This pull request is the first part of migrating the indexing and broadcasting logic from xarray.core.variable to xarray.namedarray. I intend to open follow-up pull requests to address additional changes related to this refactoring, as outlined in the proposal for decoupling lazy indexing functionality from NamedArray.

• [ ] Closes #xxxx
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [ ] Towards #8856
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

this PR builds towards

• https://github.com/pydata/xarray/pull/8870
• https://github.com/pydata/xarray/pull/8856

the primary objective is to partially address

1. Implement fall back .oindex, .vindex properties on BackendArray base class. These will simply rewrap the key tuple with the appropriate *Indexer object, and pass it on to __getitem__ or __setitem__. These methods will also raise DeprecationWarning so that external backends will know to migrate to .oindex, and .vindex over the next year.

• [ ] Closes #xxxx
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

As noted in https://github.com/pydata/xarray/issues/8909, the use of .oindex and .vindex properties in coding/* appears to have broken some backends (e.g. scipy). This PR reverts those changes. We plan to bundle these changes into a separate backends feature branch (see this comment, which will be merged once we are confident about its impact on downstream dependencies.

• [ ] Closes #8909
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [ ] Closes #xxxx
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [ ] Towards #8856
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

this is another follow-up to previous Prs

• 8780

• 8750

• 8790

, continuing the proposal outlined in plan for decoupling lazy indexing functionality from NamedArray

Cc @maxrjones for visibility

• [ ] Closes #xxxx
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

this is a follow-up to previous PRs (#8780 and #8750), continuing the efforts outlined in the plan for decoupling lazy indexing functionality from NamedArray. the primary focus of this PR is the removal of vectorized and orthogonal indexing logic from the __getitem__ method. Now, __getitem__ exclusively handles basic indexers, aligning with points 3 and 4 of the plan:

• as per point 3, lazy indexing arrays will now implement __getitem__ solely for basic indexing. For orthogonal indexing, .oindex will be used, and for vectorized indexing, .vindex will be utilized.
• following point 4, IndexingAdapter classes have been updated to consistently implement __getitem__, .oindex, and .vindex

• [ ] Closes #xxxx
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• 8827 follow up to fix https://readthedocs.org/projects/xray/builds/23743522/

• [ ] #8750 follow up
• [ ] Tests added
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

this pull request represents an initial effort to refactor the indexing logic within indexing.py by separating it from variable.py. to enhance the flexibility of explicitly indexed array classes, this PR introduces orthogonal ~~and vectorized~~ indexing methods—.oindex[key] ~~and .vindex[key[~~. these changes are a part of a broader initiative aimed at enabling indexing support for namedarray

• [x] Towards https://github.com/pydata/xarray/issues/8238
• [ ] Tests added
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

edit 1: - to minimize the blast radius of this PR, i decided to just focus on oindex[key] method in this PR. i intend to introduce vindex[key] in a separate PR

TODO:

• [x] https://github.com/pydata/xarray/pull/8750#discussion_r1491169460
• [ ] Enable passing the underlying tuple key

this pull request removes the "unnecessary optimization" introduced in https://github.com/pydata/xarray/pull/3588 that attempts to rewrite the indexer key for BasicIndexer and OuterIndexer

• [x] Closes #8377
• [ ] Tests added
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

This PR contains the design document for "named-array" a proposed new package aimed at providing a lightweight, efficient array structure with named dimensions. the main objective of named-array is to provide a more accessible version of Xarray's Variable class without the heavy dependencies like Pandas.

~~## proposed name~~

~~The "Xarray-lite" name is a placeholder for the time being. we have considered two main candidates for the package's name: "narray" and "yarray." feedback on the proposed names or other suggestions is welcome.~~

Related issues:

• https://github.com/pydata/xarray/issues/5081
• https://github.com/pydata/xarray/issues/3981

this PR copies the dtypes module from the xarray.core package to the xarray.namedarray package. It also recreates the ReprObject class from the utils module within the xarray.namedarray package

These changes are part of the ongoing work in: - https://github.com/pydata/xarray/pull/8244

• [ ] towards #8238
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• ref: https://github.com/xarray-contrib/xarray.dev/issues/66#issuecomment-1664561089

Cc @jhamman

really like how good these look on both light/dark backgrounds:

• [ ] towards fixing #8681
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

this is the initial phase of our plan to split the .set_dims() method into two distinct methods .expand_dims() and .broadcast_to() as outlined in the design document. - [x] add expand_dims() - [x] add broadcast_to() - [x] add .transpose() and .T to NamedArray - [x] update whats-new.rst

• [x] Closes #2620
• [x] 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

• [ ] #8365 follow up
• [ ] Tests added

• [ ] towards #8238
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

this pull request introduces the initial prototype for NamedArray as discussed in issue #8073. the intention is to approach the implementation in multiple phases through separate pull requests. the main objectives of this PR are as follows:

1. creation of a new class, xarray.named_array.NamedArray, which will encompass most of the necessary functionality for the lightweight version of xarray.Variable. As a result, xarray.Variable will inherit from this class and implement specific functionality related to xarray.

