summaryrefslogtreecommitdiff
path: root/tools
Commit message (Collapse)AuthorAgeFilesLines
* MAINT: move the option to build_src, simplifies logicmattip2019-09-201-2/+2
|
* BUILD: mostly restore verbose output on CI (except for pip install)mattip2019-09-202-3/+3
|
* Merge pull request #14525 from mattip/pypySebastian Berg2019-09-191-1/+1
|\ | | | | BUILD, MAINT: run tests with verbose for PyPY, also do not leak resource
| * BUILD, MAINT: run tests with verbose for PyPY, also do not leak resourcemattip2019-09-161-1/+1
| |
* | DOC,MAINT: Build release notes during CircleCI stepSebastian Berg2019-09-151-0/+16
|/ | | | | For the moment, this requires te use for the development version of towncrier.
* Merge pull request #14453 from rgommers/no-cython-in-sdistCharles Harris2019-09-081-0/+1
|\ | | | | BLD: remove generated Cython files from sdist
| * BLD: remove generated Cython files from sdistRalf Gommers2019-09-081-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Rationale: it has happened multiple times in the recent past that users needed to use a more recent Cython than was used to generate the sdist of the most recent release available on PyPI (e.g. for building with a not-yet-released version of Python for which a Cython fix just landed). The dependency specification and packaging tools have matured to the point where it should be fine to require Cython to be installed. Also, we have wheels (and conda-forge packages) for Windows, Linux and macOS, so this change won't be visible at all to the average user. Closes gh-13790 Also adds pyproject.toml to MANIFEST.in, it was missing.
* | MAINT: Only allow using Cython module when cythonizing.Charles Harris2019-09-021-7/+4
|/ | | | | | | | | The command line version of `cython` may point to a different installation of Cython than that installed in the Python running the cythonize script. Because the Cython version can be critical, requiring that the cython used comes from a known place makes it less likely that the wrong version will be used. This treats Cython as a build dependency rather than a free standing compiler.
* BLD: Update RELEASE_WALKTHROUGH and cythonize.Charles Harris2019-09-011-9/+6
| | | | | | Make it less likely that the wrong cython version is used for building or cythonizing NumPy. This is motivated by the upcoming Python 3.8 release that requires the cython 0.29.13 as of 3.8.0b4.
* Merge pull request #14378 from mattip/test-dependenciesCharles Harris2019-09-014-25/+6
|\ | | | | BUILD: move all test dependencies to ./test_requirements.txt
| * BUILD: move all test dependencies to ./test_requirements.txtmattip2019-08-274-25/+6
| |
* | Use *sys.version_info[:2]Hugo2019-08-261-1/+1
| |
* | BUG: Fix for Python 4Hugo2019-08-261-1/+1
| |
* | BUG: Fix for Python 3.10Hugo2019-08-261-1/+1
|/
* TST, MAINT: bump to OpenBLAS 0.3.7 stableTyler Reddy2019-08-211-4/+4
| | | | | | | | | | | | * for those cases where our CI uses an OpenBLAS backend for testing it has now been updated to use v0.3.7 stable release instead of a commit hash on the development branch preceding that release * the pre-built OpenBLAS binaries are pulled in from either the MacPython ecosystem (most architectures) or from a custom build script for ARMv8 and ppc64le architectures used on gcc compile farm machines (documented in openblas_support module)
* TST: Disable `--durations` flag for pytest.Charles Harris2019-08-182-2/+4
| | | | | This is (hopefully) temporary. Current versions of pytest abort when this flag is passed and the tests run with python3.6-dbg.
* TST: Pin pytest and use python3.6-dbg.Charles Harris2019-08-171-1/+1
| | | | | | | Pytest-5.1.0 segfaults on python3.5-dbg. Because we will be dropping 3.5 for NumPy 1.18 we might as well use Python3.6 for the debug testing. That cures the previous segfault, but exposes another so we also pin pytest to 5.0.1.
* TST: Don't test with nose for Python 3.5+.Charles Harris2019-07-281-1/+1
| | | | | Nose imports imp, which has a PendingDeprecationWarning since Python 3.4.
