
  mask=[[False,  True,  True,  True],
        [ True, False,  True,  True],
        [ True,  True, False,  True],
        [ True,  True,  True, False]],
  fill_value=999999)

:warning: Warning, np.kron output now follows ufunc ordering (multiply) to determine the output class type

>>> class myarr(np.ndarray):
>>>    __array_priority__ = -1
>>> a = np.ones([2, 2])
>>> ma = myarray(a.shape, a.dtype, a.data)
>>> type(np.kron(a, ma)) == np.ndarray
False # Before it was True
>>> type(np.kron(a, ma)) == myarr
True

(gh-21262)

String comparisons now supported in ufuncs

The comparison ufuncs [np.equal]{.title-ref}, [np.greater]{.title-ref}, etc. now support unicode and byte string inputs (dtypes S and U). Due to this change a FutureWarning is now given when comparing unicode to byte strings. Such comparisons always returned False and continue to do so at this time.

(gh-21716)

Performance improvements and changes

Faster np.loadtxt

numpy.loadtxt is now generally much faster than previously as most of it is now implemented in C.

(gh-20580)

Faster reduction operators

Reduction operations like numpy.sum, numpy.prod, numpy.add.reduce, numpy.logical_and.reduce on contiguous integer-based arrays are now much faster.

(gh-21001)

Faster np.where

numpy.where is now much faster than previously on unpredictable/random input data.

(gh-21130)

Faster operations on NumPy scalars

Many operations on NumPy scalars are now significantly faster, although rare operations (e.g. with 0-D arrays rather than scalars) may be slower in some cases. However, even with these improvements users who want the best performance for their scalars, may want to convert a known NumPy scalar into a Python one using scalar.item().

(gh-21188)

Faster np.kron

numpy.kron is about 80% faster as the product is now computed using broadcasting.

(gh-21354)

Checksums

MD5

527ffa92c1e964bfcfd09497a319090e  numpy-1.23.0rc3-cp310-cp310-macosx_10_9_x86_64.whl
14538d5c5a9f07c7c54b8f27ffcfcdf8  numpy-1.23.0rc3-cp310-cp310-macosx_11_0_arm64.whl
2914affc99c1d00a1e31689dfe8200c1  numpy-1.23.0rc3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a72038b74d0fe5d43967fa3c3f44a71a  numpy-1.23.0rc3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
b1c01a974cf10fdefe7c346aa09bc20a  numpy-1.23.0rc3-cp310-cp310-win32.whl
b1ae6d311ad01f15a7cf28dbb906d2a4  numpy-1.23.0rc3-cp310-cp310-win_amd64.whl
ea734f44ae2b10c9cf1530eece6c58d6  numpy-1.23.0rc3-cp38-cp38-macosx_10_9_x86_64.whl
8bc2051b1172627cbb7ca503a75fcdd8  numpy-1.23.0rc3-cp38-cp38-macosx_11_0_arm64.whl
495efaf5566fd119a0f0b4c9339449ed  numpy-1.23.0rc3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
1b93121d0c339b35dc83b92fa67a3545  numpy-1.23.0rc3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
50433ef3beaab89d5485d0fe7cbbaaaa  numpy-1.23.0rc3-cp38-cp38-win32.whl
9014a7e33d01fd4d36298836b3199eac  numpy-1.23.0rc3-cp38-cp38-win_amd64.whl
13661a5d83fd2bb423514681d0d00de4  numpy-1.23.0rc3-cp39-cp39-macosx_10_9_x86_64.whl
11386871be967bcb5331d8289b1186a6  numpy-1.23.0rc3-cp39-cp39-macosx_11_0_arm64.whl
43e6fa27b8ab7cb367103d1b890ca7c5  numpy-1.23.0rc3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
b9aa6e6cb2ad37993ea4c70e7edeec4c  numpy-1.23.0rc3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
030669015597201aca83e8cb2d2f82ed  numpy-1.23.0rc3-cp39-cp39-win32.whl
ffab9ff64e6a08784dfb0eab430d4158  numpy-1.23.0rc3-cp39-cp39-win_amd64.whl
758408262bfc96942193e22cbfdab6dc  numpy-1.23.0rc3-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
27fb60bd1cd54e785aa63b1a29b32fb1  numpy-1.23.0rc3-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
b8819d1885f39a67917faf131e716a16  numpy-1.23.0rc3-pp38-pypy38_pp73-win_amd64.whl
b6e3a4784ce0ab3d1a9d850fde5049a3  numpy-1.23.0rc3.tar.gz

SHA256

5716acbd372407b046d168228217533d48f46b1fb76a788a01ce7b446d85e054  numpy-1.23.0rc3-cp310-cp310-macosx_10_9_x86_64.whl
65ac805918a4793aa50a0f366f0a31aa67e85a3e9a28fbdc49bec076ea9d59dd  numpy-1.23.0rc3-cp310-cp310-macosx_11_0_arm64.whl
3947ff9ea7f50d44a0c278fb9c9b43109e6b2e02273b4b3341f3eea9fe31cc76  numpy-1.23.0rc3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
8635887a615e8d3cf0ce35c401b30ccf7818eb58c9ac282561b2363d85729150  numpy-1.23.0rc3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
68d6f3794c78b37faf015c9d60207aaf156edfaf96b1d8c01a2694953b979b8f  numpy-1.23.0rc3-cp310-cp310-win32.whl
c3237a338d59e271145e9773020b49dfb77314b749bde67783c45d946346a13f  numpy-1.23.0rc3-cp310-cp310-win_amd64.whl
db5a2b42d09cc0661fec75b4e3daabff91a4858e67ee106fb793ebcd02ebd7b6  numpy-1.23.0rc3-cp38-cp38-macosx_10_9_x86_64.whl
fd999d0b503dce88fd97153986c3970154c965cfd1c9835f610e89f0dda3bda3  numpy-1.23.0rc3-cp38-cp38-macosx_11_0_arm64.whl
bc58787ee33ff7828be36b656b08ddaefb7b24339aacd2e21fb28fc49b44008c  numpy-1.23.0rc3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
fae0b80e6b85e3a72a8f7dab8af91199da31b6d0d86fab36529945f23f806d35  numpy-1.23.0rc3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
ca12e7773b4f76addf2f91dad61307426f9c1fdf8da5c00b49080af56a9eeff8  numpy-1.23.0rc3-cp38-cp38-win32.whl
f403bea354d3c892ab1da29ab9d14d51149491a32731b32a94608e3dcd514533  numpy-1.23.0rc3-cp38-cp38-win_amd64.whl
849e7aa12c1df7e9f26d1936029710bb6050bc31c9cc9772d83c30e3fc7cf8e9  numpy-1.23.0rc3-cp39-cp39-macosx_10_9_x86_64.whl
e9ad17a823c106dbed8675f4d831381fe99e9df9945485792d8762c186259dcd  numpy-1.23.0rc3-cp39-cp39-macosx_11_0_arm64.whl
03d5111e85d6371ffc3678824542a8df69ea206f16f83de76ac10231d40c006b  numpy-1.23.0rc3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
b757b22fab2a3ff1f41d0f2cb69bed337eda211db7a2c981b3f93180e4aa68ea  numpy-1.23.0rc3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
87fd874392c7f8ee23829b4ac6737b19a20aeff2418169afc3e6da00c0e23bc1  numpy-1.23.0rc3-cp39-cp39-win32.whl
f159357529470285dff46e69d97e3c9694b0bd75229e4592edb02f8b812c7a4c  numpy-1.23.0rc3-cp39-cp39-win_amd64.whl
a0be42a51fbb6087049dfa9b036568e9674a25fb4615f2bf60f189f435691aa4  numpy-1.23.0rc3-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
da170cde43519e5490a1887898719e881575d0f0924ef35fe66d990246db6f0e  numpy-1.23.0rc3-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
33aa550e001817c8075c56a3b350fb3687cba66ee55dfbb09085d0d0fd63f6e9  numpy-1.23.0rc3-pp38-pypy38_pp73-win_amd64.whl
6a8fc5573cc8cb8108cec555bca5745d2798c54eef107d478b4320c1f6542102  numpy-1.23.0rc3.tar.gz

v1.23.0rc2

4 months ago

NumPy 1.23.0 Release Notes

The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. The highlights are:

  • Implementation of loadtxt in C, greatly improving its performance.
  • Exposing DLPack at the Python level for easy data exchange.
  • Changes to the promotion and comparisons of structured dtypes.
  • Improvements to f2py.

See below for the details,

New functions

  • A masked array specialization of ndenumerate is now available as numpy.ma.ndenumerate. It provides an alternative to numpy.ndenumerate and skips masked values by default.

    (gh-20020)

  • numpy.from_dlpack has been added to allow easy exchange of data using the DLPack protocol. It accepts Python objects that implement the __dlpack__ and __dlpack_device__ methods and returns a ndarray object which is generally the view of the data of the input object.

    (gh-21145)

Deprecations

  • Setting __array_finalize__ to None is deprecated. It must now be a method and may wish to call super().__array_finalize__(obj) after checking for None or if the NumPy version is sufficiently new.

    (gh-20766)

  • Using axis=32 (axis=np.MAXDIMS) in many cases had the same meaning as axis=None. This is deprecated and axis=None must be used instead.

    (gh-20920)

  • The hook function PyDataMem_SetEventHook has been deprecated and the demonstration of its use in tool/allocation_tracking has been removed. The ability to track allocations is now built-in to python via tracemalloc.

    (gh-20394)

  • numpy.distutils has been deprecated, as a result of distutils itself being deprecated. It will not be present in NumPy for Python >= 3.12, and will be removed completely 2 years after the release of Python 3.12 For more details, see distutils-status-migration{.interpreted-text role="ref"}.

    (gh-20875)

Expired deprecations

  • The NpzFile.iteritems() and NpzFile.iterkeys() methods have been removed as part of the continued removal of Python 2 compatibility. This concludes the deprecation from 1.15.

    (gh-16830)

  • The alen and asscalar functions have been removed.

    (gh-20414)

  • The UPDATEIFCOPY array flag has been removed together with the enum NPY_ARRAY_UPDATEIFCOPY. The associated (and deprecated) PyArray_XDECREF_ERR was also removed. These were all deprecated in 1.14. They are replaced by WRITEBACKIFCOPY, that requires calling PyArray_ResoveWritebackIfCopy before the array is deallocated.

    (gh-20589)

  • Exceptions will be raised during array-like creation. When an object raised an exception during access of the special attributes __array__ or __array_interface__, this exception was usually ignored. This behaviour was deprecated in 1.21, and the exception will now be raised.

    (gh-20835)

  • Multidimensional indexing with non-tuple values is not allowed. Previously, code such as arr[ind] where ind = [[0, 1], [0, 1]] produced a FutureWarning and was interpreted as a multidimensional index (i.e., arr[tuple(ind)]). Now this example is treated like an array index over a single dimension (arr[array(ind)]). Multidimensional indexing with anything but a tuple was deprecated in NumPy 1.15.

    (gh-21029)

  • Changing to a dtype of different size in F-contiguous arrays is no longer permitted. Deprecated since Numpy 1.11.0. See below for an extended explanation of the effects of this change.

    (gh-20722)

New Features

crackfortran has support for operator and assignment overloading

crackfortran parser now understands operator and assignment definitions in a module. They are added in the body list of the module which contains a new key implementedby listing the names of the subroutines or functions implementing the operator or assignment.

(gh-15006)

f2py supports reading access type attributes from derived type statements

As a result, one does not need to use public or private statements to specify derived type access properties.

(gh-15844)

New parameter ndmin added to genfromtxt

This parameter behaves the same as ndmin from numpy.loadtxt.

(gh-20500)

np.loadtxt now supports quote character and single converter function

numpy.loadtxt now supports an additional quotechar keyword argument which is not set by default. Using quotechar='"' will read quoted fields as used by the Excel CSV dialect.

Further, it is now possible to pass a single callable rather than a dictionary for the converters argument.

(gh-20580)

Changing to dtype of a different size now requires contiguity of only the last axis

Previously, viewing an array with a dtype of a different item size required that the entire array be C-contiguous. This limitation would unnecessarily force the user to make contiguous copies of non-contiguous arrays before being able to change the dtype.

This change affects not only ndarray.view, but other construction mechanisms, including the discouraged direct assignment to ndarray.dtype.

This change expires the deprecation regarding the viewing of F-contiguous arrays, described elsewhere in the release notes.

(gh-20722)

Deterministic output files for F2PY

For F77 inputs, f2py will generate modname-f2pywrappers.f unconditionally, though these may be empty. For free-form inputs, modname-f2pywrappers.f, modname-f2pywrappers2.f90 will both be generated unconditionally, and may be empty. This allows writing generic output rules in cmake or meson and other build systems. Older behavior can be restored by passing --skip-empty-wrappers to f2py. f2py-meson{.interpreted-text role="ref"} details usage.

(gh-21187)

keepdims parameter for average

The parameter keepdims was added to the functions numpy.average and numpy.ma.average. The parameter has the same meaning as it does in reduction functions such as numpy.sum or numpy.mean.

(gh-21485)

Compatibility notes

1D np.linalg.norm preserves float input types, even for scalar results

Previously, this would promote to float64 when the ord argument was not one of the explicitly listed values, e.g. ord=3:

>>> f32 = np.float32([1, 2])
>>> np.linalg.norm(f32, 2).dtype
dtype('float32')
>>> np.linalg.norm(f32, 3)
dtype('float64')  # numpy 1.22
dtype('float32')  # numpy 1.23

This change affects only float32 and float16 vectors with ord other than -Inf, 0, 1, 2, and Inf.

(gh-17709)

Changes to structured (void) dtype promotion and comparisons

In general, NumPy now defines correct, but slightly limited, promotion for structured dtypes by promoting the subtypes of each field instead of raising an exception:

>>> np.result_type(np.dtype("i,i"), np.dtype("i,d"))
dtype([('f0', '<i4'), ('f1', '<f8')])

For promotion matching field names, order, and titles are enforced, however padding is ignored. Promotion involving structured dtypes now always ensures native byte-order for all fields (which may change the result of np.concatenate) and ensures that the result will be "packed", i.e. all fields are ordered contiguously and padding is removed. See structured_dtype_comparison_and_promotion{.interpreted-text role="ref"} for further details.

The repr of aligned structures will now never print the long form including offsets and itemsize unless the structure includes padding not guaranteed by align=True.

In alignment with the above changes to the promotion logic, the casting safety has been updated:

  • "equiv" enforces matching names and titles. The itemsize is allowed to differ due to padding.
  • "safe" allows mismatching field names and titles
  • The cast safety is limited by the cast safety of each included field.
  • The order of fields is used to decide cast safety of each individual field. Previously, the field names were used and only unsafe casts were possible when names mismatched.

The main important change here is that name mismatches are now considered "safe" casts.

(gh-19226)

NPY_RELAXED_STRIDES_CHECKING has been removed

NumPy cannot be compiled with NPY_RELAXED_STRIDES_CHECKING=0 anymore. Relaxed strides have been the default for many years and the option was initially introduced to allow a smoother transition.

(gh-20220)

np.loadtxt has recieved several changes

The row counting of numpy.loadtxt was fixed. loadtxt ignores fully empty lines in the file, but counted them towards max_rows. When max_rows is used and the file contains empty lines, these will now not be counted. Previously, it was possible that the result contained fewer than max_rows rows even though more data was available to be read. If the old behaviour is required, itertools.islice may be used:

import itertools
lines = itertools.islice(open("file"), 0, max_rows)
result = np.loadtxt(lines, ...)

While generally much faster and improved, numpy.loadtxt may now fail to converter certain strings to numbers that were previously successfully read. The most important cases for this are:

  • Parsing floating point values such as 1.0 into integers will now fail
  • Parsing hexadecimal floats such as 0x3p3 will fail
  • An _ was previously accepted as a thousands delimiter 100_000. This will now result in an error.

If you experience these limitations, they can all be worked around by passing appropriate converters=. NumPy now supports passing a single converter to be used for all columns to make this more convenient. For example, converters=float.fromhex can read hexadecimal float numbers and converters=int will be able to read 100_000.

Further, the error messages have been generally improved. However, this means that error types may differ. In particularly, a ValueError is now always raised when parsing of a single entry fails.

(gh-20580)

Improvements

ndarray.__array_finalize__ is now callable

This means subclasses can now use super().__array_finalize__(obj) without worrying whether ndarray is their superclass or not. The actual call remains a no-op.

(gh-20766)

Add support for VSX4/Power10

With VSX4/Power10 enablement, the new instructions available in Power ISA 3.1 can be used to accelerate some NumPy operations, e.g., floor_divide, modulo, etc.

(gh-20821)

np.fromiter now accepts objects and subarrays

The numpy.fromiter function now supports object and subarray dtypes. Please see he function documentation for examples.

(gh-20993)

Math C library feature detection now uses correct signatures

Compiling is preceded by a detection phase to determine whether the underlying libc supports certain math operations. Previously this code did not respect the proper signatures. Fixing this enables compilation for the wasm-ld backend (compilation for web assembly) and reduces the number of warnings.

