Closed sierra-moxon closed 3 years ago
@sierra-moxon The models I spot-checked look good! Unfortunately there's a bug we discovered in the ART workshop today that affects Noctua Form and ShEx. I'm testing a fix for this right now. Once it's merged into ontobio the models would need to be regenerated. Sorry!
@sierra-moxon OK, that bug's been fixed in ontobio so, if you git pull from master
and regen, you should be good. Unfortunately, you'd also have to reset the modelstate
to "production" again. Grr, need to add that option to the gocamgen code.
@dustine32 - I regenerated the ZFIN models after updating to the latest version of ontobio to pull in the fix. I updated this branch accordingly so you should be able to see the differences here? Could you tell me what I'm looking for to see if it all worked properly with the fix? thanks a bunch! :). I also put in a PR for dev so that we could see it on noctua-dev (and I'm running the shex locally as well).
@sierra-moxon Oh, that fix should've added providedBy
properties to all NamedIndividual
s, like:
<http://model.geneontology.org/ZFIN_ZDB-GENE-000125-12/5d91e57a-b95d-4482-a269-1d15fd486dd9> a GO:0008285,
owl:NamedIndividual ;
dc:contributor "GOC:zfin_curators" ;
dc:date "2004-09-17" ;
ns2:providedBy "http://zfin.org" .
But I just see:
<http://model.geneontology.org/ZFIN_ZDB-GENE-000125-12/5d91e57a-b95d-4482-a269-1d15fd486dd9> a GO:0008285,
owl:NamedIndividual ;
dc:contributor "GOC:zfin_curators" ;
dc:date "2004-09-17" .
Either code didn't update or there's another bug preventing this. I can eventually try the ZFIN GPAD myself.
@dustine32 - ok, another attempt - these should have 'production' status, and providedBy tag. (see accompanying dev PR: https://github.com/geneontology/noctua-models/pull/187) :) Would you mind taking a quick look? (not ready to merge until rest of testing has been done).
Here we go!
round 2 attempt for a clean PR into master of all ZFIN models, without spurious deletes that we saw in: https://github.com/geneontology/noctua-models/pull/182 The large number of commits is just a result of the number of files being too long to add in a single commit. Travis is failing because the logging output is too long - should we just move to github actions?