- specfile:
* updated requirements versions
* added new requirements: packaging and flox (recommends)
* require python >= 3.8
* update line numbers in patches
* changed README.rst->README.md
- update to version 2022.6.0:
* This release brings a number of bug fixes and improvements, most
notably a major internal refactor of the indexing functionality,
the use of flox in groupby operations, and experimental support
for the new Python Array API standard. It also stops testing
support for the abandoned PyNIO.
* Much effort has been made to preserve backwards compatibility as
part of the indexing refactor. We are aware of one unfixed issue.
* New Features
+ Add Dataset.dtypes, core.coordinates.DatasetCoordinates.dtypes,
core.coordinates.DataArrayCoordinates.dtypes properties: Mapping
from variable names to dtypes. (PR6706) By Michael Niklas.
+ Initial typing support for groupby(), rolling(), rolling_exp(),
coarsen(), weighted(), resample(), (PR6702) By Michael Niklas.
+ Experimental support for wrapping any array type that conforms
to the python array api standard. (PR6804) By Tom White.
* Bug fixes
+ save_mfdataset() now passes **kwargs on to Dataset.to_netcdf(),
allowing the encoding and unlimited_dims options with
save_mfdataset(). (GH6684) By Travis A. O’Brien.
+ Fix backend support of pydap versions <3.3.0 (GH6648,
PR6656). By Hauke Schulz.
+ Dataset.where() with drop=True now behaves correctly with mixed
dimensions. (GH6227, PR6690) By Michael Niklas.
+ Accommodate newly raised OutOfBoundsTimedelta error in the
development version of pandas when decoding times outside the
range that can be represented with nanosecond-precision values
(GH6716, PR6717). By Spencer Clark.
+ open_dataset() with dask and ~ in the path now resolves the home
directory instead of raising an error. (GH6707, PR6710) By
Michael Niklas.
+ DataArrayRolling.__iter__() with center=True now works
correctly. (GH6739, PR6744) By Michael Niklas.
* Internal Changes
+ xarray.core.groupby, xarray.core.rolling,
xarray.core.rolling_exp, xarray.core.weighted and
xarray.core.resample modules are no longer imported by
default. (PR6702)
- changes from version 2022.06.0rc0:
* This pre-release brings a number of bug fixes and improvements,
most notably a major internal refactor of the indexing
functionality and the use of flox in groupby operations. It also
stops testing support for the abandoned PyNIO.
* Known Regressions
+ reset_coords(drop=True) does not create indexes (GH6607)
* New Features
+ The zarr backend is now able to read NCZarr. By Mattia Almansi.
+ Add a weighted quantile method to DatasetWeighted and
DataArrayWeighted (PR6059). By Christian Jauvin and David Huard.
+ Add a create_index=True parameter to Dataset.stack() and
DataArray.stack() so that the creation of multi-indexes is
optional (PR5692). By Benoît Bovy.
+ Multi-index levels are now accessible through their own, regular
coordinates instead of virtual coordinates (PR5692). By Benoît
Bovy.
+ Add a display_values_threshold option to control the total
number of array elements which trigger summarization rather than
full repr in (numpy) array detailed views of the html repr
(PR6400). By Benoît Bovy.
+ Allow passing chunks in kwargs form to Dataset.chunk(),
DataArray.chunk(), and Variable.chunk(). (PR6471) By Tom
Nicholas.
+ Add core.groupby.DatasetGroupBy.cumsum() and
core.groupby.DataArrayGroupBy.cumsum(). By Vladislav Skripniuk
and Deepak Cherian. (PR3147, PR6525, GH3141)
+ Expose inline_array kwarg from dask.array.from_array in
open_dataset(), Dataset.chunk(), DataArray.chunk(), and
Variable.chunk(). (PR6471)
+ Expose the inline_array kwarg from dask.array.from_array() in
open_dataset(), Dataset.chunk(), DataArray.chunk(), and
Variable.chunk(). (PR6471) By Tom Nicholas.
+ polyval() now supports Dataset and DataArray args of any shape,
is faster and requires less memory. (PR6548) By Michael Niklas.