(gh-21154)

np.kron now maintains subclass information

np.kron maintains subclass information now such as masked arrays while computing the Kronecker product of the inputs

>>> x = ma.array([[1, 2], [3, 4]], mask=[[0, 1], [1, 0]])
>>> np.kron(x,x)
masked_array(
  data=[[1, --, --, --],
        [--, 4, --, --],
        [--, --, 4, --],
        [--, --, --, 16]],
  mask=[[False,  True,  True,  True],
        [ True, False,  True,  True],
        [ True,  True, False,  True],
        [ True,  True,  True, False]],
  fill_value=999999)

:warning: Warning, np.kron output now follows ufunc ordering (multiply) to determine the output class type

>>> class myarr(np.ndarray):
>>>    __array_priority__ = -1
>>> a = np.ones([2, 2])
>>> ma = myarray(a.shape, a.dtype, a.data)
>>> type(np.kron(a, ma)) == np.ndarray
False # Before it was True
>>> type(np.kron(a, ma)) == myarr
True

(gh-21262)

Performance improvements and changes

Faster np.loadtxt

numpy.loadtxt is now generally much faster than previously as most of it is now implemented in C.

(gh-20580)

Faster reduction operators

Reduction operations like numpy.sum, numpy.prod, numpy.add.reduce, numpy.logical_and.reduce on contiguous integer-based arrays are now much faster.

(gh-21001)

Faster np.where

numpy.where is now much faster than previously on unpredictable/random input data.

(gh-21130)

Faster operations on NumPy scalars

Many operations on NumPy scalars are now significantly faster, although rare operations (e.g. with 0-D arrays rather than scalars) may be slower in some cases. However, even with these improvements users who want the best performance for their scalars, may want to convert a known NumPy scalar into a Python one using scalar.item().

(gh-21188)

Faster np.kron

numpy.kron is about 80% faster as the product is now computed using broadcasting.

(gh-21354)

Checksums

MD5

e82b403f04ff9696804199b727c76a16  numpy-1.23.0rc2-cp310-cp310-macosx_10_9_x86_64.whl
8c54a8d758645f3cc4cd38d3e8749b53  numpy-1.23.0rc2-cp310-cp310-macosx_11_0_arm64.whl
9cb69a99ecc262a0618d1022abefddf3  numpy-1.23.0rc2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
fdc8f769f9d7f49621810dd6e93c4b7b  numpy-1.23.0rc2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
b4e4780b9893cafb7fdf9e77a917621a  numpy-1.23.0rc2-cp310-cp310-win32.whl
72522edc1356681cb2282c0ab846e185  numpy-1.23.0rc2-cp310-cp310-win_amd64.whl
e41ddd480f1a4df9fbd19613e8fe0279  numpy-1.23.0rc2-cp38-cp38-macosx_10_9_x86_64.whl
c072a129bed27a890f40fe4ae92c85d1  numpy-1.23.0rc2-cp38-cp38-macosx_11_0_arm64.whl
3403871b2b11afb71e157da8b5e77bde  numpy-1.23.0rc2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
4bfc243c9d418c3012790c916821b1a1  numpy-1.23.0rc2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
71da5294cd1c7787c502b5dd3868e810  numpy-1.23.0rc2-cp38-cp38-win32.whl
82050d9ffdd0c949bdbb431cadd1e594  numpy-1.23.0rc2-cp38-cp38-win_amd64.whl
ba5b01751b659d1d7db38ebd98c43230  numpy-1.23.0rc2-cp39-cp39-macosx_10_9_x86_64.whl
3f135b3b0e0bb45739c422c3a7eaf71d  numpy-1.23.0rc2-cp39-cp39-macosx_11_0_arm64.whl
d4c2891a953d0bbc746aaaf21e3b39b2  numpy-1.23.0rc2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
cd9afa01451169e5b09037db846948fb  numpy-1.23.0rc2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
e58528377aabfde966bb965fc4df24d3  numpy-1.23.0rc2-cp39-cp39-win32.whl
b07488787ebc63d8ba7885625231f14a  numpy-1.23.0rc2-cp39-cp39-win_amd64.whl
11c6aa572a7337e0245043bd7e442db1  numpy-1.23.0rc2-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
9ab414aa7e01e45faf2e28c7312d16e6  numpy-1.23.0rc2-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
d1f5b0c80a9eea0b57df5a5b03d2a981  numpy-1.23.0rc2-pp38-pypy38_pp73-win_amd64.whl
00884782fc17920a470641ae5c19ad31  numpy-1.23.0rc2.tar.gz

SHA256

2e66decdea13ae8091ba480209dd5ce31261fa3b021ec06b30bd2f4a304861b7  numpy-1.23.0rc2-cp310-cp310-macosx_10_9_x86_64.whl
fde47931544086a648b12ee7c9ccf30edd6c6db776005fb07e4a019a04980042  numpy-1.23.0rc2-cp310-cp310-macosx_11_0_arm64.whl
c308afc8ec782badd073999385a6c93c27ee68e6c0991697394d4fd56566af1f  numpy-1.23.0rc2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
1f22f6f3cb7094ad77c8d352e4bfd2c1db1c38bc08d0b6c74e9b46343c53b052  numpy-1.23.0rc2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
279dce16b143bc50d49bab52dc279d6ab5b0edc7f4d2cc7edaf6a547586bda7e  numpy-1.23.0rc2-cp310-cp310-win32.whl
04e4dbe6b777e977813e7ff5f43aa030ef4f6f75cbc1a4504d3135942b5c12fe  numpy-1.23.0rc2-cp310-cp310-win_amd64.whl
9793feff4758c68502f7652fab08e5ec427d9973d26014767cc15c1b1d885f56  numpy-1.23.0rc2-cp38-cp38-macosx_10_9_x86_64.whl
804293d9bdf33f9c9fb0b4a753f9e84114bb0ad538d184fc579b30782326c827  numpy-1.23.0rc2-cp38-cp38-macosx_11_0_arm64.whl
c335800064f04e0b474b64779ab234ae23c0a5b2f5a06284bb07d297d73692bd  numpy-1.23.0rc2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
9810b840a751b6f0c73c21fb2a50e306d7d0be4114cded4c7d069e142ce488cf  numpy-1.23.0rc2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
dbc987d14f46ae4c476068543d3ad2a20e7ebcb06b211eb4292224dc136eb01d  numpy-1.23.0rc2-cp38-cp38-win32.whl
020218fc82390f1d537cb193d6f1449a919ec97df69b5a64c0a1d017486e0032  numpy-1.23.0rc2-cp38-cp38-win_amd64.whl
785d6520f7bf10ff188762bc460579d6a31c11f960976b2a29efc383b0346572  numpy-1.23.0rc2-cp39-cp39-macosx_10_9_x86_64.whl
ae7e8801b93124a6b0becedc06285ddbaca2daab2d30e35ea413d3bec252717a  numpy-1.23.0rc2-cp39-cp39-macosx_11_0_arm64.whl
1c881827ff0ad7d607047c19a075a7d7c7125cc103fb969a9200bad26175fb9d  numpy-1.23.0rc2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
6fbd492bead87ab83240c56b3490ac301595ab1399ace3e3c1b7c130e3529358  numpy-1.23.0rc2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
82e69890c394a4e1cbcaf12b47d8477bbac4635866fc46a77670abbe4bb4085d  numpy-1.23.0rc2-cp39-cp39-win32.whl
f5a1c7c45ff29db501f9e38a360aedd833e355c14c75155ba2bd46ee3799e30a  numpy-1.23.0rc2-cp39-cp39-win_amd64.whl
b7be00b0a76384490845395714e62f597e64bc6dc8f8a14be0e96034dde3667c  numpy-1.23.0rc2-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
d17f7feb2cca596daa4b3dae86b611a13e9ace061e6583a8db21841f529ca891  numpy-1.23.0rc2-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
548f4d86aa259a448f2da0c07df070bf1f71b68c1f84b1356d4a2ed832598758  numpy-1.23.0rc2-pp38-pypy38_pp73-win_amd64.whl
8e8a88657c028b8b77f3df6f266a5e6ffb4419cbc3dfb525cbbb80ba710f5da2  numpy-1.23.0rc2.tar.gz

v1.23.0rc1

4 months ago

NumPy 1.23.0 Release Notes

The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. The highlights are:

  • Implementation of loadtxt in C, greatly improving its performance.
  • Exposing DLPack at the Python level for easy data exchange.
  • Changes to the promotion and comparisons of structured dtypes.
  • Improvements to f2py.

