sha,message,author_date,committer_date,raw_author,raw_committer,repo,author,committer aaa778cffb89baaece31882e03a7f4af0adfe798,"call `np.cross` with 3D vectors only (#8993) * convert the expected parameters to arrays * manually zero-pad to avoid calling `np.cross` with 2D vectors",2024-05-03T15:22:25Z,2024-05-03T15:22:25Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 748bb3a328a65416022ec44ced8d461f143081b5,"avoid a couple of warnings in `polyfit` (#8939) * use `numpy.finfo` instead of `numpy.core.finfo` * add `copy` to the signature of `__array__` * don't pass `copy` to `asarray` The `copy` kwarg appears to only exist in `numpy>=2`. * try making the typing the same for all overloads * change the typing for `NamedArray.__array__` * fix the signature of `__array__` * revert the change to the signature of `__array__` * ignore all deprecation warnings about `__array__` gaining a new kwarg I don't know enough about typing to ignore typing issues within a protocol, so that was the only way I could figure out how to get this PR unstuck. Once we know what to do here, we can remove the ignore in a new PR. * add back the `copy` kwarg if the protocol does not have type hints * Update xarray/core/dataset.py --------- Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2024-05-01T15:34:20Z,2024-05-01T15:34:20Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 bfcb0a7a97ee83d6fb93e67ef7a5127b59135464,"clean up the upstream-dev setup script (#8986) * try not to build `numcodecs` in CI [skip-ci] [skip-rtd] * don't remove `numcodecs` from the upstream-dev environment * also don't remove `pint`, it's not in the environment at the moment * [skip-ci][skip-rtd] * don't build `bottleneck` without build isolation `bottleneck` uses the release candidate on PyPI to build `numpy>=2` compatible wheels. * also build `cftime` without build isolation (same as `bottleneck`: it builds against `numpy>=2.0.0rc1`) * remove the build-deps * [skip-ci][skip-rtd] * try removing the upstream-dev build of `pyarrow` [skip-ci][skip-rtd] Supposedly `pyarrow>=16.0` is compatible with `numpy>=2`, but I don't know whether the version on `conda-forge` is built against the release candidate of `numpy`. * Revert ""try removing the upstream-dev build of `pyarrow`"" [skip-ci][skip-rtd] This reverts commit 48a656f0ead2271446c7e56a699782f3d91fe169.",2024-04-30T20:59:56Z,2024-04-30T20:59:56Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 a05fe82a16a29234d9e3c6f1cab550e5e7a34034,"more engine environment tricks in preparation for `numpy>=2` (#8978) * build-deps for numcodecs * temporarily remove `pydap` from the upstream-dev CI * try adding back `h5netcdf` * skip h5netcdf ros3 tests if h5py was compiled without support for ros3 * invert the condition * also invert the other condition * replace `numpy.core.defchararray.add` with `numpy.strings.add` * use `numpy.char.add` instead `numpy.strings` exists since `numpy>=2`",2024-04-29T14:56:21Z,2024-04-29T14:56:21Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 5a35ca42d5372980c6cf5ccb6c6767324dee5ebf,use `nan` instead of `NaN` (#8961),2024-04-21T22:01:03Z,2024-04-21T22:01:03Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 af085283024472d7e5c7f653eb1786791cf6afc7,"adapt more tests to the copy-on-write behavior of pandas (#8940) * mirror the `Dataset` `copy_coords` test * avoid modifying the values of a `pandas` series",2024-04-13T14:44:50Z,2024-04-13T14:44:50Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 aa08785265265c270ff2450812e1cefc1cd2124f,use `pd.to_timedelta` instead of `TimedeltaIndex` (#8938),2024-04-13T12:32:13Z,2024-04-13T12:32:13Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 14fe7e077a59d530f18ea2e4d7c38fb4a62c0b19,"try to get the `upstream-dev` CI to complete again (#8823) * install upstream-dev h5py and enable h5netcdf again * try installing nightly `pyarrow` [skip-ci] * use `micromamba` to speed up the force-removal of packages * add a missing trailing backslash [skip-ci] * revert the enabling of upstream-dev h5netcdf [skip-ci] * also build `numexpr` * replace the removed `numpy.array_api` with `array-api-strict` * don't install upstream-dev `h5py` * remove `numexpr` from the environment instead [skip-ci] * ignore the missing typing in `array_api_strict` * restore the old `numpy.array_api` import for `numpy<2` * ignore the redefinition of `Array`",2024-03-12T16:04:52Z,2024-03-12T16:04:52Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 8e1dfcf571e77701d5669fd31f1576aed1007b45,new whats-new section (#8767),2024-02-20T10:35:34Z,2024-02-20T10:35:34Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 626648a16568a19d4a90e425c851570a7b3ecac0,actually release (#8766),2024-02-18T23:06:21Z,2024-02-18T23:06:21Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 537f4a74ffc58d1dbe65b1e1d32ccce6740fb9b0,"release summary for 2024.02.0 (#8764) * move author names to a separate line * style fixes * first attempt at a summary * add contributors * move entry to the right place * more style * reflow * remove additional indentation * point intersphinx to the new home of cubed * add flox to intersphinx * [skip-ci]",2024-02-18T22:52:14Z,2024-02-18T22:52:14Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 d34ebff531b5c02ead95b9dcd9b1c8fdf101260a,"implement `isnull` using `full_like` instead of `zeros_like` (#7395) * use `full_like` to implement `isnull` on non-nullable dtypes * remove the unused import of `zeros_like` * add tests to make sure non-nullable dtypes still return all-`False` * remove the now unused `zeros_like`",2024-01-23T18:29:13Z,2024-01-23T18:29:13Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 96f9e948d85bb0c1b1a511971606d23e0f7cc122,"import from the new location of `normalize_axis_index` if possible (#8483) * import from the new location of `normalize_axis_index` if possible * disable build isolation for `cftime` [skip-ci] [skip-rtd] * actually apply the `--no-build-isolation` [skip-ci] [skip-rtd] * install `cython`, which is needed to build `cftime` * use `micromamba` instead of `conda` for removing packages * switch the main uninstall back to `conda` * temporarily remove `cf_units`, `h5py`, and `netcdf4` * also remove `hdf5` [skip-ci] * try building `numcodecs` from github and without build isolation * build `cftime` separately from `numcodecs` and install more deps [skip-ci] * don't uninstall packages that are already removed [skip-ci] * also build `bottleneck` without build isolation * use `module_available` instead of eagerly importing * add a `minversion` kwarg to `module_available` * use `module_available` in `nputils` as well * more type ignores * move the type ignores to the import * ignore complaints about unused ignores (the point here is compatibility between multiple versions) * a couple of other `type: ignore` comments (only relevant for nightly)",2024-01-18T15:34:56Z,2024-01-18T15:34:56Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 e91ee89edbdec4299897e60c9d97d0cf915353c9,"run CI on `python=3.12` (#8605) * enable a special 3.12 ci job * add the special environment files * update the trove classifiers * remove `pint` from the 3.12 environment The reason is that `pint` currently does not explicitly define `dtype`, causing our runtime checkable protocols not to match.",2024-01-17T21:54:12Z,2024-01-17T21:54:12Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 c47f35db4e73507b4ac729ec23b0faf8ceb5ace1,"ignore a `DeprecationWarning` emitted by `seaborn` (#8576) * move the ignore from the cli to `xarray.tests` * ignore `seaborn` using deprecated API",2023-12-30T22:10:07Z,2023-12-30T22:10:07Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 7c1bb8dffcfc2fba08b470676884200f035b5051,"explicitly skip using `__array_namespace__` for `numpy.ndarray` (#8526) * explicitly skip using `__array_namespace__` for `numpy.ndarray` * actually use the array in question --------- Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2023-12-06T17:58:46Z,2023-12-06T17:58:46Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 b6eaf436f7b120f5b6b3934892061af1e9ad89fe,"migrate the other CI to python 3.11 (#8416) * migrate the additional CI to py311 * migrate the upstream-dev CI to python 3.11 * switch to the default environment file * convert a left-over use of `provision-with-micromamba` * silence the `cgi` deprecation warning from `pydap` --------- Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2023-11-17T15:27:21Z,2023-11-17T15:27:21Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 28053ba09cb8262f4ba3e73fb884ea4f236c5704,"remove `cdms2` (#8441) * remove `cdms2` from all environments * remove any code related to `cdms2` * remove the `mypy` override for `cdms2` * remove the special environment files for python 3.11 * whats-new * more info on the replacement * remove the `cdms2` methods from the api docs * remove all mentions of UV-CDAT / cdms2 from the faq * also change the binder environment",2023-11-14T17:15:49Z,2023-11-14T17:15:49Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 bac90ab067d7437de875ec57d90d863169e70429,"adapt to NEP 51 (#8064) * convert string and bytes items to standard python types * [test-upstream] * modify the expected error message",2023-09-25T04:46:48Z,2023-09-25T04:46:48Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 e00e31b9ea7fcd562f75f6c657d02c4f81867095,"fix the failing docs (#8188) * don't try to build with `pandas=2.1.0` one of the docstrings we inherit from `pandas` is malformatted in that version * reorder the project metadata sections * source the exhaustive list of optional deps from `pyproject.toml` * link to the new location of the colormap tutorial in matplotlib * also convert the narrative documentation to `pyproject.toml`",2023-09-15T13:26:24Z,2023-09-15T13:26:24Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 2951ce0215f14a8a79ecd0b5fc73a02a34b9b86b,"fix miscellaneous `numpy=2.0` errors (#8117) * replace `np.unicode_` with `np.str_` * replace `np.NaN` with `np.nan` * replace more instances of `np.unicode_` note that with more modern versions of `numpy` the `.astype(np.str_)` don't actually change the dtype, so maybe we can remove those. * more instances of renamed / removed dtypes * more dtype replacements * use `str.encode(encoding)` instead of `bytes(str, encoding)` * explicitly import `RankWarning` * left-over `np.RankWarning` * use `float` instead of the removed `np.float_` * ignore missing stubs for `numpy.exceptions` --------- Co-authored-by: Kai Mühlbauer <kmuehlbauer@wradlib.org> Co-authored-by: Mathias Hauser <mathause@users.noreply.github.com> Co-authored-by: Maximilian Roos <5635139+max-sixty@users.noreply.github.com> Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2023-09-11T03:55:51Z,2023-09-11T03:55:51Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 ab096b05821a039f7a5b416e83227462625687ff,"unpin `numpy` (#8061) * unpin `numpy` * dispatch to `np.empty_like` if there's no missing value * make sure there's no code path without `create_template` * declare `create_template` as callable * ignore the intentionally wrong typing * mark the ignore as intentional, even if it might not be used this important because the fix in `numpy` that now means we *don't* need it anymore has been around for less than 3 months (requires a sufficiently new version of `mypy`). * also directly check that `Variable.unstack` does not raise warnings * fix the unstack test",2023-08-17T18:14:21Z,2023-08-17T18:14:21Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 6b1ff6d13bf360df786500dfa7d62556d23e6df9,"join together duplicate entries in the text `repr` (#7225) * group the variable names of common indexes * implement option 1: separate with just newlines * implement option 2: list-like presentation of the indexes * implement option 3: light box components in front of the coord names * implement option 4: light box components after the coord names * fix the repr of `Indexes` * only add combine markers for xarray indexes * select option 3 Support for the box components should be common enough that we can use it (and the worst that can happen is that those characters are displayed with a replacement glyph – usually a diamond with a question mark) * improve typing Co-authored-by: Illviljan <14371165+Illviljan@users.noreply.github.com> * whats-new.rst * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * simplify the visual hint construction * improve the existing index repr test * check that the hints are constructed correctly * convert the names to tuples * revert the typing of `index` * more conversion to tuple * type hint the instance variable * whats-new entry * don't type-check the monkeypatching * adjust the typing, again * use a subclass instead of monkeypatching and type-ignores --------- Co-authored-by: Illviljan <14371165+Illviljan@users.noreply.github.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2023-07-20T21:13:56Z,2023-07-20T21:13:56Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 4a79b32d80bb6e2200754c4b65f5d17530b2ab4c,"manually unshallow the repository on RTD (#7961) RTD is deprecating the feature flag we made use of before.",2023-07-05T15:44:09Z,2023-07-05T15:44:09Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 073a57f372496533966645691712e988605f1b13,"use trusted publishers instead of a API token (#7899) * use trusted publishers instead of a API token * same for TestPyPI",2023-06-16T08:37:07Z,2023-06-16T08:37:07Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 f4e0523be0ce8babaa8eff38365e5308b1fdb76b,"don't use `CacheFileManager.