✨ Add Curator for tiledbsoma stores

[Koncopd]

This PR adds SOMACurator class to curate tiledbsoma stores.

curator = ln.Curator.from_tiledbsoma(
    "./curate.tiledbsoma",
    {"RNA": ("var_id", bt.Gene.symbol)},
    categoricals={"cell_type": bt.CellType.name},
    organism="human",
)
curator.validate()
curator.standardize("all")
curator.add_new_from("all")
curator.save_artifact()

[github-actions]

🚀 Deployed on https://674edefc172d89071fbe998d--lamindb-qnwk.netlify.app

[codecov]

Codecov Report

Attention: Patch coverage is 99.05660% with 2 lines in your changes missing coverage. Please review.

Project coverage is 92.90%. Comparing base (4c27162) to head (c6472e9).
Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
lamindb/_curate.py	99.05%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2228      +/-   ##
==========================================
+ Coverage   92.71%   92.90%   +0.18%     
==========================================
  Files          55       55              
  Lines        6947     7159     +212     
==========================================
+ Hits         6441     6651     +210     
- Misses        506      508       +2     

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

[falexwolf]

This is good from my end. The bigger questions are for @sunnyosun and @Zethson about conventions for the user-facing API design and code organizations.

This is another thread.

@sunnyosun should also review this PR though.

[falexwolf]

let's please not expose using_key for any new Curator class. It should disappear from all.

(Unless you tell me we absolutely need this @Zethson @sunnyosun, I continue to think that this is a horribly complicated design choice and a very bad name for an argument. I'll make a new GitHub issue on this topic or try to find where we previously discussed this).

[sunnyosun]

Agreed, let's hide it.

[Zethson]

I think it's useful to extend the CxG curator. By default, it uses the laminlabs/cellxgene instance but when extending it (e.g. with perturbations) you want to now not curate against laminlabs/cellxgene but your own instance.

I've also never liked the name but one of you wanted it renamed from using to using_key.

[Koncopd]

Removed the description of the argument, put it to the bottom.

[falexwolf]

This is impressive, Sergei!

All my comments are minor. Last request is to populate the PR description.

[Zethson]

Awesome! I recognized quite many parts of the code and we can look out for ways to consolidate them in the future I hope.

[sunnyosun]

Agreed, let's hide it.

[Zethson]

Suggested change

	exclude: A dictionary mapping column names to values to exclude.
	exclude: A dictionary mapping column names to values to exclude from validation.
	When specific :class:~bionty.Source instances are pinned and may lack default values (e.g., "unknown" or "na"), using the exclude parameter ensures they are not validated.

@Koncopd . Written on Github, so please check syntax + formatting.

[Koncopd]

added

[Zethson]

Suggested change

	>>> curate = ln.Curator.from_tiledbsoma(
	>>> curator = ln.Curator.from_tiledbsoma(

Thought we're using that everywhere.

[Koncopd]

fixed