* MAINT: bump ARMv8 / POWER8 OpenBLAS in CITyler Reddy2019-07-201-3/+3
| | | | | * require OpenBLAS 0.3.7.dev in ARMv8 and POWER8 CI runs to match ecosystem / wheels
* Merge pull request #14002 from tylerjereddy/expand_openblas_ver_checkMatti Picus2019-07-164-7/+36
|\ | | | | TST, MAINT, BUG: expand OpenBLAS version checking
| * TST, MAINT: expand OpenBLAS version checkingTyler Reddy2019-07-144-7/+36
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * add a function to tools/openblas_support.py that may be used to test OpenBLAS version available through NumPy in a manner agnostic to any specific CI service * expand OpenBLAS version checking to include ppc64le and POWER8 CI runs; a separate PR will be provided to bump these versions to 0.3.7.dev * OpenBLAS install was broken on ppc64le Travis CI after migration to tools/openblas_support.py; fixed that * Azure CI config no longer uses a single-line 280-character Python "program" assigned to an environment variable to check OpenBLAS version; instead, it now leverages tools/openblas_support.py to perform the same operation using a conventional Python function
* | DEV: Fix spelling error in npy_tempita kwargkai-striega2019-07-151-28/+28
|/ | | | | | | | | | | | | | | | | | | | | | | This commit fixes the spelling of the ``delimeter`` kwarg to ``delimiter`` (note the second e -> i) in the npy_tempita module. This npy_tempita vendorises Tempita a minimal templating language. The spelling has been fixed in the official documentation [1] and source code [2]. This commit fixes that fix by aligning npy_tempita with the reference documentation. Although the vendored version is used in NumPy the delimeter kwarg is not. This raised the consideration of reverting to either the original version or Cython's non-vendored version [3]. Both were decided against because: 1. The original is not actively maintained and required some fixes. 2. Cython's Tempita is an internal Cython tool and Cython may not guarantee that it will be present. [1] https://pyrocore.readthedocs.io/en/latest/tempita.html [2] https://bitbucket.org/ianb/tempita/src/default/tempita/__init__.py [3] https://github.com/cython/cython/blob/master/Cython/Tempita/_tempita.py
* MAINT, TST: remove test-installed-numpy.pyTyler Reddy2019-07-132-68/+2
| | | | | | * replace usage of test-installed-numpy.py with runtests.py; they serve similar purposes and the latter is more commonly used / maintained
* ENH, BUILD: refactor all OpenBLAS downloads into a single, testable filemattip2019-07-093-19/+178
|
* Merge pull request #13772 from mattip/windllCharles Harris2019-06-261-0/+42
|\ | | | | BUILD: use numpy-wheels/openblas_support.py to create _distributor_init.py
| * MAINT: add script to local repomattip2019-06-201-0/+42
| |
* | DOC: Change random to uninitialized and unpredictable in empty and ↵Kexuan Sun2019-06-151-1/+2
|/ | | | | | empty_like (#13771) * DOC: Change "random" to "uninitialized" in doctests
* CI: run asv check on TravisPauli Virtanen2019-05-301-0/+1
|
* Merge pull request #13163 from mattip/randomgenSebastian Berg2019-05-281-33/+52
|\ | | | | | | | | ENH: randomgen This merges randomgen into numpy, which was originally developed at https://github.com/bashtage/randomgen and provides a new and improved API for random number generation with much new and improved functionality.
