issues

When opening files from path-like objects (str, pathlib.Path), the backend machinery (_dataset_from_backend_dataset) sets the "source" encoding. This is useful if we need the original path for additional processing, like writing to a similarly named file, or to extract additional metadata. This would be useful as well when using fsspec to open remote files.

In this PR, I'm extracting the path attribute that most fsspec objects have to set that value. I've considered using isinstance checks instead of the getattr-with-default, but the list of potential classes is too big to be practical (at least 4 classes just within fsspec itself).

If this sounds like a good idea, I'll update the documentation of the "source" encoding to mention this feature.

• [x] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst

Most notably: - bump python to 3.10 - bump sphinx to at least 5.0 - remove the pydata-sphinx-theme pin: sphinx-book-theme pins to a exact minor version so pinning as well does not change anything - xref https://github.com/executablebooks/sphinx-book-theme/issues/686

~Edit: it seems this is blocked by sphinx-book-theme pinning sphinx to >=3,<5. They already changed the pin, we're just waiting on a release~

It seems there is a numbagg package on conda-forge now.

Not sure what to do about the min-all-deps CI, but given that the most recent version of numbagg happened more than 12 months ago (more than 18 months, even) maybe we can just bump it to the version on conda-forge?

We started to pin the version of black used in the environment of blackdoc, but the version becomes out-of-date pretty quickly. The new hook I'm adding here is still experimental, but pretty limited in what it can destroy (the pre-commit configuration) so for now we can just review any new autoupdate PRs from the pre-commit-ci a bit more thoroughly.

In order to fix the failing tests in #4902 we need to expose combine_attrs to be able to properly construct the expected result (to be passed through to the combine function).

This overlaps with the fallback of the attrs_file code, which I removed for now. Maybe combine_attrs="override" would be better?

• [x] Tests added
• [x] Passes pre-commit run --all-files
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst

This attempts to fix the RTD timeouts. I think these are due to a warning from matplotlib, which we can probably just ignore for now.

• [x] attempts to close #4249
• [ ] Tests added
• [x] Passes isort . && black . && mypy . && flake8
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

This is part of the effort to add support for pint (see #3594) to Dataset objects (although it will probably be a test-only PR, just like #3643).

• [x] Tests added
• [x] Passes isort -rc . && black . && mypy . && flake8
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

The list of failing tests from #3594: * Dataset methods - __init__: Needs unit support in IndexVariable, and merge does not work yet (test bug is also possible) - aggregation: xarray does not implement __array_function__ (see #3917) - rank: depends on bottleneck and thus only works with numpy.array - ffill, bfill: uses bottleneck - interpolate_na: uses numpy.vectorize, which does not support NEP-18, yet - equals, identical: works (but no units / unit checking in IndexVariable) - broadcast_like: works (but no units / unit checking in IndexVariable) - to_stacked_array: no units in IndexVariable - sel, loc: no units in IndexVariable - interp, reindex: partially blocked by IndexVariable. reindex works with units in data, but interp uses scipy - interp_like, reindex_like: same as interp / reindex - quantile: works, but needs pint >= 0.12 - groupby_bins: needs pint >= 0.12 (for isclose) - rolling: uses numpy.lib.stride_tricks.as_strided - rolling_exp: uses numbagg (supports NEP-18, but pint doesn't support its functions)

As pointed out by @TomAugspurger in https://github.com/pydata/xarray/issues/4133#issuecomment-641332231, there are pre-built nightly wheels for numpy, scipy and pandas in the scipy-wheels-nightly repository.

Not sure how frequently these are updated, though, at least the numpy wheel doesn't really seem to be built daily.

• [x] Closes #4133

As promised in #3493, this adds integration tests for units. I'm doing this now rather than later since I encountered a few cases in #3643 where a increased test coverage for variables would have been helpful.

• [x] Tests added
• [x] Passes black . && mypy . && flake8