__del__` on interpreter shutdown (#7880) * store a reference to the acquire function on the object This keeps the function from being garbage collected before we can call it. * whats-new.rst * make `lock_acquire` a static method * fix rst link * comment on the existence of the static method `lock_acquire` * also store a reference to `OPTIONS` * revert storing additional references * remove `CachingFileManager.__del__` on interpreter shutdown * rephrase the changelog [skip-ci] * [skip-ci] --------- Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2023-06-06T15:14:36Z,2023-06-06T15:14:36Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 46ef0ca86c8a8cd9e28a1a38c45e2ebb7d022ae0,retire the TestPyPI workflow (#7889),2023-06-04T18:46:13Z,2023-06-04T18:46:13Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 960f15ce35826b85745adf5a7b5d8f7decbfa16d,"move to `setup-micromamba` (#7878) * transition to setup-micromamba * transition the last few ci to setup-micromamba * remove the obsolete and unused `parse_logs` script * don't specify `environment-file: false`",2023-06-01T16:21:55Z,2023-06-01T16:21:55Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 d4edd70190f3e57a39e2cf330a411b5a17543ee7,"deprecate the `cdms2` conversion methods (#7876) * emit deprecation warnings for the cdms2 conversion methods * add the deprecation notice to the docstring * check that the deprecation warnings are properly raised * actually check that the deprecation message mentions cdms2 * skip cdms2 checks for new numpy / python versions * whats-new entry * make the deprecation message consistent * functions → methods * adjust the skip condition * maintenance mode → deprecated",2023-05-29T19:01:20Z,2023-05-29T19:01:20Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 b319d8671b178407927b5878ed878ffab22b6352,"add `numba` to the py3.11 environment (#7867) * add `numba` back to the py311 environment * try cdms2 * remove `cdms2`, it's not available for `python=3.11`",2023-05-28T06:30:09Z,2023-05-28T06:30:09Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 fcb81756a1d81f334e84f35e4b6cb59a448e367e,"defer to `numpy` for the expected result (#7875) * simplify the test for unary funcs * use `full_like` to construct the expected result * Revert ""use `full_like` to construct the expected result"" This reverts commit a14aeb51f6787a93ce12958bc3e9634fb0e84cdb.",2023-05-27T19:53:06Z,2023-05-27T19:53:06Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 fd0ac3e09a09567a4b584f1429219ff72b5fa4dd,"adapt the `pint` + `dask` test to the newest version of `pint` (#7855) * use the registry to create the quantities * skip the `pint` + `dask` test before `pint=0.21` * also parse the static version",2023-05-25T17:24:33Z,2023-05-25T17:24:33Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 6d17fa0bea187ccba7247dd2bd9b6179636017b0,"adjust the deprecation policy for python (#7793) * extend the policy interval for python to 30 months * extend the hard-coded list of python release dates * changelog",2023-05-01T22:26:55Z,2023-05-01T22:26:55Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 6d98dbd644ac264c9c67db78d9a5ae011019b0ec,remove the `black` hook (#7756),2023-04-14T16:36:18Z,2023-04-14T16:36:18Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 388387b834a9763d24c5d028ec7727a4a56a71f3,"reword the what's new entry for the `pandas` 2.0 dtype changes (#7755) * move and rewrite the what's new entry for the pandas 2.0 dtype PR * [skip-ci]",2023-04-14T14:30:49Z,2023-04-14T14:30:49Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 db2d414c6af236be2df4f5ca353e154b2f4cf885,"`pandas=2.0` support (#7724) * unpin pandas in the package metadata * unpin pandas in the ci environments [skip-rtd] * create the input arrays in the parametrization * split `test_sel_float` into variants * skip the float16 variant if pandas>=2.0 is installed * [skip-rtd] * add tests for `days_in_month` and its alias * make sure the name and dtype match the expected * actually verify that the dtype stays the same * apply the dtype for non-dask * always use `int32` to follow `pandas=2.0` * back to `int64` * same for the test * final undo of `int64` → `int32` * update the comment to make more sense * simplify the conversion of the expected data * change back to the old condition * cast float16 to float64 when creating indexes (but warn anyways) * convert float16 to float64 when selecting using arrays (all except 0d) * move the float16 variant to a separate test this allows us to check for expected warnings * explicitly type the kwargs as a mapping of str → str * reword the warning message * restore the pin * [skip-ci] [skip-rtd] * rerun to make sure we don't introduce failures with `pandas<2` * changelog",2023-04-12T13:04:11Z,2023-04-12T13:04:11Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 a61d73b39c5e62631378348e49fc6febe2595622,"use the `files` interface instead of the deprecated `read_binary` (#7664) * use the `files` interface instead of the deprecated `read_binary` * use `read_text(encoding=...)