Open balhoff opened 1 year ago
As long as things are going in and out of Noctua as they currently are (i.e. "typeless" JSON), there shouldn't be any issue with things built on top of Minerva. That said, there are API pieces that currently make use of the sparql endpoints, so it's possible that a change in data type could make a difference. Tagging @sierra-moxon and @dustine32 for feedback.
We have long had comments pointing out that contributor and provider values are strings representing IRIs: https://github.com/geneontology/go-shapes/blob/f5b117b6f12bdffe72595f4409c2d18817709152/shapes/go-cam-shapes.shex#L120-L124
When we implemented model copy in Minerva, we used an IRI value for the
prov:wasDerivedFrom
property. This doesn't seem to cause any problems in Noctua. Is there anything stopping us from changing these to IRIs? (besides migrating the data and updating Minerva)IRI values are much better for triplestore queries, such as finding all models contributed to by a particular person.