| * BUILD: move files out of _randomgenmattip2019-05-201-33/+52
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | first cut at building randomgen upgrade 'cythonize' and fix absolute imports to relative define NPY_NO_DEPRECATED_API and fix other warnings enable pgc64 by always using PCG_EMULATED_MATH refactor so import randomgen works add TODO comments for pcg64 improvements fix imports, module name in setup.py; remove _testing make cythonize non-recursive, restore examples to proper place update to randomgen 7bca296c0b9 replace mtrand with LegacyGenerator, tweak for compatibility port f879ef4 to fix GH10839 minimized difference between generator.pyx and _legacy.pyx fix namespace in doctests, mark results that are random update to randomgen commit 95c8cdd1c Incorporate testing of edge cases into main tests Rename test files to describe their purpose Import import locations to reflect numpy paths Correct tolerance on float32 tests Remove set_printoptions Remove complex normal Remove future imports Pull in BasicRNG source changes from original author Small doc fixes _mtrand => _rand Improve consistency of nan handling Prevent nans prducing values from int functions add randomgen documentation to the tree
* | TST: bump / verify OpenBLAS in CITyler Reddy2019-05-211-9/+1
|/ | | | | | | | | * bump OpenBLAS version to 0.3.7.dev in Azure CI to pick up AVX512 SkylakeX fixes and match wheels * use ctypes code to verify appropriate version of OpenBLAS
* TEST: add duration report to testsmattip2019-05-062-2/+6
|
* BLD, TST: implicit func errorsTyler Reddy2019-04-301-0/+1
| | | | | | | | * Travis CI will now fail NumPy builds at the compilation stage when implicit-function-declaration is detected; otherwise, C functions without included prototypes may be allowed to pass through with a warning
* Fix deprecated ``assertEquals()``Nicola Soranzo2019-04-236-30/+30
| | | | See https://docs.python.org/2/library/unittest.html#deprecated-aliases
* MAINT: use openblas, gfortran, tweak azure and comment mro (from review)mattip2019-04-211-4/+48
|
* MAINT: move gc.collect to function (from review)mattip2019-04-191-0/+4
|
* TEST: update for PyPymattip2019-04-172-10/+11
|
* DEV: add pypy3 to azure CImattip2019-04-161-0/+10
|
* TST: use POWER8 OpenBLAS for CITyler Reddy2019-04-051-0/+11
|
* DOC: fixes from reviewmattip2019-02-281-1/+1
|
* MAINT: Convert property to @propertyAlex Watt2019-02-241-10/+10
|
* ENH: __array_function__ updates for NumPy 1.17.0Stephan Hoyer2019-01-221-1/+0
| | | | | | | | | | - Always enable __array_function__ overrides. - Remove special cases for Python 2 compatibility. - Document these changes in 1.17.0-notes.rst. It will be good to see ASV numbers to understand the performance implications of these changes. If need be, we can speed up NumPy functions internally by using non-dispatched functions (with ``.__wrapped__``).
* TEST: activate doctests on doc.structured_arrays, fix for pure doc modulemattip2019-01-141-1/+5
|
* ENH: add "max difference" messages to np.testing.assert_array_equal (#12591)Stephan Hoyer2019-01-011-0/+1
| | | | | | | | | | | | | | | | | | | | | | Example behavior: >>> x = np.array([1, 2, 3]) >>> y = np.array([1, 2, 3.0001]) >>> np.testing.assert_allclose(x, y) AssertionError: Not equal to tolerance rtol=1e-07, atol=0 Mismatch: 33.3% Max absolute difference: 0.0001 Max relative difference: 3.33322223e-05 x: array([1, 2, 3]) y: array([1. , 2. , 3.0001]) Motivation: when writing numerical algorithms, I frequently find myself experimenting to pick the right value of `atol` and `rtol` for `np.testing.assert_allclose()`. If I make the tolerance too generous, I risk missing regressions in accuracy, so I usually try to pick the smallest values for which tests pass. This change immediately reveals appropriate values to use for these parameters, so I don't need to guess and check.
* TST: Update `travis-test.sh` for C99Charles Harris2018-12-251-2/+1
| | | | | | | Most of this was already done, but we were still raising an error for declaration after a statement because the Windows Python 2.7 compiler did not allow it. We can fix this now as NumPy >= 1.17 has dropped Python 2.7 support.
* Merge pull request #12589 from charris/fix-changelog-py3kCharles Harris2018-12-211-2/+4
|\ | | | | MAINT: Update changelog.py for Python 3.
| * MAINT: Update changelog.py for Python 3.Charles Harris2018-12-191-2/+4
| | | | | | | | | | | | | | | | - Don't use UTF8Writer in Python 3. - Don't use print() for blank lines. This is left compatible with Python 2 as someone may use it with that version and there is no need to force the change.
* | ENH: refactor __array_function__ pure Python implementationStephan Hoyer2018-12-191-1/+1
|/
* MAINT: include recfunctions moduleTyler Reddy2018-12-141-0/+1
| | | | | | | | | * added lib.recfunctions to refguide PUBLIC_SUBMODULES, as the doctests were otherwise not getting executed * fixed a failing doctest in recfunctions after above activation
* TST, DOC: enable refguide_checkTyler Reddy2018-12-141-0/+951
| | | | | | | | * ported the refguide_check module from SciPy for usage in NumPy docstring execution/ verification; added the refguide_check run to Azure Mac OS CI * adjusted NumPy docstrings such that refguide_check passes