` instead of `read_bytes(...).decode(...)` Co-authored-by: Michael Niklas <mick.niklas@gmail.com> --------- Co-authored-by: Illviljan <14371165+Illviljan@users.noreply.github.com> Co-authored-by: Michael Niklas <mick.niklas@gmail.com>",2023-03-30T14:58:42Z,2023-03-30T14:58:42Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 635b9d0f2db4cc3ff005983c4af46ef6801258e2,"add timeouts for tests (#7657) * install `pytest-timeout` into all environments * don't also install it into the upstream-dev environment * set a timeout of 5 minutes This is plenty for a test to run: we shouldn't have tests that take more than 30 seconds, but the windows / macos runners tend to be slower than the linux ones. * add `pytest-timeout` to the remaining test environments * add `pytest-timeout` to the ignored dependencies [skip-rtd] * reduce the timeout from 5 minutes to 90 seconds * double the test timeout",2023-03-24T15:49:22Z,2023-03-24T15:49:22Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 821dc24b5f3ed91b843a634bf8513a26046269ef,ignore the `pkg_resources` deprecation warning (#7594),2023-03-08T13:41:53Z,2023-03-08T13:41:53Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 43ba095712de12c957e0a4acf956df01d84b2046,"update the docs environment (#7442) * use python 3.10 in the docs environment * require a more recent `sphinx` version * [skip-ci] * remove the unused pydata-sphinx-theme [skip-ci] * pin `sphinx-book-theme` to a more recent version [skip-ci] * try working around the theme bug [skip-ci] * use string interpolation markers for the extlink text * try specifying the project of a intersphinx link * remove the empty `extra_navbar` and `navbar_footer_text` theme options * [skip-ci] --------- Co-authored-by: Joe Hamman <joe@earthmover.io> Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com> Co-authored-by: Illviljan <14371165+Illviljan@users.noreply.github.com>",2023-03-03T10:14:13Z,2023-03-03T10:14:13Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 f128f248f87fe0442c9b213c2772ea90f91d168b,"RTD maintenance (#7477) * upgrade the os * get the correct version in builds Even if we don't display it in the built documentation anymore, some functionality like `pandas.DataFrame.to_xarray` have a version check. * print the last 5 commits [skip-rtd] [skip-ci] * remove the ""unshallow"" apparently, it does not make sense for complete repos, so git fails. * also remove the history print The idea was to check whether there was a special commit on top of the most recent commit of the PR (this seems to be the case on GH actions). There's none, so we don't need the print. * [skip-ci] [skip-rtd] * [skip-ci] [skip-rtd]",2023-01-25T14:58:46Z,2023-01-25T14:58:46Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 c3032cdcf732d3387778d8718476cf7b68ef0c5c,"fix the RTD build skipping feature (#7476) * treat the pattern as a fixed string * skip-rtd [skip-ci] Intentionally misspell to make sure the skip is *not* triggered. * [skip-rtd] [skip-ci]",2023-01-25T11:17:56Z,2023-01-25T11:17:56Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 9b3014649380b537c35054732eaf9aabfc80e9c2,"allow skipping RTD builds (#7470) * configure rtd to skip builds if `[skip-rtd]` is found in the subject * properly cancel the build early * try if it actually works [skip-rtd]",2023-01-24T16:09:50Z,2023-01-24T16:09:50Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 8be93496a4ef8d09889b835ea57c94d02f933f00,"create separate environment files for `python=3.11` (#7469) * add new environment files for python 3.11 These should be deleted after all the remaining dependencies (cdms2, numba, numbagg, sparse) support python 3.11. * enable python 3.11 * use the special environment files for python 3.11 * officially support `python=3.11`",2023-01-23T13:03:12Z,2023-01-23T13:03:12Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 19838dce11faeea8dc629cea1607a646ee7be86e,"install `numbagg` from `conda-forge` (#7415) * install `numbagg` from conda-forge * temporarily pin `numpy<1.24` to get CI to pass Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2023-01-20T19:46:42Z,2023-01-20T19:46:42Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 