This data as json
| sha | message | author_date | committer_date | raw_author | raw_committer | repo | author | committer |
|---|---|---|---|---|---|---|---|---|
| a7fb5a9fa1a2b829181ea9e4986b959f315350dd | use `dask.array.apply_gufunc` in `xr.apply_ufunc` (#4060) * ENH: use `dask.array.apply_gufunc` in `xr.apply_ufunc` for multiple outputs when `dask='parallelized'`, add/fix tests * DOC: Update docstring and whats-new.rst * WIP: apply_gufunc * WIP: apply_gufunc -> reinstate dask='allowed' as per @mathause, adapting tests * WIP: apply_gufunc -> add test for GH #4015, fix test for sparse meta checking * WIP: apply_gufunc -> remove unused `input_dims` * Update xarray/core/computation.py Co-authored-by: Mathias Hauser <mathause@users.noreply.github.com> * Update xarray/core/computation.py Co-authored-by: Mathias Hauser <mathause@users.noreply.github.com> * Update xarray/core/computation.py Co-authored-by: Mathias Hauser <mathause@users.noreply.github.com> * WIP: use dask_gufunc_kwargs, keep vectorize first but only for non-dask-gufunc, rework docstrings, adapt tests * DOC: add reference to internal changes in whats-new.rst * FIX: mypy * FIX: vectorize inside `apply_variable_ufunc` * TST: add tests from #4022 from @mathause * FIX: address black issue * FIX: xfail test for dask < 2.3 * WIP: apply changes in response to @mathause's review comments * WIP: remove line * WIP: catch different chunksize error and allow_rechunk, docstring fixes * WIP: remove comment * WIP: style issues * WIP: revert catch, revert test, add tests without output_dtypes * WIP: fix signature in apply_ufunc->apply_gufunc, handle output_sizes, handle dask version, fix tests * WIP: fix tuple * WIP: add dims_map to _UFuncSignature, adapt output_sizes to fit for apply_gufunc * WIP: black * WIP: raise ValueError if output_sizes dimension mismatch * WIP: raise ValueError if output_sizes is missing for given output_core_dims * WIP: simplify if/else * FIX: resolve conflicts prior merge with master * FIX: combine if's as per review * FIX: pass `vectorize` and `output_dtypes` kwargs explicitely into `apply_variable_ufunc` as per review suggestion * FIX: pass `vectorize` and `output_dtypes` kwargs explicitely into `da.apply_gufunc` * FIX: address review comments of @keewis and @mathause * FIX: black * FIX: `vectorize` not needed in if-clause * FIX: set DeprecationWarning and stacklevel=2 * FIX: use FutureWarning for user visibility * FIX: remove comment as suggested Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com> Co-authored-by: Mathias Hauser <mathause@users.noreply.github.com> Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com> | 2020-08-19T06:57:55Z | 2020-08-19T06:57:55Z | b1298f176a5cd512085eefae3e16f385bfba2d2a | cd792325681cbad9f663f2879d8b69f1edbb678f | 13221727 | 5821660 | 19864447 |