2. removal of all imports of internal Xarray classes and utilities from the xarray.named_array sub-package. the aim is to eliminate dependencies on anything within xarray.core. This will facilitate the future goal of moving named_array into its own standalone package, with xarray then depending on it.

3. [ ] Towards #3981

4. [x] Tests added
5. [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
6. [ ] New functions/methods are listed in api.rst

• [x] Closes #6649
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

This PR introduces small changes to enhance code maintainability...

• [x] Passes pre-commit run --all-files

• [ ] Addresses https://github.com/pydata/xarray/pull/5139#issuecomment-817362317
• [x] Passes pre-commit run --all-files

Here's a snippet of how this looks like when a pull request is created on the same main repository:

Cc @keewis

Libraries, including xarray, register backends via entrypoints. Xarray adds an EntryPoint "group" in its setup.py. Third party libraries add items to that group in their setup.pys.

Xarray and the third-party library don't need to talk to each other directly to register a backend. They talk through pkg_resources.

For instance, hvplot/holoviews would add an EntryPoint for xarray_plotting_backends in its setup.py:

python entry_points={"xarray_plotting_backends": ['holoviews = hvplot:plotting']}

and xarray users should be able to achieve what @jsignell proposed in #3553:

```python import xarray as xr air = xr.tutorial.open_dataset('air_temperature').load().air xr.set_options(plotting_backend='holoviews')

air.isel(time=500).plot() ```

• [ ] Closes #3553
• [x] Tests added
• [x] Passes black . && mypy . && flake8
• [ ] Fully documented, including whats-new.rst for all changes and api.rst for new API

• [x] Closes #7299
• [x] Tests added
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

This PR removes the existing team page from the docs and links to the newly created https://xarray.dev/team on the landing page

• [ ] Addresses https://github.com/pydata/xarray/pull/6884#issuecomment-1207094758
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [ ] #6530 follow-up

This PR supersedes https://github.com/pydata/xarray/pull/4540. This is still a very rough draft :)

• Closes
• [x] Closes #4473
• [x] Closes #4498
• [x] Closes #659
• [x] Closes #2237
• [x] xref https://github.com/pangeo-data/pangeo/issues/271
• [x] Tests added
• [x] Passes pre-commit run --all-files
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst

TODO: - [x] flox vs legacy tests with nans - [x] test passing flox-specific kwargs like engine and method.

• [x] Addresses https://github.com/pydata/xarray/issues/6027#issuecomment-1048737617

Cc @benbovy & @max-sixty

This PR ensures all CI workflows are only enabled on the main repository. A few workflows were still being triggered on forks.

• [x] Closes #6222
• [ ] Tests added
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [ ] Closes #6027

Preview:

We've been using pre-commit for the last month or so, and it seems to be working fine. This PR removes the redundant pre-commit hooks GHA workflow.

• [ ] Towards https://github.com/pydata/xarray/issues/5940

This fixes https://github.com/pydata/xarray/issues/5232#issuecomment-833729171

Currently, the CI runs both on a user's fork and pydata/xarray. This leads to duplicate runs and the user is flooded with GitHub notifications in case of failures... This PR addresses this issue by making sure the workflows are triggered on pydata/xarray only.

It turns out that pushes on main branch result in an invalid package version. To address this issue, I am disabling the push event trigger on the release workflow

Cc @keewis

```html Uploading xarray-0.17.1.dev108+gf5e4fd5f-py3-none-any.whl

0%| | 0.00/796k [00:00<?, ?B/s] 73%|███████▎ | 584k/796k [00:00<00:00, 3.39MB/s] 100%|██████████| 796k/796k [00:00<00:00, 1.52MB/s] Content received from server:

400 '0.17.1.dev108+gf5e4fd5f' is an invalid value for Version. Error: Can't use PEP 440 local versions. See packaging.python.org/specifications/core-metadata for more information.

