Bump curation version to v0.26.0 (SCRUM-3367)

[mluypaert]

No description provided.

[markquintontulloch]

Do we really want to change diseaseRelation to relation everywhere? This will presumably require UI changes too. Or do we just want to change it in the indexer where pulling the information from the curation system and creating the search documents?

[mluypaert]

Do we really want to change diseaseRelation to relation everywhere? This will presumably require UI changes too. Or do we just want to change it in the indexer where pulling the information from the curation system and creating the search documents?

Well why did we change the property name but not the display name in the curation application? Changing backend name but not display name is an anti-pattern to my opinion and should be avoided unless there is a good reason to.

[markquintontulloch]

Do we really want to change diseaseRelation to relation everywhere? This will presumably require UI changes too. Or do we just want to change it in the indexer where pulling the information from the curation system and creating the search documents?

Well why did we change the property name but not the display name in the curation application? Changing backend name but not display name is an anti-pattern to my opinion and should be avoided unless there is a good reason to.

We changed the property name in the curation system because it was at odds with the LinkML model. The search documents do not follow the LinkML model. I agree that consistency is good, but think that this requires a bit more thought as to what we change and when. For example, why replace diseaseRelationNegation with relationNegation when this is something that is constructed for the search document, has nothing to do with the LinkmL model, and will probably result in more UI failures if changed?

[markquintontulloch]

diseaseRelationNegation is constructed from the negated field and the relation field. It is not part of the LinkML model and, since it is specific to disease annotations, I don't see any compelling reason to change the name. From a quick check, the UI code only seems to refer to diseaseRelationNegation and never directly to diseaseRelation/relation. As such, I'd be inclined to keep diseaseRelationNegation as it is so that this doesn't cause breaking changes in the UI repo.

That said, I'm not particularly familiar with the code in either this repo or the UI one so I will leave it to @cmpich or @oblodgett to approve this PR or request changes.

[mluypaert]

Alright, thanks, mind approving the UI PR as well @oblodgett? (alliance-genome/agr_ui#1213)