Implement QC pipeline for monarch.owl

monarch-initiative / monarch-ontology

Top level monarch importer ontology

4 stars 3 forks source link

Implement QC pipeline for monarch.owl #11

Open cmungall opened 6 years ago

cmungall commented 6 years ago

[ ] Ensure ontology is coherent without having to do disjoint removal as a pre-processing step
[ ] Report/Avoid unintended class merges (see https://github.com/monarch-initiative/mondo-build/issues/52)
[ ] Ensure each ontology has an ontologyIRI that matches it's import (avoid issues like #10)
[ ] Avoid consuming stale portions of external ontologies in import (@balhoff)
[ ] build an owl-zip product and deploy on OSF/S3, with purl (avoid network latency issues when doing operations with large import chains)
[ ] create basic reports for ontology QC team to easily follow up on issues
[ ] integrate into monarch jenkins (wrap in Docker?)

matentzn commented 6 years ago

@cmungall I would be interested in contributing to this. Do you have any updates on the current state of the QC pipeline? Maybe we should schedule a short meet to discuss your vision for this.

cmungall commented 6 years ago

we don't really have much of an existing QC pipeline for upheno.

I would just start with the existing makefile, start adding more checks, have someone actually check the jenkins results and follow up where changes need to be made.

Ideally checks could be implemented in ROBOT. We are moving towards a large library of shared SPARQL checks. These can be extended with ontology-specific ones. We're also looking at ShEx (@dougl1sqrd, @balhoff).

@drseb has thought a lot about this and will have ideas; we also want to make sure zp is nicely integrated.

We want to tie this in with Monarch releases, cc @kshefchek