See below for the details,

New functions

  • A masked array specialization of ndenumerate is now available as numpy.ma.ndenumerate. It provides an alternative to numpy.ndenumerate and skips masked values by default.

    (gh-20020)

  • numpy.from_dlpack has been added to allow easy exchange of data using the DLPack protocol. It accepts Python objects that implement the __dlpack__ and __dlpack_device__ methods and returns a ndarray object which is generally the view of the data of the input object.

    (gh-21145)

Deprecations

  • Setting __array_finalize__ to None is deprecated. It must now be a method and may wish to call super().__array_finalize__(obj) after checking for None or if the NumPy version is sufficiently new.

    (gh-20766)

  • Using axis=32 (axis=np.MAXDIMS) in many cases had the same meaning as axis=None. This is deprecated and axis=None must be used instead.

    (gh-20920)

  • The hook function PyDataMem_SetEventHook has been deprecated and the demonstration of its use in tool/allocation_tracking has been removed. The ability to track allocations is now built-in to python via tracemalloc.

    (gh-20394)

  • numpy.distutils has been deprecated, as a result of distutils itself being deprecated. It will not be present in NumPy for Python >= 3.12, and will be removed completely 2 years after the release of Python 3.12 For more details, see distutils-status-migration{.interpreted-text role="ref"}.

    (gh-20875)

Expired deprecations

  • The NpzFile.iteritems() and NpzFile.iterkeys() methods have been removed as part of the continued removal of Python 2 compatibility. This concludes the deprecation from 1.15.

    (gh-16830)

  • The alen and asscalar functions have been removed.

    (gh-20414)

  • The UPDATEIFCOPY array flag has been removed together with the enum NPY_ARRAY_UPDATEIFCOPY. The associated (and deprecated) PyArray_XDECREF_ERR was also removed. These were all deprecated in 1.14. They are replaced by WRITEBACKIFCOPY, that requires calling PyArray_ResoveWritebackIfCopy before the array is deallocated.

    (gh-20589)

  • Exceptions will be raised during array-like creation. When an object raised an exception during access of the special attributes __array__ or __array_interface__, this exception was usually ignored. This behaviour was deprecated in 1.21, and the exception will now be raised.

    (gh-20835)

  • Multidimensional indexing with non-tuple values is not allowed. Previously, code such as arr[ind] where ind = [[0, 1], [0, 1]] produced a FutureWarning and was interpreted as a multidimensional index (i.e., arr[tuple(ind)]). Now this example is treated like an array index over a single dimension (arr[array(ind)]). Multidimensional indexing with anything but a tuple was deprecated in NumPy 1.15.

    (gh-21029)

  • Changing to a dtype of different size in F-contiguous arrays is no longer permitted. Deprecated since Numpy 1.11.0. See below for an extended explanation of the effects of this change.

    (gh-20722)

New Features

crackfortran has support for operator and assignment overloading

crackfortran parser now understands operator and assignment definitions in a module. They are added in the body list of the module which contains a new key implementedby listing the names of the subroutines or functions implementing the operator or assignment.

(gh-15006)

f2py supports reading access type attributes from derived type statements

As a result, one does not need to use public or private statements to specify derived type access properties.

(gh-15844)

New parameter ndmin added to genfromtxt

This parameter behaves the same as ndmin from numpy.loadtxt.

(gh-20500)

np.loadtxt now supports quote character and single converter function

numpy.loadtxt now supports an additional quotechar keyword argument which is not set by default. Using quotechar='"' will read quoted fields as used by the Excel CSV dialect.

Further, it is now possible to pass a single callable rather than a dictionary for the converters argument.

(gh-20580)

Changing to dtype of a different size now requires contiguity of only the last axis

Previously, viewing an array with a dtype of a different item size required that the entire array be C-contiguous. This limitation would unnecessarily force the user to make contiguous copies of non-contiguous arrays before being able to change the dtype.

This change affects not only ndarray.view, but other construction mechanisms, including the discouraged direct assignment to ndarray.dtype.

This change expires the deprecation regarding the viewing of F-contiguous arrays, described elsewhere in the release notes.

(gh-20722)

Deterministic output files for F2PY

For F77 inputs, f2py will generate modname-f2pywrappers.f unconditionally, though these may be empty. For free-form inputs, modname-f2pywrappers.f, modname-f2pywrappers2.f90 will both be generated unconditionally, and may be empty. This allows writing generic output rules in cmake or meson and other build systems. Older behavior can be restored by passing --skip-empty-wrappers to f2py. f2py-meson{.interpreted-text role="ref"} details usage.

(gh-21187)

keepdims parameter for average

The parameter keepdims was added to the functions numpy.average and numpy.ma.average. The parameter has the same meaning as it does in reduction functions such as numpy.sum or numpy.mean.

(gh-21485)

Compatibility notes

1D np.linalg.norm preserves float input types, even for scalar results

Previously, this would promote to float64 when the ord argument was not one of the explicitly listed values, e.g. ord=3:

>>> f32 = np.float32([1, 2])
>>> np.linalg.norm(f32, 2).dtype
dtype('float32')
>>> np.linalg.norm(f32, 3)
dtype('float64')  # numpy 1.22
dtype('float32')  # numpy 1.23

This change affects only float32 and float16 vectors with ord other than -Inf, 0, 1, 2, and Inf.

(gh-17709)

Changes to structured (void) dtype promotion and comparisons

In general, NumPy now defines correct, but slightly limited, promotion for structured dtypes by promoting the subtypes of each field instead of raising an exception:

>>> np.result_type(np.dtype("i,i"), np.dtype("i,d"))
dtype([('f0', '<i4'), ('f1', '<f8')])

For promotion matching field names, order, and titles are enforced, however padding is ignored. Promotion involving structured dtypes now always ensures native byte-order for all fields (which may change the result of np.concatenate) and ensures that the result will be "packed", i.e. all fields are ordered contiguously and padding is removed. See structured_dtype_comparison_and_promotion{.interpreted-text role="ref"} for further details.

The repr of aligned structures will now never print the long form including offsets and itemsize unless the structure includes padding not guaranteed by align=True.

In alignment with the above changes to the promotion logic, the casting safety has been updated:

  • "equiv" enforces matching names and titles. The itemsize is allowed to differ due to padding.
  • "safe" allows mismatching field names and titles
  • The cast safety is limited by the cast safety of each included field.
  • The order of fields is used to decide cast safety of each individual field. Previously, the field names were used and only unsafe casts were possible when names mismatched.

The main important change here is that name mismatches are now considered "safe" casts.

(gh-19226)

NPY_RELAXED_STRIDES_CHECKING has been removed

NumPy cannot be compiled with NPY_RELAXED_STRIDES_CHECKING=0 anymore. Relaxed strides have been the default for many years and the option was initially introduced to allow a smoother transition.

(gh-20220)

np.loadtxt has recieved several changes

The row counting of numpy.loadtxt was fixed. loadtxt ignores fully empty lines in the file, but counted them towards max_rows. When max_rows is used and the file contains empty lines, these will now not be counted. Previously, it was possible that the result contained fewer than max_rows rows even though more data was available to be read. If the old behaviour is required, itertools.islice may be used:

import itertools
lines = itertools.islice(open("file"), 0, max_rows)
result = np.loadtxt(lines, ...)

While generally much faster and improved, numpy.loadtxt may now fail to converter certain strings to numbers that were previously successfully read. The most important cases for this are:

  • Parsing floating point values such as 1.0 into integers will now fail
  • Parsing hexadecimal floats such as 0x3p3 will fail
  • An _ was previously accepted as a thousands delimiter 100_000. This will now result in an error.

If you experience these limitations, they can all be worked around by passing appropriate converters=. NumPy now supports passing a single converter to be used for all columns to make this more convenient. For example, converters=float.fromhex can read hexadecimal float numbers and converters=int will be able to read 100_000.

Further, the error messages have been generally improved. However, this means that error types may differ. In particularly, a ValueError is now always raised when parsing of a single entry fails.

(gh-20580)

Improvements

ndarray.__array_finalize__ is now callable

This means subclasses can now use super().__array_finalize__(obj) without worrying whether ndarray is their superclass or not. The actual call remains a no-op.

(gh-20766)

Add support for VSX4/Power10

With VSX4/Power10 enablement, the new instructions available in Power ISA 3.1 can be used to accelerate some NumPy operations, e.g., floor_divide, modulo, etc.

(gh-20821)

np.fromiter now accepts objects and subarrays

The numpy.fromiter function now supports object and subarray dtypes. Please see he function documentation for examples.

