Closed misialq closed 4 months ago
Hey @lizgehret, do you think you could look this over? I think I'll still need to add some tests for the new type introduced here but the rest of the code should be ready.
Just FYI, @nbokulich - these are changes we talked about today, in case you want to check those out.
Hey @lizgehret, I'm done with the changes here - I added one test for the new directory format and refactored some paths to test files for consistency.
Hmmm this is very weird - some tests are failing due to some directories supposedly not being there. I looked through setup.py and the tests themselves but cannot really see what the problem is. They also pass locally for me. Do you spot what's wrong @lizgehret? 👀
i think i found the issue(s) @misialq - letting the tests run now to confirm!
Aaaaaaah, all these commas 🤦 Thanks for fixing! Feel free to merge any time 🚀
This PRs refactors some of the
GenomeData
types and introduces new types which will be required in q2-moshpit. Below is the summary of the changes:SampleData[BLAST6]
type - this was used to store orthologs detected by EggNOG but will be replaced bySampleData[Orthologs]
(see below)SampleData[Orthologs]
andGenomeData[Orthologs]
to store the orthologs produced by EggNOG+Diamond; the BLAST6 variant was not really the best choice as the table produced by EggNOG bears only a degree of similarity to a "proper" BLAST6 result table; using the separate type for those orthologs will allow us to be more explicit and clear about the contents (the type actually already existed but we did not use it)OG
andKEGG
types - we are not using those anywhere and will not; they had been designed in the beginning of the project, before we knew exactly what outputs from functional annotation we would want to store and howGenomeData[DNASequence]
to represent collections of genome sequences fetched from various sources; we want to be able to store each genome in a separate file so theFeatureData[Sequence]
was not a good candidate