b93dae4079daf0fc4c042fd0d699c16624430cdc,"remove `numbagg` and `numba` from the upstream-dev CI (#7416) * temporarily remove `numba` from the upstream-dev environment * don't explicitly install the matplotlib dependency * use `pip` to remove `numbagg` * [test-upstream] * remove `numba` and `numbagg` first [test-upstream] Co-authored-by: Joe Hamman <joe@earthmover.io>",2023-01-04T20:07:28Z,2023-01-04T20:07:28Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 db68db6793bdd10f740e1ff7f68d821e853e3d73,"`keep_attrs` for pad (#7267) * make use of `keep_attrs` in `Variable.pad` * propagate `keep_attrs` * resolve the default `keep_attrs` in `Dataset.pad` * add tests for `Variable.pad` * add tests for `Dataset.pad` and keep the `Dataset` attrs * actually pad the variable * also check that untouched variables are not modified * add tests for `DataArray.pad` * update whats-new.rst * Set True by default * move the whats-new.rst entry to the dev section Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com> Co-authored-by: Illviljan <14371165+Illviljan@users.noreply.github.com>",2022-12-12T15:59:41Z,2022-12-12T15:59:41Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 cc7e09a3507fa342b3790b5c109e700fa12f0b17,pin mypy to a known good version (#7269),2022-11-08T16:49:12Z,2022-11-08T16:49:12Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 4b65245b79b3549b057824fd2aa3228aa0a47226,"use the moving release tag of `issue-from-pytest-log` (#7260) * use the moving release tag * [skip-ci] [test-upstream]",2022-11-07T14:32:02Z,2022-11-07T14:32:02Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 93d1eabca4392eeab6a7cbe34e583a7195520dc8,actually test different python versions (#7242),2022-10-31T13:16:45Z,2022-10-31T13:16:45Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 ca57e5cd984e626487636628b1d34dca85cc2e7c,"use the new action to create an issue from the output of reportlog (#7212) * use the new action to create an issue from the output of reportlog * move report generation from a separate job to a step * [test-upstream] * remove the output declaration [skip-ci][test-upstream] * specify a version for the action [skip-ci][test-upstream] * consistent step name spelling [skip-ci]",2022-10-26T09:12:42Z,2022-10-26T09:12:42Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 519abb7bde020f2b27cb7e3dfddec8c6eecb7722,"use the moving release tag of ci-trigger (#7213) * use the moving release tag of ci-trigger * [test-upstream]",2022-10-25T14:29:50Z,2022-10-25T14:29:50Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 8f34b32cd77c7e7e2e559c1ced9adf3100ecf40d,"indexes section in the HTML repr (#7185) * borrow `Index._repr_inline_` and `inline_index_repr` from a different PR * implement a crude HTML repr for indexes * css fixes and tweaks * escape inline repr * show one section item per unique index Show the names of all corresponding coordinates in the same item (separated by line return). * also add the indexes section to the DataArray html repr * make the grouper a bit easier to understand * add documentation for the display_expand_indexes option * collapse the indexes section by default * update the HTML formatting tests * Update xarray/core/options.py Co-authored-by: Benoit Bovy <bbovy@gfz-potsdam.de> * whats-new.rst * improve the index preview contrast * add the repr_inline PR to the changelog [skip-ci] Co-authored-by: Benoit Bovy <benbovy@gmail.com> Co-authored-by: Benoit Bovy <bbovy@gfz-potsdam.de>",2022-10-19T21:12:45Z,2022-10-19T21:12:45Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 7379923de756a2bcc59044d548f8ab7a68b91d4e,"use `_repr_inline_` for indexes that define it (#7183) * use the inline repr, if available * remove the unused max_width argument * also remove the unused is_index parameter * add tests for the index summarizer * add a basic _repr_inline_ implementation * don't delegate to __repr__ since _repr_inline_ is now guaranteed to exist * update the test to check the default implementation * don't strip the `Index` suffix * undo the code simplification: pandas indexes use the same code path and the raw pandas indexes don't define `_repr_inline_`. * move the inline index repr formatting to a separate function * properly use the max_width setting * fix the definition of the `_repr_inline_` method * disable `mypy` for the index formatting test It does not allow reassigning the method, and it also complains about an attribute created in the constructor. * styling [skip-ci] * trigger CI * add a