(gh-20993)

Math C library feature detection now uses correct signatures

Compiling is preceded by a detection phase to determine whether the underlying libc supports certain math operations. Previously this code did not respect the proper signatures. Fixing this enables compilation for the wasm-ld backend (compilation for web assembly) and reduces the number of warnings.

(gh-21154)

np.kron now maintains subclass information

np.kron maintains subclass information now such as masked arrays while computing the Kronecker product of the inputs

>>> x = ma.array([[1, 2], [3, 4]], mask=[[0, 1], [1, 0]])
>>> np.kron(x,x)
masked_array(
  data=[[1, --, --, --],
        [--, 4, --, --],
        [--, --, 4, --],
        [--, --, --, 16]],
  mask=[[False,  True,  True,  True],
        [ True, False,  True,  True],
        [ True,  True, False,  True],
        [ True,  True,  True, False]],
  fill_value=999999)

:warning: Warning, np.kron output now follows ufunc ordering (multiply) to determine the output class type

>>> class myarr(np.ndarray):
>>>    __array_priority__ = -1
>>> a = np.ones([2, 2])
>>> ma = myarray(a.shape, a.dtype, a.data)
>>> type(np.kron(a, ma)) == np.ndarray
False # Before it was True
>>> type(np.kron(a, ma)) == myarr
True

(gh-21262)

Performance improvements and changes

Faster np.loadtxt

numpy.loadtxt is now generally much faster than previously as most of it is now implemented in C.

(gh-20580)

Faster reduction operators

Reduction operations like numpy.sum, numpy.prod, numpy.add.reduce, numpy.logical_and.reduce on contiguous integer-based arrays are now much faster.

(gh-21001)

Faster np.where

numpy.where is now much faster than previously on unpredictable/random input data.

(gh-21130)

Faster operations on NumPy scalars

Many operations on NumPy scalars are now significantly faster, although rare operations (e.g. with 0-D arrays rather than scalars) may be slower in some cases. However, even with these improvements users who want the best performance for their scalars, may want to convert a known NumPy scalar into a Python one using scalar.item().

(gh-21188)

Faster np.kron

numpy.kron is about 80% faster as the product is now computed using broadcasting.

(gh-21354)

Checksums

MD5

c67b4cc1de8a0753bc65765a508aa0e3  numpy-1.23.0rc1-cp310-cp310-macosx_10_9_x86_64.whl
9daceb162c46298986b5fceb13f10e54  numpy-1.23.0rc1-cp310-cp310-macosx_11_0_arm64.whl
ac07046b70001710d8d3243d9b5d0389  numpy-1.23.0rc1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
14269d197cd6aac02655d43aa10ba108  numpy-1.23.0rc1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
ba5fc5cd776549afc353e0043f6c6f5a  numpy-1.23.0rc1-cp310-cp310-win32.whl
0e1b00f156f32aabde1e29607c709a24  numpy-1.23.0rc1-cp310-cp310-win_amd64.whl
34b5a9f3abeb9f6e9c6fbd494305d53d  numpy-1.23.0rc1-cp38-cp38-macosx_10_9_x86_64.whl
bb825cf372822daad1e440577e324042  numpy-1.23.0rc1-cp38-cp38-macosx_11_0_arm64.whl
45296e6b41691c4265c90562c94254aa  numpy-1.23.0rc1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
b10b131e5c0576629ab99829301d6fba  numpy-1.23.0rc1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
bc6210254087b73715d8c6a79dafa3b8  numpy-1.23.0rc1-cp38-cp38-win32.whl
d9b7fb5a539a738309a717051f13e41a  numpy-1.23.0rc1-cp38-cp38-win_amd64.whl
d2cec33489c96dfc489bb00353d351fe  numpy-1.23.0rc1-cp39-cp39-macosx_10_9_x86_64.whl
e4982f3bf3d4acab67cb61d3d0e2f85b  numpy-1.23.0rc1-cp39-cp39-macosx_11_0_arm64.whl
781a79ea9f1683a579a1ef27d809a8e0  numpy-1.23.0rc1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
8c03ea50c2baa172e8252d10dea73498  numpy-1.23.0rc1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
61829dec0785cf72b8f5bc92dc44d828  numpy-1.23.0rc1-cp39-cp39-win32.whl
94c7c971ed5eedb1b61a09bcfc123617  numpy-1.23.0rc1-cp39-cp39-win_amd64.whl
b5c6f674b468e7fc513882563391efec  numpy-1.23.0rc1-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
41ab4b757478c8e244018c37bcb52bb3  numpy-1.23.0rc1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
f2082734772a6a7afbe3568e7b2ad458  numpy-1.23.0rc1-pp38-pypy38_pp73-win_amd64.whl
faf6a08cda5696b96acb670c433495e5  numpy-1.23.0rc1.tar.gz

SHA256

a72607e12891615a314a892f8d21301b930f211841f0084d269baa1eb31710b4  numpy-1.23.0rc1-cp310-cp310-macosx_10_9_x86_64.whl
4784a81089c75a941dcc013b09290dfb22768780a1f3525667328d09a3338116  numpy-1.23.0rc1-cp310-cp310-macosx_11_0_arm64.whl
d4ebdeb0e2a57bbd28c3258a562ee011775127427eea833613744af1a66c3e11  numpy-1.23.0rc1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a2dfb54cb1c6470918a3c02da77706f28977cb7eac4b76cc40b14942c8634615  numpy-1.23.0rc1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
1567c488f9ef97341c5937b4140a45ac37e0592c43ba2c59d3e49ff7d5da90b0  numpy-1.23.0rc1-cp310-cp310-win32.whl
78df1fc2ddf543508b5358dd24ac68ee693599e5df0d136062b9ec21ba7643cf  numpy-1.23.0rc1-cp310-cp310-win_amd64.whl
020c6d8476fced48f42629f46996af8a07bc725cb821081205c4422eacaa2283  numpy-1.23.0rc1-cp38-cp38-macosx_10_9_x86_64.whl
efd26eecd1ada0c8dadc5b221c52086ccc72e4cb0707e451889ef3b62c14163c  numpy-1.23.0rc1-cp38-cp38-macosx_11_0_arm64.whl
8cf3f61984777a830eef452d8b04338795691949214e6cafc46c5236900cd1f5  numpy-1.23.0rc1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
96738ae39db010502564325ce2f4aff4f42b75adf64f3ccb2b19214e9be1c01c  numpy-1.23.0rc1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
33b233f59d9430a27c2a58a056f32259eadf9584f41c6ec02c493c3aeb90f844  numpy-1.23.0rc1-cp38-cp38-win32.whl
edf0a82d285e18418e3915fece2cf0f4f31e84fe62271331fbaafbbc7d57e9ee  numpy-1.23.0rc1-cp38-cp38-win_amd64.whl
a71f1602bf84d0a2fb5d586a2d8c31f29fbca9253ae1eecf46b7059fa265eb79  numpy-1.23.0rc1-cp39-cp39-macosx_10_9_x86_64.whl
05000d27fd135dd0aab90acaf96652991c070dda688739097ac2dea92189f9f0  numpy-1.23.0rc1-cp39-cp39-macosx_11_0_arm64.whl
ebe07758ac3e7402290f43d379f6d79d81a247488561743490cf2e5b64351ba6  numpy-1.23.0rc1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
3cd05784cdcd09114c2f6186bb99af7f5ee65ffd720dae9990722a94309b17ea  numpy-1.23.0rc1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
6ccb79435d4501b35ed3d807e1bf7345e42f68b25fbf720ade9c74c7196360f9  numpy-1.23.0rc1-cp39-cp39-win32.whl
a8fbe61e09565fa2f7bca076627ea0efbf50ab689c35af5082c5d94fb24b30ee  numpy-1.23.0rc1-cp39-cp39-win_amd64.whl
7a45352476e92c1958ce513fa84b508d59dd8e6ffe0e6f6cceebfc0f3c06d086  numpy-1.23.0rc1-pp38-pypy38_pp73-macosx_10_9_x86_64.whl
4f15768493ecf23c5d82e5542642a36764e551c7744781268c7c221f26c7ffd6  numpy-1.23.0rc1-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
edf0b720c8ba3d35b23c71c0cd13df34290be87b42f0e10d0ec2f1639cda2692  numpy-1.23.0rc1-pp38-pypy38_pp73-win_amd64.whl
3a09d0f564f59b6da54f592909d3fdbd50b492ef9fbe6d699043c992538ba0e0  numpy-1.23.0rc1.tar.gz

v1.22.4

4 months ago

NumPy 1.22.4 Release Notes

NumPy 1.22.4 is a maintenance release that fixes bugs discovered after the 1.22.3 release. In addition, the wheels for this release are built using the recently released Cython 0.29.30, which should fix the reported problems with debugging.

