build: use uv and ruff

[slevang]

So this would be replacing poetry with yet another new shiny tool, but I would argue the benefits are worth it. uv is a drop-in replacement for pip, and ruff is their linter/formatter tool. Both are incredibly fast (written in rust), have a lot of convenient options, and are seeing rapid adoption across the python ecosystem.

One really nice feature of uv is the --resolution arg which I've used here to run tests on a "minimal" version, using lowest supported python version of 3.10 and lowest specified versions of all our primary dependencies. Then the 3.11 and 3.12 jobs run with --resolution=highest which will always have us testing against new versions.

The substantive changes of this PR are the pyproject.toml and the GH Actions. Everything else is only the result of running ruff's linter (almost all of them auto-fixed). Mostly unused imports, trimming whitespace, and a few other random things like truthy/falsy checks and unused f-strings.

I did mess with the docs a little to get that running, and rebuilt all the sphinx. Hopefully didn't screw anything up there.

Only thing I haven't tested is the build_command and uploading of wheels to pypi, but it should work fine.

[codecov]

Codecov Report

Attention: Patch coverage is 62.50000% with 9 lines in your changes are missing coverage. Please review.

❗ No coverage uploaded for pull request base (main@5c554a6). Click here to learn what that means.

❗ Current head cc4e73f differs from pull request most recent head 4dbe872. Consider uploading reports for the commit 4dbe872 to get more accurate results

Files	Patch %	Lines
xeofs/preprocessing/stacker.py	0.00%	6 Missing ⚠️
xeofs/models/gwpca.py	0.00%	2 Missing ⚠️
xeofs/utils/xarray_utils.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #160   +/-   ##
=======================================
  Coverage        ?   25.60%           
=======================================
  Files           ?       41           
  Lines           ?     3007           
  Branches        ?        0           
=======================================
  Hits            ?      770           
  Misses          ?     2237           
  Partials        ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[nicrie]

Hey @slevang that's awesome ! I heard about these things recently coming out from Astral and would love to give it a try and see how they work. Expecting my first child literally today :) - so I won't have time to review anytime soon. I'm OK giving it a try 👍

[slevang]

Amazing, congrats! Hope all goes smoothly, and enjoy those early days 🥰

[nicrie]

Nice! This is so much cleaner!

[nicrie]

That's great - looking forward to see it in action! Thank you @slevang !

[nicrie]

One additional thought: we probably want to update the contributing guide, specifically the section about how to create a local env. Instead of poetry, would you prefer pip install -e . and pip install -e .[docs] or rather uv?

[slevang]

Good point. I went with basic pip to keep things simple. I also had to add those two system dependencies to the conda env creation to get the docs dependencies to install properly.