Update db-xrefs.yaml

[suzialeksander]

• do the generic_urls (http://flybase.org -> https://www.alliancegenome.org/members/flybase ?) also need to be changed? 796 1559 2320 2331 2331 2755 2862 2904

• 1566 1572 are the MGI:MGI and may not be correct

Identifier type not in Alliance:

• SGD no SGD_REF
• WB no RNAi or protein page. no WB_REF
• no WBbt, WBls, WBPhenotypes
• XB morpholino oligo, transgene, line

[suzialeksander]

For #1943

[kltm]

@balhoff I'm noticing here that some rdf_uri_prefixs trailing slashes and some don't; do you know what the use is here and whether that matters?

[kltm]

@suzialeksander Okay, good news / bad news.
The good news is that I believe you did that URL changes correctly. 🎉
The bad news is that they cannot work with our current system. This is part of a consistent misunderstanding of how we do CURIEs. Essentially, for this to work, we'd need a single route at the Alliance that would get CURIEed identifier to the right location
See: #1906 (comment)

[balhoff]

@balhoff I'm noticing here that some rdf_uri_prefixs trailing slashes and some don't; do you know what the use is here and whether that matters?

@kltm I don't think trailing slashes are always needed, but I would expect some sort of delimiter (e.g., # or _).

[kltm]

@balhoff I'm noticing here that some rdf_uri_prefixs trailing slashes and some don't; do you know what the use is here and whether that matters?

@kltm I don't think trailing slashes are always needed, but I would expect some sort of delimiter (e.g., # or _).

@balhoff What should these look like then for our use case? Poking around, I'm not sure where they're even getting used (and we see a mix in db-xrefs.yaml as well, originating here e1372ce).

[kltm]

@suzialeksander @balhoff Okay, after talking to @pgaudet , I think the PR is okay to take now--I made a bunch of changes, so it would be best to look around.

@suzialeksander While the spirit of what you had is there, I did have to back out the changes to WB. There needs to be a single place for WB to go at the Alliance, just like there is with the others and like what WB has now. I don't know if we want to go ahead with what we have here or wait until we sort out the WB-specific issues (tagging @vanaukenk ).

[vanaukenk]

@kltm I don't know what would have to happen at the Alliance for things to work the way they do at WB. Is this a question for the Alliance technical team(s)?
Also, looking at the diff for this PR, is the situation with MGI the same?

[kltm]

@vanaukenk @ukemi Ugh, yes--that's the same issue. I'm going to revert MGI as well.

I don't know the way forward here as I have no "specialist" PM to have contact with anymore. This is something we can look into next week?

That said, if we can't do MGI and WB for the moment, it might be more confusing than not to let this PR in.

[ukemi]

Apologies for not being on board, but if I go to the file and use the URLs that were reverted they seem to resolve correctly.
For example:
https://www.alliancegenome.org/gene/MGI:1345277
https://www.alliancegenome.org/allele/MGI:3590672

However, a general MGI identifier, for example a genotype link, doesn't work:
https://www.alliancegenome.org/allele/MGI:2167526

[suzialeksander]

This does seem to be a mess, and I won't be remotely offended if this PR gets rejected and start fresh with a new one. @kltm I'll leave that to you, let me know if you do want me to try again- I think we're waiting for a report back from the Alliance though?

[kltm]

@suzialeksander I think it's probably best to bail on this PR for now and revisit after the meeting. I've converted it to "draft", so it's unlikely to get merged. Let's keep it around for a little for evidence of effort on #1943 and to give us a place to restart our work in a couple of weeks.

[kltm]

Noting from @sierra-moxon that the rdf_uri_prefix is consumed directly from out db-xrefs.yaml by prefixmaps.