Align to two different organisms for same project if requested

[gerikson]

See project 3512, some samples are human and others are mouse

[katsikora]

Project 3580, 3603

[adRn-s]

FWIW, https://parkour.ie-freiburg.mpg.de/api/requests/3512/get_flowcell/? shows that all samples in project 3512 are on the same flowcell (good to know first), then https://parkour.ie-freiburg.mpg.de/api/analysis_list/analysis_list/?flowcell_id=HNFT5DRX5 shows "organisms strings" (triplet) for each sample. The triplet is composed of name, which is the old string returned until 2024, plus: label for naming downstream analysis directories, and a yaml key or filepath for downstream analyses e.g. dorado, cellranger, snakePipes, etc.

edit: s/now/2024

[WardDeb]

Note that this used to be supported already (cfr. A00931, NB501361 as examples).
My guess is the change in label handling is what's to blame here (#103 , #104).

[katsikora]

yes, I rember it used to work and the code looks like it should work.
2 things:

1. the workflow command is run on all samples and not on the subset of (sampleid, sample name,...) tuples that match the organism and library type
2. ET Phone Home errors out

[katsikora]

to record here what I was discussing with Adrian,
the loop through the items of analysisTypes loops through organism_label while the resulting workflow command receives organism as input. I believe this is this might be the source of the issue.

[katsikora]

I have the fix working now, I'll push it to the repo next week.