Open colleenXu opened 1 year ago
Good to add this discussion to the TRAPI call agenda :)
Thanks @colleenXu for your comments. I'm leaving @sierra-moxon to mainly comment on this (as she has above). I do note, however, your mention of the type: object
versus $ref
. In fact, we should likely remove the type: object
property and move the $ref
to after the example:
tag, since the $ref
effectively effectively overwrites any subschema tags which follow it (see see $ref and Sibling Elements on https://swagger.io/docs/specification/using-ref/).
I've now submitted a PR for this: https://github.com/NCATSTranslator/ReasonerAPI/pull/428. @sierra-moxon, if we need to adjust anything else in this particular MetaEdge.association
property, we can perhaps use this PR to achieve this?
Still not sure that this was addressed in #428. @colleenXu is this still a concern? Not a schema issue, but an implementation discussion still needs to happen.
I'm not sure that the other points have been addressed:
I noticed that there's something that seems unrelated to MetaQualifiers in the MetaQualifier PR: this new
association
property on MetaEdges (see below).In https://github.com/NCATSTranslator/ReasonerAPI/pull/387/files:
I am wondering if:
association
will inherit the BiolinkEntity schema and will betype:string
, nottype:object
. Istype:object
then an error in the schema?