The Python versions supported for this release are 3.8-3.10. Note that the Mac wheels are now based on OS X 10.15 rather than 10.6 that was used in previous NumPy release cycles.

Contributors

A total of 12 people contributed to this release. People with a "+" by their names contributed a patch for the first time.

  • Alexander Shadchin
  • Bas van Beek
  • Charles Harris
  • Hood Chatham
  • Jarrod Millman
  • John-Mark Gurney +
  • Junyan Ou +
  • Mariusz Felisiak +
  • Ross Barnowski
  • Sebastian Berg
  • Serge Guelton
  • Stefan van der Walt

Pull requests merged

A total of 22 pull requests were merged for this release.

  • #21191: TYP, BUG: Fix np.lib.stride_tricks re-exported under the...
  • #21192: TST: Bump mypy from 0.931 to 0.940
  • #21243: MAINT: Explicitly re-export the types in numpy._typing
  • #21245: MAINT: Specify sphinx, numpydoc versions for CI doc builds
  • #21275: BUG: Fix typos
  • #21277: ENH, BLD: Fix math feature detection for wasm
  • #21350: MAINT: Fix failing simd and cygwin tests.
  • #21438: MAINT: Fix failing Python 3.8 32-bit Windows test.
  • #21444: BUG: add linux guard per #21386
  • #21445: BUG: Allow legacy dtypes to cast to datetime again
  • #21446: BUG: Make mmap handling safer in frombuffer
  • #21447: BUG: Stop using PyBytesObject.ob_shash deprecated in Python 3.11.
  • #21448: ENH: Introduce numpy.core.setup_common.NPY_CXX_FLAGS
  • #21472: BUG: Ensure compile errors are raised correclty
  • #21473: BUG: Fix segmentation fault
  • #21474: MAINT: Update doc requirements
  • #21475: MAINT: Mark npy_memchr with no_sanitize("alignment") on clang
  • #21512: DOC: Proposal - make the doc landing page cards more similar...
  • #21525: MAINT: Update Cython version to 0.29.30.
  • #21536: BUG: Fix GCC error during build configuration
  • #21541: REL: Prepare for the NumPy 1.22.4 release.
  • #21547: MAINT: Skip tests that fail on PyPy.

Checksums

MD5

a19351fd3dc0b3bbc733495ed18b8f24  numpy-1.22.4-cp310-cp310-macosx_10_14_x86_64.whl
0730f9e196f70ad89f246bf95ccf05d5  numpy-1.22.4-cp310-cp310-macosx_10_15_x86_64.whl
63c74e5395a2b31d8adc5b1aa0c62471  numpy-1.22.4-cp310-cp310-macosx_11_0_arm64.whl
f99778023770c12f896768c90f7712e5  numpy-1.22.4-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
757d68b0cdb4e28ffce8574b6a2f3c5e  numpy-1.22.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
50becf2e048e54dc5227dfe8378aae1e  numpy-1.22.4-cp310-cp310-win32.whl
79dfdc29a4730e44d6df33dbea5b35b0  numpy-1.22.4-cp310-cp310-win_amd64.whl
8fd8f04d71ead55c2773d1b46668ca67  numpy-1.22.4-cp38-cp38-macosx_10_15_x86_64.whl
41a7c6240081010824cc0d5c02900fe6  numpy-1.22.4-cp38-cp38-macosx_11_0_arm64.whl
6bc066d3f61da3304c82d92f3f900a4f  numpy-1.22.4-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
86d959605c66ccba11c6504f25fff0d7  numpy-1.22.4-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
ae0405894c065349a511e4575b919e2a  numpy-1.22.4-cp38-cp38-win32.whl
c9a731d08081396b7a1b66977734d2ac  numpy-1.22.4-cp38-cp38-win_amd64.whl
4d9b97d74799e5fc48860f0b4a3b255a  numpy-1.22.4-cp39-cp39-macosx_10_14_x86_64.whl
c99fa7e04cb7cc23f1713f2023b4e489  numpy-1.22.4-cp39-cp39-macosx_10_15_x86_64.whl
dda3815df12b8a99c6c3069f69997521  numpy-1.22.4-cp39-cp39-macosx_11_0_arm64.whl
9b7c5b39d5611d92b66eb545d44b25db  numpy-1.22.4-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
90fc45eaf8b8c4fac3f3ebd105a5a856  numpy-1.22.4-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
9562153d4a83d773c20eb626cbd65cde  numpy-1.22.4-cp39-cp39-win32.whl
711b23acce54a18ce74fc80f48f48062  numpy-1.22.4-cp39-cp39-win_amd64.whl
ab803b24ea557452e828adba1b986af3  numpy-1.22.4-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
09b3a41ea0b9bc20bd1691cf88f0b0d3  numpy-1.22.4.tar.gz
b44849506fbb54cdef9dbb435b2b1987  numpy-1.22.4.zip

SHA256

ba9ead61dfb5d971d77b6c131a9dbee62294a932bf6a356e48c75ae684e635b3  numpy-1.22.4-cp310-cp310-macosx_10_14_x86_64.whl
1ce7ab2053e36c0a71e7a13a7475bd3b1f54750b4b433adc96313e127b870887  numpy-1.22.4-cp310-cp310-macosx_10_15_x86_64.whl
7228ad13744f63575b3a972d7ee4fd61815b2879998e70930d4ccf9ec721dce0  numpy-1.22.4-cp310-cp310-macosx_11_0_arm64.whl
43a8ca7391b626b4c4fe20aefe79fec683279e31e7c79716863b4b25021e0e74  numpy-1.22.4-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a911e317e8c826ea632205e63ed8507e0dc877dcdc49744584dfc363df9ca08c  numpy-1.22.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
9ce7df0abeabe7fbd8ccbf343dc0db72f68549856b863ae3dd580255d009648e  numpy-1.22.4-cp310-cp310-win32.whl
3e1ffa4748168e1cc8d3cde93f006fe92b5421396221a02f2274aab6ac83b077  numpy-1.22.4-cp310-cp310-win_amd64.whl
59d55e634968b8f77d3fd674a3cf0b96e85147cd6556ec64ade018f27e9479e1  numpy-1.22.4-cp38-cp38-macosx_10_15_x86_64.whl
c1d937820db6e43bec43e8d016b9b3165dcb42892ea9f106c70fb13d430ffe72  numpy-1.22.4-cp38-cp38-macosx_11_0_arm64.whl
d4c5d5eb2ec8da0b4f50c9a843393971f31f1d60be87e0fb0917a49133d257d6  numpy-1.22.4-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
64f56fc53a2d18b1924abd15745e30d82a5782b2cab3429aceecc6875bd5add0  numpy-1.22.4-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
fb7a980c81dd932381f8228a426df8aeb70d59bbcda2af075b627bbc50207cba  numpy-1.22.4-cp38-cp38-win32.whl
e96d7f3096a36c8754207ab89d4b3282ba7b49ea140e4973591852c77d09eb76  numpy-1.22.4-cp38-cp38-win_amd64.whl
4c6036521f11a731ce0648f10c18ae66d7143865f19f7299943c985cdc95afb5  numpy-1.22.4-cp39-cp39-macosx_10_14_x86_64.whl
b89bf9b94b3d624e7bb480344e91f68c1c6c75f026ed6755955117de00917a7c  numpy-1.22.4-cp39-cp39-macosx_10_15_x86_64.whl
2d487e06ecbf1dc2f18e7efce82ded4f705f4bd0cd02677ffccfb39e5c284c7e  numpy-1.22.4-cp39-cp39-macosx_11_0_arm64.whl
f3eb268dbd5cfaffd9448113539e44e2dd1c5ca9ce25576f7c04a5453edc26fa  numpy-1.22.4-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
37431a77ceb9307c28382c9773da9f306435135fae6b80b62a11c53cfedd8802  numpy-1.22.4-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
cc7f00008eb7d3f2489fca6f334ec19ca63e31371be28fd5dad955b16ec285bd  numpy-1.22.4-cp39-cp39-win32.whl
f0725df166cf4785c0bc4cbfb320203182b1ecd30fee6e541c8752a92df6aa32  numpy-1.22.4-cp39-cp39-win_amd64.whl
0791fbd1e43bf74b3502133207e378901272f3c156c4df4954cad833b1380207  numpy-1.22.4-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
b4308198d0e41efaa108e57d69973398439c7299a9d551680cdd603cf6d20709  numpy-1.22.4.tar.gz
425b390e4619f58d8526b3dcf656dde069133ae5c240229821f01b5f44ea07af  numpy-1.22.4.zip

v1.21.6

5 months ago

NumPy 1.21.6 Release Notes

NumPy 1.21.6 is a very small release that achieves two things:

  • Backs out the mistaken backport of C++ code into 1.21.5.
  • Provides a 32 bit Windows wheel for Python 3.10.

