Mock run of new (proposed) SOP noctua-model replacement and updates

[kltm]

This is to be a standing ticket to support and discuss mock runs of current and proposed SOPs for https://github.com/orgs/geneontology/projects/87 . The reason to consolidate is to 1) make it look more like how it will be run in the future and 2) prevent the proliferation of expensive branches of noctua-models. Currently, we'll be testing:

• replaced_by ontology updates
• relations updates (i.e. Implement replace-terms command. #478
• in-place sparql updates (most recently, to change purls back to CURIEs)

Replaces #479 and #476 .

[kltm]

@vanaukenk It would be good to decide on a standard commit message for relation updates.

[kltm]

@balhoff In your example given at https://github.com/geneontology/minerva/pull/478/files#diff-aefb43779e951b6f961416f1195b5d9b753eca8a12a29f64a7a2cd8fb432e63dR226, you have a file class-replacements.tsv. where/what is this file? Is this for future use? (When testing, it seems to be required.)

[vanaukenk]

@kltm I'm thinking that it might be useful to have a generic commit message for any changes that are not made using a 'replaced by' tag in an ontology.

How about: "Automated change YYYY-MM-DD: CURIE:xxxxxxx updated to CURIE:yyyyyyy" ?

[kltm]

@vanaukenk We can't really have variables in a git commit message in this case as it is for the entire code tree at that moment in time, rather than individual files, so individual CURIEs would not make much sense.

[vanaukenk]

@kltm - I thought you were referring to the actual comment in the models (where I'm still wondering if we should use 'replaced by' for these relations updates since that has a specific meaning in ontologies.. @balhoff - any thoughts?)

For a git commit message, then, how about:
'Updating RO relations as specified in noctua-models/migration/replace_relations.tsv'

I am open to suggestions here, if there's something you and @balhoff think would be more informative or appropriate.

[kltm]

@vanaukenk Sounds good to me. I think my only comment would be something a step outward: geneontology/noctua-models#233 (comment)

[balhoff]

@balhoff In your example given at https://github.com/geneontology/minerva/pull/478/files#diff-aefb43779e951b6f961416f1195b5d9b753eca8a12a29f64a7a2cd8fb432e63dR226, you have a file class-replacements.tsv. where/what is this file? Is this for future use? (When testing, it seems to be required.)

Yes that is to be able to replace classes as well as relations. We should just create an empty file alongside the relations one for now.

[kltm]

Thank you for the clarification @balhoff ; I've updated the SOP.
The changes connected with just a relations change are available here: https://github.com/geneontology/noctua-models/commits/issue-minerva-483-test-updates

[kltm]

@vanaukenk A replaced_by changes set is available here: https://github.com/geneontology/noctua-models/commits/issue-minerva-483-test-updates
(I'm mostly sure I did this right, but there was some fiddliness.)