ENH: allow SampleData[MAGs] as input to eggnog-diamond-search and classify-kraken2

[misialq]

Closes #110.
Closes #112.
Depends on qiime2/q2-types#311.

Some sample MAGs can be found here and dereplicated MAGs here.

[misialq]

Hey @gregcaporaso, do you think someone from your team could have a look/test this (particularly the Kraken2 part)? 🙏

[github-actions]

This PR/issue depends on:

• ENH: add sample_dict method to MultiFASTADirectoryFormat qiime2/q2-types#311

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.82%. Comparing base (2b3b03d) to head (3ef2ed9).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #125      +/-   ##
==========================================
+ Coverage   97.74%   97.82%   +0.07%     
==========================================
  Files          64       64              
  Lines        3415     3534     +119     
==========================================
+ Hits         3338     3457     +119     
  Misses         77       77              

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

[Sann5]

Hey Michal! I all looks good to me. I already made some stand alone comments. As general suggestion I think I would have prefered if this was two PRs (one per action), but I imagine maybe you had a reason for doing them together.

• ✅ I tries the actions with some SampleData[MAGs] artifacts and it all works smoothly.
• ⚠️ All tests pass individually but when I run the whole test suite together I get error in ~12 tests for having too many files open. I think this is rather a limitation of my personal computer since this does not come up in the CI (or your own machine I suppose). Even so, this might prove annoying in the future for developers with weak computers like mine.

[misialq]

Hey @Sann5, thanks for your review! The behaviour you described, though, seems very strange - I never saw any of that when running the tests locally. Also, I don't see how this should happen as all the tests should be running one-by-one and things should get cleaned up between those runs. Could you provide more details as to that behaviour?:

• is it always the same set of tests?
• which files would those tests be using?
• how exactly are you running those?

Thanks! I'll merge this regardless as this should not be related to that PR but it would be nice if you could share some more details 🙏