The provision of the 32 bit wheel is intended to make life easier for oldest-supported-numpy.

Checksums

MD5

5a3e5d7298056bcfbc3246597af474d4  numpy-1.21.6-cp310-cp310-macosx_10_9_universal2.whl
d981d2859842e7b62dc93e24808c7bac  numpy-1.21.6-cp310-cp310-macosx_10_9_x86_64.whl
171313893c26529404d09fadb3537ed3  numpy-1.21.6-cp310-cp310-macosx_11_0_arm64.whl
5a7a6dfdd43069f9b29d3fe6b7f3a2ce  numpy-1.21.6-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a9e25375a72725c5d74442eda53af405  numpy-1.21.6-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
6f9a782477380b2cdb7606f6f7634c00  numpy-1.21.6-cp310-cp310-win32.whl
32a73a348864700a3fa510d2fc4350b7  numpy-1.21.6-cp310-cp310-win_amd64.whl
0db8941ebeb0a02cd839d9cd3c5c20bb  numpy-1.21.6-cp37-cp37m-macosx_10_9_x86_64.whl
67882155be9592850861f4ad8ba36623  numpy-1.21.6-cp37-cp37m-manylinux_2_12_i686.manylinux2010_i686.whl
c70e30e1ff9ab49f898c19e7a6492ae6  numpy-1.21.6-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
e32dbd291032c7554a742f1bb9b2f7a3  numpy-1.21.6-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
689bf804c2cd16cb241fd943e3833ffd  numpy-1.21.6-cp37-cp37m-win32.whl
0062a7b0231a07cb5b9f3d7c495e6fe4  numpy-1.21.6-cp37-cp37m-win_amd64.whl
0d08809980ab497659e7aa0df9ce120e  numpy-1.21.6-cp38-cp38-macosx_10_9_universal2.whl
3c67d14ea2009069844b27bfbf74304d  numpy-1.21.6-cp38-cp38-macosx_10_9_x86_64.whl
5f0e773745cb817313232ac1bf4c7eee  numpy-1.21.6-cp38-cp38-macosx_11_0_arm64.whl
fa8011e065f1964d3eb870bb3926fc99  numpy-1.21.6-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl
486cf9d4daab59aad253aa5b84a5aa83  numpy-1.21.6-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
88509abab303c076dfb26f00e455180d  numpy-1.21.6-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
f7234e2ef837f5f6ddbde8db246fd05b  numpy-1.21.6-cp38-cp38-win32.whl
e1063e01fb44ea7a49adea0c33548217  numpy-1.21.6-cp38-cp38-win_amd64.whl
61c4caad729e3e0e688accbc1424ed45  numpy-1.21.6-cp39-cp39-macosx_10_9_universal2.whl
67488d8ccaeff798f2e314aae7c4c3d6  numpy-1.21.6-cp39-cp39-macosx_10_9_x86_64.whl
128c3713b5d1de45a0f522562bac5263  numpy-1.21.6-cp39-cp39-macosx_11_0_arm64.whl
50e79cd0610b4ed726b3bf08c3716dab  numpy-1.21.6-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl
bd0c9e3c0e488faac61daf3227fb95af  numpy-1.21.6-cp39-cp39-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
aa5e9baf1dec16b15e481c23f8a23214  numpy-1.21.6-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a2405b0e5d3f775ad30177296a997092  numpy-1.21.6-cp39-cp39-win32.whl
f0d20eda8c78f957ea70c5527954303e  numpy-1.21.6-cp39-cp39-win_amd64.whl
9682abbcc38cccb7f56e48aacca7de23  numpy-1.21.6-pp37-pypy37_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
6aa3c2e8ea2886bf593bd8e0a1425c64  numpy-1.21.6.tar.gz
04aea95dcb1d256d13a45df42173aa1e  numpy-1.21.6.zip

SHA256

8737609c3bbdd48e380d463134a35ffad3b22dc56295eff6f79fd85bd0eeeb25  numpy-1.21.6-cp310-cp310-macosx_10_9_universal2.whl
fdffbfb6832cd0b300995a2b08b8f6fa9f6e856d562800fea9182316d99c4e8e  numpy-1.21.6-cp310-cp310-macosx_10_9_x86_64.whl
3820724272f9913b597ccd13a467cc492a0da6b05df26ea09e78b171a0bb9da6  numpy-1.21.6-cp310-cp310-macosx_11_0_arm64.whl
f17e562de9edf691a42ddb1eb4a5541c20dd3f9e65b09ded2beb0799c0cf29bb  numpy-1.21.6-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
5f30427731561ce75d7048ac254dbe47a2ba576229250fb60f0fb74db96501a1  numpy-1.21.6-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
d4bf4d43077db55589ffc9009c0ba0a94fa4908b9586d6ccce2e0b164c86303c  numpy-1.21.6-cp310-cp310-win32.whl
d136337ae3cc69aa5e447e78d8e1514be8c3ec9b54264e680cf0b4bd9011574f  numpy-1.21.6-cp310-cp310-win_amd64.whl
6aaf96c7f8cebc220cdfc03f1d5a31952f027dda050e5a703a0d1c396075e3e7  numpy-1.21.6-cp37-cp37m-macosx_10_9_x86_64.whl
67c261d6c0a9981820c3a149d255a76918278a6b03b6a036800359aba1256d46  numpy-1.21.6-cp37-cp37m-manylinux_2_12_i686.manylinux2010_i686.whl
a6be4cb0ef3b8c9250c19cc122267263093eee7edd4e3fa75395dfda8c17a8e2  numpy-1.21.6-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
7c4068a8c44014b2d55f3c3f574c376b2494ca9cc73d2f1bd692382b6dffe3db  numpy-1.21.6-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
7c7e5fa88d9ff656e067876e4736379cc962d185d5cd808014a8a928d529ef4e  numpy-1.21.6-cp37-cp37m-win32.whl
bcb238c9c96c00d3085b264e5c1a1207672577b93fa666c3b14a45240b14123a  numpy-1.21.6-cp37-cp37m-win_amd64.whl
82691fda7c3f77c90e62da69ae60b5ac08e87e775b09813559f8901a88266552  numpy-1.21.6-cp38-cp38-macosx_10_9_universal2.whl
643843bcc1c50526b3a71cd2ee561cf0d8773f062c8cbaf9ffac9fdf573f83ab  numpy-1.21.6-cp38-cp38-macosx_10_9_x86_64.whl
357768c2e4451ac241465157a3e929b265dfac85d9214074985b1786244f2ef3  numpy-1.21.6-cp38-cp38-macosx_11_0_arm64.whl
9f411b2c3f3d76bba0865b35a425157c5dcf54937f82bbeb3d3c180789dd66a6  numpy-1.21.6-cp38-cp38-manylinux_2_12_i686.manylinux2010_i686.whl
4aa48afdce4660b0076a00d80afa54e8a97cd49f457d68a4342d188a09451c1a  numpy-1.21.6-cp38-cp38-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
d6a96eef20f639e6a97d23e57dd0c1b1069a7b4fd7027482a4c5c451cd7732f4  numpy-1.21.6-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
5c3c8def4230e1b959671eb959083661b4a0d2e9af93ee339c7dada6759a9470  numpy-1.21.6-cp38-cp38-win32.whl
bf2ec4b75d0e9356edea834d1de42b31fe11f726a81dfb2c2112bc1eaa508fcf  numpy-1.21.6-cp38-cp38-win_amd64.whl
4391bd07606be175aafd267ef9bea87cf1b8210c787666ce82073b05f202add1  numpy-1.21.6-cp39-cp39-macosx_10_9_universal2.whl
67f21981ba2f9d7ba9ade60c9e8cbaa8cf8e9ae51673934480e45cf55e953673  numpy-1.21.6-cp39-cp39-macosx_10_9_x86_64.whl
ee5ec40fdd06d62fe5d4084bef4fd50fd4bb6bfd2bf519365f569dc470163ab0  numpy-1.21.6-cp39-cp39-macosx_11_0_arm64.whl
1dbe1c91269f880e364526649a52eff93ac30035507ae980d2fed33aaee633ac  numpy-1.21.6-cp39-cp39-manylinux_2_12_i686.manylinux2010_i686.whl
d9caa9d5e682102453d96a0ee10c7241b72859b01a941a397fd965f23b3e016b  numpy-1.21.6-cp39-cp39-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
58459d3bad03343ac4b1b42ed14d571b8743dc80ccbf27444f266729df1d6f5b  numpy-1.21.6-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
7f5ae4f304257569ef3b948810816bc87c9146e8c446053539947eedeaa32786  numpy-1.21.6-cp39-cp39-win32.whl
e31f0bb5928b793169b87e3d1e070f2342b22d5245c755e2b81caa29756246c3  numpy-1.21.6-cp39-cp39-win_amd64.whl
dd1c8f6bd65d07d3810b90d02eba7997e32abbdf1277a481d698969e921a3be0  numpy-1.21.6-pp37-pypy37_pp73-manylinux_2_12_x86_64.manylinux2010_x86_64.whl
d4efc6491a1cdc00f9eca9bf2c1aa13671776f6941c7321ddf75b45c862f0c2c  numpy-1.21.6.tar.gz
ecb55251139706669fdec2ff073c98ef8e9a84473e51e716211b41aa0f18e656  numpy-1.21.6.zip

