Control subsets

[pbuttigieg]

Linked to #440 #331 #346 #295.

There's a high likelihood that we're going to have a mass inflation of subsets as we interface with new communities and groups.

@cmungall it would be cumbersome to keep adding in-subset APs to the terms, could we use something similar to the import seeds to drive subset generation? That way, each subsetter can identify the terms that they need in a separate document.

[cmungall]

Yes, this would be easy to do.

What would the preferred format be? 2 column CSV of CURIE, label? This would be optionally extensible. E.g additional columns could say "also include all subsets". Worth looking at the OntoFox syntax for this.

[pbuttigieg]

2 column CSV of CURIE, label? This would be optionally extensible. E.g additional columns could say "also include all subsets". Worth looking at the OntoFox syntax for this.

That sounds good - and the include capability would streamline things even more.

[cmungall]

@DoctorBud's taboggan tool would be useful here.

@dougli1sqrd can work on the initial SPARQL to first extract the subsets embedded in the ontology into a csv. We'll then make the csv the source, and do the reverse operation to package the subset back into the owl as part of the release (Eric, this is related to what we're doing in GO).

[pbuttigieg]

@cmungall @DoctorBud @dougli1sqrd
Looks like we have a few more subsets in the ontology - will we be able to use this approach soon?
If it's a simple CSV format, users can probably maintain their own subsets with PRs, which would be quite cool.