Fixing flaky tests

[hannahbaumann]

Fixes #43 and #67

This PR:

• Closes netcdfs both in the code and in the test
• Adds pooch caching back in
• Updates the test; the new simulation_skipped dataset is from CDK2, which has two chains. Since the multichain fix is not in yet, this test currently has large RMSDs
• Removes flaky argument in tests
• Uses the skipped test dataset in more tests for not having to use the bigger one everywhere.

[codecov]

Codecov Report

❌ Patch coverage is 97.43590% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 87.94%. Comparing base (5260e3b) to head (f42d063).
⚠️ Report is 7 commits behind head on main.

Files with missing lines	Patch %	Lines
src/openfe_analysis/rmsd.py	97.14%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #70      +/-   ##
==========================================
- Coverage   88.16%   87.94%   -0.23%     
==========================================
  Files           7        7              
  Lines         338      340       +2     
==========================================
+ Hits          298      299       +1     
- Misses         40       41       +1     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[atravitz]

@hannahbaumann you'll want to build pooch from main (like we are here in openfe) if you're seeing issues with fetching data.

[IAlibay]

Couple of things, still looking into that positions thing.

[IAlibay]

Do we still have a test somewhere to warn for the old netcdf file types?

If not, can we mock one?

[hannahbaumann]

The old one was the one with >500MB, I added a new test test_determine_position_indices_warns_for_old_nc that mocks this and gives a warning.

[hannahbaumann]

pre-commit.ci autofix

[hannahbaumann]

@IAlibay : After I removed the different u.trajectory.close(), the tests were flaky again with workers crashing, sometimes for macos, sometimes for ubuntu. I added the closing back in and now tests are stable again. I haven't seen the crashes locally.

[IAlibay]

@IAlibay : After I removed the different u.trajectory.close(), the tests were flaky again with workers crashing, sometimes for macos, sometimes for ubuntu. I added the closing back in and now tests are stable again. I haven't seen the crashes locally.

That means things aren't being deleted properly.. I'm not sure we can fix it in this PR, can you open an issue and we can revisit later?

[hannahbaumann]

@IAlibay , I opened an issue here #84 !

[IAlibay]

lgtm thanks!

just the one thing that needs to be done, will merge it directlt