comment explaining why we need the `repr` fallback",2022-10-19T14:06:47Z,2022-10-19T14:06:47Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 b80d97381e224b7f8cc83e4082d3b5b659c30617,"display the indexes in the string reprs (#6795) * add indexes to the string repr * switch indexes_repr to use _mapping_repr * only print the indexes if there are any * make use of col_width and max_rows * display the wrapped index to get rid of the test collection errors * skip default indexes by default * update the doctest for indexes Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>",2022-10-12T16:52:53Z,2022-10-12T16:52:53Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 d3c15fe34ee32934ca41d6e00138bfea3879e787,"use a hook to synchronize the versions of `black` (#7153) * use the synchronization hook * [skip-ci]",2022-10-12T08:00:06Z,2022-10-12T08:00:06Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 8417f495e6b81a60833f86a978e5a8080a619aa0,"rely on `numpy`'s version of `nanprod` and `nansum` (#6873) Co-authored-by: dcherian <deepak@cherian.net>",2022-08-09T14:55:20Z,2022-08-09T14:55:20Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 0c8a78b3080884d985eed74789ee5da62ed043e0,skip creating a cupy-backed IndexVariable (#6872),2022-08-03T15:04:55Z,2022-08-03T15:04:55Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 cc183652bf6e1273e985e1c4b3cba79c896c1193,"resolve the timeouts on RTD (#6727) * explicitly set the dataset for demonstration with `to_dict` * show the dataset for reference",2022-06-25T11:00:50Z,2022-06-25T11:00:50Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 a1b052360b3fae74ebe19961ed83720b0434dd00,"try to import `UndefinedVariableError` from the new location (#6701) * try to import `UndefinedVariableError` from the new location * [test-upstream] * mention when to clean up [test-upstream]",2022-06-22T10:33:24Z,2022-06-22T10:33:24Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 d7931f9014a26e712ff5f30c4082cf0261f045d3,"use `pytest-reportlog` to generate upstream-dev CI failure reports (#6699) * rely on pytest-reportlog to produce the log file * use the new script to generate the report * install the dependencies of the parser script * temporarily enable report generation [test-upstream] * actually upload the log file [skip-ci][test-upstream] * print the created report [skip-ci][test-upstream] * Undo the temporary changes [test-upstream]",2022-06-16T08:11:21Z,2022-06-16T08:11:21Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 a7799fb38a6fec0252d7d4e2d0243d647e70c454,"use micromamba instead of mamba (#6674) * use micromamba instead of mamba * specify the version of the micromamba GA * use a commit hash instead, cache-env has never been released yet * properly access the environment variable * use micromamba in the additional CI * uses micromamba in the upstream-dev CI [test-upstream] * synchronise the minimum tested pandas version with setup.cfg * fix the workflow * try if removing the environment cache works [skip-ci] * synchronise the minimum numpy version * don't cache the upstream-dev environment [test-upstream] * specify the environment name and don't use a file [skip-ci] * trigger ci * use conda-forge * set manual cache keys * fix the mypy job's caching * fix the date call * assign a name to the extra step to get the date * rerun all CI [test-upstream] * include the python version in the cache key * rerun CI",2022-06-10T11:32:59Z,2022-06-10T11:32:59Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 27d6266f0c6bd5b810171812d7df67dabf6ce97e,"install the development version of `matplotlib` into the upstream-dev CI (#6675) * install matplotlib from pypi.anaconda.org/scipy-wheels-nightly/ * [skip-ci][test-upstream] * install pytest-timeout at the end of the file [test-upstream] * install matplotlib before pandas [test-upstream] * temporarily install contourpy This will not be necessary anymore once `matplotlib>=3.6.0` has been released. * [skip-ci][test-upstream]",2022-06-10T11:25:30Z,2022-06-10T11:25:30Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447 b6c5be2e35d20c95d7ba323abcdd272850049a3b,"release notes for the pre-release (#6676) * add a section for the release candidate * add the known regressions * add a small summary of the most notable changes. * pynio * rephrase [skip-ci] * whats-new.rst fixes [skip-ci] * rephrase and remove the mention of `numpy-groupies`",2022-06-09T14:41:31Z,2022-06-09T14:41:31Z,f48b8f0d94dbaa4588e3a965a600d38e311eab17,cd792325681cbad9f663f2879d8b69f1edbb678f,13221727,14808389,19864447