v1.22.3

6 months ago

NumPy 1.22.3 Release Notes

NumPy 1.22.3 is a maintenance release that fixes bugs discovered after the 1.22.2 release. The most noticeable fixes may be those for DLPack. One that may cause some problems is disallowing strings as inputs to logical ufuncs. It is still undecided how strings should be treated in those functions and it was thought best to simply disallow them until a decision was reached. That should not cause problems with older code.

The Python versions supported for this release are 3.8-3.10. Note that the Mac wheels are now based on OS X 10.14 rather than 10.9 that was used in previous NumPy release cycles. 10.14 is the oldest release supported by Apple.

Contributors

A total of 9 people contributed to this release. People with a "+" by their names contributed a patch for the first time.

  • @GalaxySnail +
  • Alexandre de Siqueira
  • Bas van Beek
  • Charles Harris
  • Melissa Weber Mendonça
  • Ross Barnowski
  • Sebastian Berg
  • Tirth Patel
  • Matthieu Darbois

Pull requests merged

A total of 10 pull requests were merged for this release.

  • #21048: MAINT: Use "3.10" instead of "3.10-dev" on travis.
  • #21106: TYP,MAINT: Explicitly allow sequences of array-likes in np.concatenate
  • #21137: BLD,DOC: skip broken ipython 8.1.0
  • #21138: BUG, ENH: np._from_dlpack: export correct device information
  • #21139: BUG: Fix numba DUFuncs added loops getting picked up
  • #21140: BUG: Fix unpickling an empty ndarray with a non-zero dimension...
  • #21141: BUG: use ThreadPoolExecutor instead of ThreadPool
  • #21142: API: Disallow strings in logical ufuncs
  • #21143: MAINT, DOC: Fix SciPy intersphinx link
  • #21148: BUG,ENH: np._from_dlpack: export arrays with any strided size-1...

Checksums

MD5

14f1872bbab050b0579e5fcd8b341b81  numpy-1.22.3-cp310-cp310-macosx_10_14_x86_64.whl
c673faa3ac8745ad10ed0428a21a77aa  numpy-1.22.3-cp310-cp310-macosx_11_0_arm64.whl
d925fff720561673fd7ee8ead0e94935  numpy-1.22.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
319f97f5ee26b9c3c06f7a2a3df412a3  numpy-1.22.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
866eae5dba934cad50eb38c8505c8449  numpy-1.22.3-cp310-cp310-win32.whl
e4c512437a6d4eb4a384225861067ad8  numpy-1.22.3-cp310-cp310-win_amd64.whl
a28052af37037f0d5c3b47f4a7040135  numpy-1.22.3-cp38-cp38-macosx_10_14_x86_64.whl
d22dc074bde64f6e91a2d1990345f821  numpy-1.22.3-cp38-cp38-macosx_11_0_arm64.whl
e8a01c2ca1474aff142366a0a2fe0812  numpy-1.22.3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
4fe6e71e7871cb31ffc4122aa5707be7  numpy-1.22.3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
1273fb3c77383ab28f2fb05192751340  numpy-1.22.3-cp38-cp38-win32.whl
001244a6bafa640d7509c85661a4e98e  numpy-1.22.3-cp38-cp38-win_amd64.whl
b8694b880a1a68d1716f60a9c9e82b38  numpy-1.22.3-cp39-cp39-macosx_10_14_x86_64.whl
ba122eaa0988801e250f8674e3dd612e  numpy-1.22.3-cp39-cp39-macosx_11_0_arm64.whl
3641825aca07cb26732425e52d034daf  numpy-1.22.3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
f92412e4273c2580abcc1b75c56e9651  numpy-1.22.3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
b38604778ffd0a17931c06738c3ce9ed  numpy-1.22.3-cp39-cp39-win32.whl
644e0b141fa36a1baf0338032254cc9a  numpy-1.22.3-cp39-cp39-win_amd64.whl
99d2dfb943327b108b2c3b923bd42000  numpy-1.22.3-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
3305c27e5bdf7f19247a7eee00ac053e  numpy-1.22.3.tar.gz
b56530be068796a50bf5a09105c8011e  numpy-1.22.3.zip

SHA256

92bfa69cfbdf7dfc3040978ad09a48091143cffb778ec3b03fa170c494118d75  numpy-1.22.3-cp310-cp310-macosx_10_14_x86_64.whl
8251ed96f38b47b4295b1ae51631de7ffa8260b5b087808ef09a39a9d66c97ab  numpy-1.22.3-cp310-cp310-macosx_11_0_arm64.whl
48a3aecd3b997bf452a2dedb11f4e79bc5bfd21a1d4cc760e703c31d57c84b3e  numpy-1.22.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
a3bae1a2ed00e90b3ba5f7bd0a7c7999b55d609e0c54ceb2b076a25e345fa9f4  numpy-1.22.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
f950f8845b480cffe522913d35567e29dd381b0dc7e4ce6a4a9f9156417d2430  numpy-1.22.3-cp310-cp310-win32.whl
08d9b008d0156c70dc392bb3ab3abb6e7a711383c3247b410b39962263576cd4  numpy-1.22.3-cp310-cp310-win_amd64.whl
201b4d0552831f7250a08d3b38de0d989d6f6e4658b709a02a73c524ccc6ffce  numpy-1.22.3-cp38-cp38-macosx_10_14_x86_64.whl
f8c1f39caad2c896bc0018f699882b345b2a63708008be29b1f355ebf6f933fe  numpy-1.22.3-cp38-cp38-macosx_11_0_arm64.whl
568dfd16224abddafb1cbcce2ff14f522abe037268514dd7e42c6776a1c3f8e5  numpy-1.22.3-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
3ca688e1b9b95d80250bca34b11a05e389b1420d00e87a0d12dc45f131f704a1  numpy-1.22.3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
e7927a589df200c5e23c57970bafbd0cd322459aa7b1ff73b7c2e84d6e3eae62  numpy-1.22.3-cp38-cp38-win32.whl
07a8c89a04997625236c5ecb7afe35a02af3896c8aa01890a849913a2309c676  numpy-1.22.3-cp38-cp38-win_amd64.whl
2c10a93606e0b4b95c9b04b77dc349b398fdfbda382d2a39ba5a822f669a0123  numpy-1.22.3-cp39-cp39-macosx_10_14_x86_64.whl
fade0d4f4d292b6f39951b6836d7a3c7ef5b2347f3c420cd9820a1d90d794802  numpy-1.22.3-cp39-cp39-macosx_11_0_arm64.whl
5bfb1bb598e8229c2d5d48db1860bcf4311337864ea3efdbe1171fb0c5da515d  numpy-1.22.3-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
97098b95aa4e418529099c26558eeb8486e66bd1e53a6b606d684d0c3616b168  numpy-1.22.3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
fdf3c08bce27132395d3c3ba1503cac12e17282358cb4bddc25cc46b0aca07aa  numpy-1.22.3-cp39-cp39-win32.whl
639b54cdf6aa4f82fe37ebf70401bbb74b8508fddcf4797f9fe59615b8c5813a  numpy-1.22.3-cp39-cp39-win_amd64.whl
c34ea7e9d13a70bf2ab64a2532fe149a9aced424cd05a2c4ba662fd989e3e45f  numpy-1.22.3-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
a906c0b4301a3d62ccf66d058fe779a65c1c34f6719ef2058f96e1856f48bca5  numpy-1.22.3.tar.gz
dbc7601a3b7472d559dc7b933b18b4b66f9aa7452c120e87dfb33d02008c8a18  numpy-1.22.3.zip