GO Xrefs are not updating in align with db-xref.yaml

[pgaudet]

Hello,

The page: http://geneontology.org/docs/go-annotation-file-gaf-format-2.1/ links to http://amigo.geneontology.org/xrefs (under 'DB (column 1)').

However http://amigo.geneontology.org/xrefs is out of date compared with the YAML file - for example Wormbase is not in the YAML file (https://github.com/geneontology/go-site/blob/master/metadata/db-xrefs.yaml) but is in the dbxref page.

Who is the intended audience for this page ? Based on this, we should decide what information we want to show users (for example do we need the information as to how to construct URLs)?

Thanks, Pascale

[suzialeksander]

These GAF files were outright copy/pasted from the old site. I've thought about making a simpler page as an overview for this GO-general-user audience, maybe linking to the wiki for the currently presented full-force breakdown of specs, but multiple pages doubles the maintenance.

Do we know why the YAML doesn't match the xrefs page? I see Wormbase/WB in both places, but a 26 March update for CARO isn't appearing (is this related to the pipeline hiccups?)

[suzialeksander]

Tagging to remind @pgaudet and I to review

[pgaudet]

@cmungall @kltm The page http://amigo.geneontology.org/xrefs is not being updated WRT the YAML files. We have 2 options:

1. Link to the yaml file directly (less user-friendly but at least it's current).
2. Create a new dynamic page

Which solution do you prefer ? We could do 1 for now and 2 when the new page is ready.

Thanks, Pascale

[kltm]

@pgaudet Can you point to the part that is not being updated there? For example: http://amigo.geneontology.org/xrefs#CARO I believe that the index rollout on release is regenerating the amigo page as well.

[lpalbou]

@kltm there is indeed a WormBase section in the amigo xrefs page which is not defined in the db-xrefs.yaml

I thought it could be related to amigo having its own db-xrefs.yaml but there is no "WormBase" section here either. Btw, did we took a decision on deprecating this duplicated YAML in amigo ?

Looking a bit further, I think the issue comes from amigo xrefs.js where xrefs seem to be hard coded

[pgaudet]

Thanks @lpalbou

It's be great if we could avoid hard-coding the info on the website.

Pascale

[kltm]

@pgaudet The re-generation of http://amigo.geneontology.org/xrefs is tied to AmiGO redeployments, which is in turn currently mostly connected to release updates. The embedded db-xrefs.yaml is an offline fallback for some older use cases and should not be active in the scenario. The JS is also a generated file and should not be active in this case.

[pgaudet]

@kltm What is the action then ?

[kltm]

I'll take a look at this after the next release (Any Day Now (tm)); if it persists, I'll dig in to see why it isn't updating.

[kltm]

Okay, let's make sure we have a concrete example to see if things are aligned. http://amigo-exp.geneontology.io/xrefs should be updating enough to get a handle on things and we can play with it no problem.

Looking at the most recent state of all things WormBase in db-xrefs.yaml (https://github.com/geneontology/go-site/blob/479ee371c9719385287e58edf5fbf8c6a78abea0/metadata/db-xrefs.yaml) we have:

• WB: https://github.com/geneontology/go-site/blob/479ee371c9719385287e58edf5fbf8c6a78abea0/metadata/db-xrefs.yaml#L3284
• WB_REF: https://github.com/geneontology/go-site/blob/479ee371c9719385287e58edf5fbf8c6a78abea0/metadata/db-xrefs.yaml#L3310
• WBbt: https://github.com/geneontology/go-site/blob/479ee371c9719385287e58edf5fbf8c6a78abea0/metadata/db-xrefs.yaml#L3320
• WBls: https://github.com/geneontology/go-site/blob/479ee371c9719385287e58edf5fbf8c6a78abea0/metadata/db-xrefs.yaml#L3333
• WBPhenotype: https://github.com/geneontology/go-site/blob/479ee371c9719385287e58edf5fbf8c6a78abea0/metadata/db-xrefs.yaml#L3343

Looking from http://amigo-exp.geneontology.io/xrefs#WormBase, all of these can be seen. Same seems to be true on production. There is one difference here in that the xrefs page unrolls the synonym into it's own "entry", but the reasons there are orthogonal to this as it is still pulling that from the db-xrefs.yaml (and I don't remember off the top of my head).

So, I guess I'm looking for a concrete example where this is not currently working, as it would seem that WormBase is, unless I'm missing something.

[vanaukenk]

@kltm - let's discuss because the URLs for the different entities under WB, e.g. variation and protein, do not look correct and I want to make sure I fix them correctly.

[kltm]

@vanaukenk Can do. Shall we work out a time to talk? We can continue this part a bit on email or skype or wherever.