+ Improved overall typing.
+ Dataset.to_dict() and DataArray.to_dict() may now optionally
include encoding attributes. (PR6635) By Joe Hamman.
+ Upload development versions to TestPyPI. By Justus Magin.
* Breaking changes
+ PyNIO support is now untested.
+ The Dataset and DataArray rename" methods do not implicitly add
or drop indexes. (PR5692). By Benoît Bovy.
+ Many arguments like keep_attrs, axis, and skipna are now keyword
only for all reduction operations like .mean. By Deepak Cherian,
Jimmy Westling.
+ Xarray’s ufuncs have been removed, now that they can be replaced
by numpy’s ufuncs in all supported versions of numpy. By
Maximilian Roos.
+ xr.polyval() now uses the coord argument directly instead of its
index coordinate. (PR6548) By Michael Niklas.
* Bug fixes
+ Dataset.to_zarr() now allows to write all attribute types
supported by zarr-python. By Mattia Almansi.
+ Set skipna=None for all quantile methods
(e.g. Dataset.quantile()) and ensure it skips missing values for
float dtypes (consistent with other methods). This should not
change the behavior (PR6303). By Mathias Hauser.
+ Many bugs fixed by the explicit indexes refactor, mainly related
to multi-index (virtual) coordinates. See the corresponding
pull-request on GitHub for more details. (PR5692). By Benoît
Bovy.
+ Fixed “unhashable type” error trying to read NetCDF file with
variable having its ‘units’ attribute not str
(e.g. numpy.ndarray) (GH6368). By Oleh Khoma.
+ Omit warning about specified dask chunks separating chunks on
disk when the underlying array is empty (e.g., because of an
empty dimension) (GH6401). By Joseph K Aicher.
+ Fixed the poor html repr performance on large multi-indexes
(PR6400). By Benoît Bovy.
+ Allow fancy indexing of duck dask arrays along multiple
dimensions. (PR6414) By Justus Magin.
+ In the API for backends, support dimensions that express their
preferred chunk sizes as a tuple of integers. (GH6333, PR6334)
By Stan West.
+ Fix bug in where() when passing non-xarray objects with
keep_attrs=True. (GH6444, PR6461) By Sam Levang.
+ Allow passing both other and drop=True arguments to
DataArray.where() and Dataset.where() (PR6466, PR6467). By
Michael Delgado.
+ Ensure dtype encoding attributes are not added or modified on
variables that contain datetime-like values prior to being
passed to xarray.conventions.decode_cf_variable() (GH6453,
PR6489). By Spencer Clark.
+ Dark themes are now properly detected in Furo-themed Sphinx
documents (GH6500, PR6501). By Kevin Paul.
+ Dataset.isel(), DataArray.isel() with drop=True works as
intended with scalar DataArray indexers. (GH6554, PR6579) By
Michael Niklas.
+ Fixed silent overflow issue when decoding times encoded with
32-bit and below unsigned integer data types (GH6589,
PR6598). By Spencer Clark.
+ Fixed .chunks loading lazy data (GH6538). By Deepak Cherian.
* Documentation
+ Revise the documentation for developers on specifying a
backend’s preferred chunk sizes. In particular, correct the
syntax and replace lists with tuples in the examples. (GH6333,
PR6334) By Stan West.
+ Mention that DataArray.rename() can rename coordinates. (GH5458,
PR6665) By Michael Niklas.
+ Added examples to Dataset.thin() and DataArray.thin() By Emma
Marshall.
* Performance
+ GroupBy binary operations are now vectorized. Previously this
involved looping over all groups. (GH5804, PR6160) By Deepak
Cherian.
+ Substantially improved GroupBy operations using flox. This is
auto-enabled when flox is installed. Use
xr.set_options(use_flox=False) to use the old
algorithm. (GH4473, GH4498, GH659, GH2237, PR271). By Deepak
Cherian, Anderson Banihirwe, Jimmy Westling.
* Internal Changes
+ Many internal changes due to the explicit indexes refactor. See
the corresponding pull-request on GitHub for more
details. (PR5692). By Benoît Bovy.