Due to issues like geneontology/noctua-models#23 and geneontology/noctua-models#18 that crop up when bad input is passed back to minerva, Noctua clients should take a more proactive roll in checking to make sure that the input is at least a little sane before making at call (until such time as Minerva can fend for itself).
Practically, as we seem to be having this issue almost exclusively with the graph editor client, the idea would be to slip in some code that would check to make sure that all individuals named have at least a colon in them somewhere before sending it back.
It may be easier to do this in the class-expression library, the widgetry (a la the macromolecular complex creator), or maybe a brute-force pre-call check.
Due to issues like geneontology/noctua-models#23 and geneontology/noctua-models#18 that crop up when bad input is passed back to minerva, Noctua clients should take a more proactive roll in checking to make sure that the input is at least a little sane before making at call (until such time as Minerva can fend for itself).
Practically, as we seem to be having this issue almost exclusively with the graph editor client, the idea would be to slip in some code that would check to make sure that all individuals named have at least a colon in them somewhere before sending it back.
It may be easier to do this in the class-expression library, the widgetry (a la the macromolecular complex creator), or maybe a brute-force pre-call check.