release v0.18.0

[dcherian]

As discussed in the meeting, we should issue a release soon with the new backend refactor and the new docs theme.

Here's a list of blockers:

• open_mfdataset: Raise if combine='by_coords' and concat_dim=None #5231
• lock kwarg needs a deprecation cycle? #5073
• Fix exception when display_expand_data=False for file-backed array. #5235

Would be nice and look done:

• Add GitHub action for publishing artifacts to PyPI #5244
• Bump versions #5258
• Surface plots #5101
• Refactor line plotting #4866 (we should let this sit on master for a while to find bugs)
• also apply combine_attrs to the attrs of the variables #4902
• Automatic duck array testing - reductions #4972 (this should probably also sit on master for a while)
• coarsen: better keep_attrs #5227
• Faster interp #4740
• Convert attribute and dimension names to strings when generating HTML repr #5149

Somewhat important, but no PR yet:

• Feature to retrieve the underlying store #5175 (as pointed out by @shoyer, this is really a new feature, not a regression, it can wait)

@TomNicholas and @alexamici volunteered to handle this. I can be online at release time to help with things if needed.

Release instructions are here: https://github.com/pydata/xarray/blob/master/HOW_TO_RELEASE.md

IIRC they'll need to be added to the PyPI list and RTD list.

[Illviljan]

May I suggest #4740, the only thing this needs I think is a rebase. I recently got a new computer and was reminded how slow the interpolation is without these changes.

[alexamici]

I propose the following timeline that is the fastest I can personally try:

• Monday May the 3rd, feature freeze, please refrain from merging new features unless they are discussed here
• during feature freeze everybody please help reviewing the PR in the list above
• Thursday the 6th, target release day, the best would be to schedule a window for a couple of hours (?) for @TomAugspurger and me to going through the release steps together with @dcherian.

@TomAugspurger and others does it work for you all?

[TomNicholas]

@alexamici I think you tagged the wrong Tom, but yes that works for me.

[max-sixty]

I think @alexamici 's plan is reasonable. I've had less of a process — no freeze etc — and that's worked fine — we do have a good test suite, and I've been really trying to make the release process a normal thing anyone can do in an hour.

But it's your call (and there's not much practical difference).

Thanks again!

[dcherian]

IIRC they'll need to be added to the PyPI list

We can avoid this by using the pypi github action thing to automatically build and upload when tagging a release on github. It uses a repo-level secret. Just a thought.

and RTD list.

I added @alexamici . @TomNicholas you need to sign up on RTD before I can add you there as a maintainer

[dopplershift]

We can avoid this by using the pypi github action thing to automatically build and upload when tagging a release on github. It uses a repo-level secret. Just a thought.

$0.02 from an outsider is that this has served us exceedingly well on MetPy. Our release process has become:

1. Close milestone
2. Adjust the auto-generated draft GitHub release (summary notes)
3. Click "publish release" -> packages uploaded to PyPI
4. Merge conda-forge update from their bots

It's almost more secure this way because the token from PyPI only has upload permissions--no need to store someone's password.

[TomNicholas]

(@dcherian I've signed up on RTD now)

[keewis]

great, you're a maintainer now (for pint-xarray, too, btw)

[dcherian]

It'd be nice to bump our minimum dependency versions if someone has the time.

[alexamici]

Form what I can see the #5258 "bump versions" looks ready, someone with knowledge about minimum versions please merge it.

After that is in, @dcherian you can probably uncomment the var._meta bit in #4740 "faster interp" and we can merge it as well. Again I'm not a user of that part of code, and I would feel more conformable if someone with more knowledge would click on the "merge" button.

This leaves only #5235 as a blocker that has a failing test, a proposed fix by @keewis. but the OP hasn't responded yet. How should we handle it? This looks ready to merge now. Sorry @tomwhite I read the timestamps wrong. Great job!