Path to this page:
Subject: CVS commit: pkgsrc/math/py-xarray
From: Adam Ciarcinski
Date: 2022-02-10 15:11:31
Message id: 20220210141131.89971FB24@cvs.NetBSD.org
Log Message:
py-xarray: updated to 0.21.1
v0.21.1 (31 January 2022)
-------------------------
This is a bugfix release to resolve
Bug fixes
~~~~~~~~~
- Add `packaging` as a dependency to Xarray
v0.21.0 (27 January 2022)
-------------------------
New Features
~~~~~~~~~~~~
- New top-level function :py:func:`cross`.
- ``keep_attrs`` support for :py:func:`where`
- Enable the limit option for dask array in the following methods \
:py:meth:`DataArray.ffill`, :py:meth:`DataArray.bfill`, :py:meth:`Dataset.ffill` \
and :py:meth:`Dataset.bfill`
Breaking changes
~~~~~~~~~~~~~~~~
- Rely on matplotlib's default datetime converters instead of pandas'
- Improve repr readability when there are a large number of dimensions in \
datasets or dataarrays by
wrapping the text once the maximum display width has been exceeded.
Deprecations
~~~~~~~~~~~~
- Removed the lock kwarg from the zarr and pydap backends, completing the \
deprecation cycle started in :issue:`5256`.
- Support for ``python 3.7`` has been dropped.
Bug fixes
~~~~~~~~~
- Preserve chunks when creating a :py:class:`DataArray` from another \
:py:class:`DataArray`
- Properly support :py:meth:`DataArray.ffill`, :py:meth:`DataArray.bfill`, \
:py:meth:`Dataset.ffill` and :py:meth:`Dataset.bfill` along chunked dimensions
- Subclasses of ``byte`` and ``str`` (e.g. ``np.str_`` and ``np.bytes_``) will \
now serialise to disk rather than raising a ``ValueError: unsupported dtype for \
netCDF4 variable: object`` as they did previously
- Fix applying function with non-xarray arguments using :py:func:`xr.map_blocks`.
- No longer raise an error for an all-nan-but-one argument to
:py:meth:`DataArray.interpolate_na` when using `method='nearest'`
- `dt.season \
<https://xarray.pydata.org/en/stable/generated/xarray.DataArray.dt.season.html>`_ \
can now handle NaN and NaT.
- Determination of zarr chunks handles empty lists for encoding chunks or \
variable chunks that occurs in certain cirumstances
Internal Changes
~~~~~~~~~~~~~~~~
- Replace ``distutils.version`` with ``packaging.version``
- Removed internal checks for ``pd.Panel``
- Add ``pyupgrade`` pre-commit hook
Files: