bioentityset/slimmer/function should provide a relationship_type argument to determine which closure to use

[deepakunni3]

Currently the GOlr query made by bioentityset/slimmer/function call is different in comparison to its analogous AmiGO query.

bioentityset/slimmer/function call is operating over isa_partof_closure instead of regulates_closure.

The solution would be to provide a relationship_type argument for this call, similar to the bioentity/goterm/{id}/genes/ call, where the relationship_type will determine which closure to use in the GOlr query:

• involved_in -> isa_partof_closure
• acts_upstream_of_or_within -> regulates_closure

@kltm @cmungall @selewis @nathandunn

[deepakunni3]

Adding comments from https://github.com/biolink/ontobio/pull/212,

In future iterations of our golr schema, we should simply name the closure fields for entity and association docs according to the gp2term relationship.

In other words, isa_partof_closure would be for GO ontology docs (since is_a and part_of are relations that hold e.g. between two processes), and involved_in and causally_upstream_of_or_within would be in assoc docs (since this holds between gp and a term)

This would obviate the need for a confusing mapping, where we say "for involved in use the isapartof closure" etc

[nathandunn]

@deepakunni3 Has this made its way into production?

[nathandunn]

https://agr-jira.atlassian.net/browse/AGR-1108

[deepakunni3]

@nathandunn The production BioLink API is still due an update.

[deepakunni3]

@nathandunn The changes should be live now.