HTTPError: 400 Bad Request from test.pypi.org/legacy '0.17.1.dev108+gf5e4fd5f' is an invalid value for Version. Error: Can't use PEP 440 local versions. See packaging.python.org/specifications/core-metadata for more information. ```

This addresses https://github.com/pydata/xarray/issues/5232#issuecomment-830242266, and will (hopefully) reduce the time it takes to cut a new release.

• [x] Passes pre-commit run --all-files

Follow up PR for #5174

• [x] Tests added
• [x] Passes pre-commit run --all-files

• [x] Closes #5248
• [x] Passes pre-commit run --all-files

As I understand it, the issue described in #5248 is caused by bad indentation... With this PR, we get the following:

@max-sixty, let me know if there are other instances that need to be fixed as part of this PR....

• [x] Passes pre-commit run --all-files

• This PR reduces the max-width for .container divs so as to address issues with RTD's multiple version badges.

• Latest: https://xarray.pydata.org/en/latest/index.html: the rtd badge box doesn't work

• This PR: https://xray--5039.org.readthedocs.build/en/5039/index.html: appears to be working fine

• Fixes CSS properties to enable appropriate toggle/responsiveness when using small screens. Unfortunately, this doesn't seem to have fixed the problem on small screens:

• [x] Passes pre-commit run --all-files

• [ ] Closes #xxxx
• [x] Passes pre-commit run --all-files
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

The styfle/cancel-workflow-action requires a token with write permissions in order to cancel duplicate runs. It turns out that the default GitHub token access made available to pull requests from forks has read permissions only. To address this issue, this PR introduces a new workflow that uses a special workflow_run trigger which allows the workflow to run to run with a token with write permissions.

• [x] Passes pre-commit run --all-files

Cc @keewis

As discussed in today's dev meeting, it'd be convenient to migrate all our CIs from azure pipelines to GitHub Actions.

• [x] Consolidate all style/type checks (isort, black, mypy, flake8) into a single workflow and use the pre-commit GitHub action.
• [x] Create a main CI workflow and parametrize it for different settings
• [x] Add minimum version policy workflow
• [x] Add Doctests workflow
• [x] Remove azure pipelines config
• [x] Update docs: remove references to azure pipelines

As discussed in today's dev meeting, this PR modifies the push and pull_request event triggers by making sure the upstream dev CI is triggered by cron schedule by default.

• [ ] 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

• [x] Closes #2840
• [x] Tests added
• [x] Passes isort . && black . && mypy . && flake8
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [x] Closes #4603
• [x] Closes #4643
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst

Towards https://github.com/pydata/xarray/issues/4549

• [ ] Closes #xxxx
• [ ] 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 GitHub action will run tests against the upstream dependencies. When the tests fail, the GH action will create an issue (with "CI" label). The body of the created issue includes:

• test summary report
• a link to the failed workflow run

Here's a couple screenshots from a toy repository:

• [x] Closes #4574
• [x] Tests added
• [x] Passes isort . && black . && mypy . && flake8
• [x] User visible changes (including notable bug fixes) are documented in whats-new.rst
• [ ] New functions/methods are listed in api.rst

• [x] Closes #3609
• [x] Tests added
• [x] Passes black . && mypy . && flake8
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

As pointed out in https://github.com/jbusecke/cmip6_preprocessing/issues/5, xarray appears to be very strict during the encoding and decoding steps even when there are (harmless) discrepancies between missing_value and _FillValue. For instance, when dtypes of missing_value and _FillValue are different, xarray gives up:

```python In [74]: from xarray.coding import variables

In [75]: import numpy as np

In [76]: import xarray as xr

In [77]: original = xr.Variable( ...: ("x",), ...: [0.0, -1.0, 1.0], ...: encoding={"_FillValue": np.float32(1e20), "missing_value": np.float64(1e20)}, ...: )

In [78]: coder = variables.CFMaskCoder()

In [79]: encoded = coder.encode(original)

ValueError Traceback (most recent call last) <ipython-input-79-9fbc3632e28b> in <module> ----> 1 encoded = coder.encode(original)

/glade/work/abanihi/devel/pangeo/xarray/xarray/coding/variables.py in encode(self, variable, name) 156 raise ValueError( 157 "Variable {!r} has multiple fill values {}. " --> 158 "Cannot encode data. ".format(name, [fv, mv]) 159 ) 160

ValueError: Variable None has multiple fill values [1e+20, 1e+20]. Cannot encode data. ```

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

Cc @dcherian

• Don't need dayofwk=-1 for cftime>=1.0.4

• [x] Closes #3426 and Fixes #3427
• [ ] Tests added
• [x] Passes black . && mypy . && flake8
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API