Bottleneck 1.3.0¶

Release date: 2019-11-12

Project Updates¶

Bottleneck has a new maintainer, Christopher Whelan (@qwhelan on GitHub).
Documentation now hosted at https://bottleneck.readthedocs.io
1.3.x will be the last release to support Python 2.7
Bottleneck now supports and is tested against Python 3.7 and 3.8. (#211, #268)
The LICENSE file has been restructured to only include the license for the Bottleneck project to aid license audit tools. There has been no change to the licensing of Bottleneck.
- Licenses for other projects incorporated by Bottleneck are now reproduced in full in separate files in the LICENSES/ directory (eg, LICENSES/NUMPY_LICENSE)
- All licenses have been updated. Notably, setuptools is now MIT licensed and no longer under the ambiguous dual PSF/Zope license.
Bottleneck now uses PEP 518 for specifying build dependencies, with per Python version specifications (#247)

Remove numpydoc package from Bottleneck source distribution
bottleneck.slow.reduce.nansum() and bottleneck.slow.reduce.ss() now longer coerce output to have the same dtype as input
Test (tox, travis, appveyor) against latest numpy (in conda)
Performance benchmarking also available via asv
versioneer now used for versioning (#213)
Test suite now uses pytest as nose is deprecated (#222)
python setup.py build_ext --inplace is now incremental (#224)
python setup.py clean now cleans all artifacts (#226)
Compiler feature support now identified by testing rather than hardcoding (#227)
The BN_OPT_3 macro allows selective use of -O3 at the function level (#223)
Contributors are now automatically cited in the release notes (#244)

Speed up bottleneck.reduce.anynan() and bottleneck.reduce.allnan() by 2x via BN_OPT_3 (#223)
All functions covered by asv benchmarks
bottleneck.nonreduce.replace() speedup of 4x via more explicit typing (#239)
bottleneck.reduce.median() up to 2x faster for Fortran-ordered arrays (#248)

Documentation fails to build on Python 3 (#170)
bottleneck.benchmark.bench() crashes on python 3.6.3, numpy 1.13.3 (#175)
bottleneck.nonreduce_axis.push() raises when n=None is explicitly passed (#178)
bottleneck.reduce.nansum() wrong output when a = np.ones((2, 2))[..., np.newaxis] same issue of other reduce functions (#183)
Silenced FutureWarning from NumPy in the slow version of move functions (#194)
Installing bottleneck onto a system that does not already have Numpy (#195)
Memory leaked when input was not a NumPy array (#201)
Tautological comparison in bottleneck.move.move_rank() removed (#207, #212)