v0.8

This release of xESMF improves support for parallelization with dask: weights can now be computed in parallel, and those weights can be applied over chunks spanning the horizontal grid dimensions. Previously, computing weights in parallel was only possible using MPI, and datasets could only be chunked over non-spatial dimensions.

These new features are the outcome of Charles Gauthier' internship at Ouranos during the summer of 2023. Thanks to Charles for his hard work and sharp analysis, which led to a permanent position at Ouranos!

What's Changed

• Remove uppercase in longitude/latitude for test by @raphaeldussin in #259
• Fix broken link by @rcaneill in #255
• Perform ci tests with python 3.11 now that numba is compatible. by @charlesgauthier-udm in #272
• Bump mamba-org/provision-with-micromamba from 15 to 16 by @dependabot in #266
• Removed 3.7 from supported versions. Build docs using 3.9 by @huard in #271
• Added w property to Regridder and SpatialAverager by @huard in #276
• Adding the ability to use dask arrays with chunks along spatial axes by @charlesgauthier-udm in #280
• Repare broken links to earthsystemcog by @huard in #292
• Parallel weight generation with Dask by @charlesgauthier-udm in #290
• Replace if statements by dict.get to reduce number of code switches by @huard in #295
• Warn of SpatialAverager error over large region and densify polygons by @charlesgauthier-udm in #293

New Contributors

• @rcaneill made their first contribution in #255
• @charlesgauthier-udm made their first contribution in #272

Full Changelog: v0.7.1...v0.8

v0.7.1

What's Changed

• Unpin and fix for Shapely 2.0 by @aulemahal in #219
• Implement workaround for setup conda problem by @raphaeldussin in #229
• Update CI and doc - fix for DataArrays by @aulemahal in #230
• Add util to build tripolar grid by @raphaeldussin in #228
• Fix ci/cd badge for build status by @pierre-manchon in #231
• Update CI for Micromamba environments by @Zeitsperre in #233
• Bump isort to 5.12.0 and add default language version to pre-commit by @huard in #237
• Document installation options for ESMpy by @mplough-kobold in #241
• Modernize the package a bit by @ocefpaf in #248
• Fix test so it works with pandas 2.0 by @huard in #251
• Minor fixes by @ocefpaf in #252

New Contributors

• @pierre-manchon made their first contribution in #231
• @mplough-kobold made their first contribution in #241

Full Changelog: v0.7.0...v0.7.1

v0.7.0

What's Changed

New features

• Add option to grid_global to vary longitudes from 0 to 360 by @huard in #207
• Modify the stop argument in arange to avoid coordinates going over the right bound by @huard in #206

Documentation

• Added section 'Adaptive Masking' to Masking.ipynb by @sol1105 in #187
• Update why.rst by @RichardScottOZ in #178
• Update index.rst by @RichardScottOZ in #179
• Typo fixes by @jsta in #191
• Add cftime and pandas to mock imports by @huard in #204
• Update README with current doc by @aulemahal in #209

Fixes

• Modify test so it works with xarray v2022-11 by @huard in #210
• Constrain numba >= 0.55.2 to make sure xESMF works with numpy 1.22. by @huard in #205
• Try importing esmpy (aliased to ESMF) for esmpy >= 8.4.0 fixes #212 by @raphaeldussin in #214

New Contributors

• @RichardScottOZ made their first contribution in #178
• @jsta made their first contribution in #191

Full Changelog: v0.6.3...v0.7.0

v0.6.3

Bug fixes

• Spatial coordinates of ds_out are kept within the regridder and transferred to the regridded DataArray or Dataset (#175). By Pascal Bourgault
• Added numba as an explicit dependency to fix installation with conda (#168). By Pascal Bourgault

Internal changes

• Use cf-xarray to guess missing CF coordinates before extracting bounds (#147). By Pascal Bourgault

v0.6.2

Bug fixes

• The introduction of sparse, with numba under the hood, restricted input data to little-endian dtypes. For big-endian dtypes, xESMF will convert to little-endian, regrid and convert back. By @aulemahal in #135
• SpatialAverager did not compute the same weights as Regridder when source cell areas were not uniform. By @huard in #128
• Refactor of how the regridding is called internally, to fix a bug with dask and sparse. By @aulemahal in #135

Internal changes

• Deprecation of regrid_numpy and regrid_dask is scheduled for 0.7.0. All checks on shape, array layout and numba support are now done at call time, rather then at computation time. By @aulemahal in #135

Relevant PRs

• Fix SpatialAverager to account for variable input grid cells areas by @huard in #128
• Add test checking that SpatialAverager does work with masked values by @huard in #130
• Regrid dataset with dask variables by @aulemahal in #122
• On-the-fly endianness conversion by @aulemahal in #125
• Unpin shapely by @aulemahal in #134
• Reduce dask task count, refactor regrid dispatch by @aulemahal in #135

Full Changelog: v0.6.1...v0.6.2

v0.6.1

Internal changes

• Weights are now stored in a xr.DataArray backed by sparse.COO, which allows to pass them as an argument to the xr.apply_ufunc and decrease memory usage when using dask.
• New dependency sparse replacing scipy.

v0.6.0

New features

• Add the skipna and na_threshold options to deal with masks over non-spatial dimensions (#29). This is useful when, for example, masks vary over time. By Stéphane Raynaud.
• Add unmapped_to_nan argument to regridder frontend. When True, this sets target cells outside the source domain to NaN instead of zero for all regridding methods except nearest neighbour (#94). By Martin Schupfner.

v0.5.3

Bug fixes

• Fix regression regarding support for non-CF-compliant coordinate names (#73). By Sam Levang
• Infer bounds dimension name using cf-xarray (#78). By Pascal Bourgault
• Do not regrid variables that are not defined over horizontal dimensions (#79). By Pascal Bourgault
• Ensure locstream dimension name is consistent with ds_out (#81). By Mattia Almansi

Documentation

• Add release instructions (#75). By David Huard
• Update Zenodo DOI badge

v0.5.2

Bug fixes

• Restore original behavior for lon/lat discovery, uses cf-xarray if lon/lat not found in dataset (#64)
• Solve issue of dimension order in dataset (#53) with (#66)

v0.5.1

• Update installation instructions to mention that PyPi only holds xesmf up to version 0.3.0 (#59)
• Regridded xarray.Dataset now preserves the name and attributes of target coordinates (#60)
• Fix doc build for API/Regridder (#61)