We (monarchinitiative.org) aggregate gene/phenotype data from lots of places. For instance on our page here we have information about a specific genotype of mouse from MGI.
Thus outgoing mappings like this:
"MGI" : "http://www.informatics.jax.org/accession/MGI:"
However, we want to also essentially advertise to others that we have information to offer about MGI entities and declare the permalinks where on our site they can be found.
Thus incoming mappings like this:
"MGI" : "https://monarchinitiative.org/resolve/MGI:"
1) Are there are other use cases like ours?
2) If using a context file, is there any convention for encoding 'incoming' versus 'outgoing'
Our RDF swerves this incoming/outgoing issue, however most of those consumers interested in our data do not have the skill or patience to build a sparql query. We are working toward HCLS dataset descriptions too; however, this effort is stalled due to issues about how to identify certain tricky entities such as institutions. Two @context files (one incoming, one outgoing) would achieve this in a snap. JSON-LD context scoping, as exemplified, can be useful for consuming either of these two contexts as needed. The question is:
1) What's the best way to distinguish incoming from outgoing in the biocontext registry?
Separate folders
Separate filename conventions
Something else?
From a maintenance perspective on Monarch's part, the incoming context file can be deterministically derived from the outgoing one, which will be the one that is hand maintained. For others like idot, this would not be so easy as the concept of a prefix is not yet well established there. Also, it may be useful to put a comment in each mapping regarding the regex we expect at that incoming link. Stuff falls in and out of identifiers and folks should be aware of what we expect.
We (monarchinitiative.org) aggregate gene/phenotype data from lots of places. For instance on our page here we have information about a specific genotype of mouse from MGI.
1) For our own internal purposes, we already have and use our JSON-LD context to declare that when we refer to MGI:2180790 we are in fact referring to the authoritative source http://www.informatics.jax.org/accession/MGI:2180790
Thus outgoing mappings like this:
"MGI" : "http://www.informatics.jax.org/accession/MGI:"However, we want to also essentially advertise to others that we have information to offer about MGI entities and declare the permalinks where on our site they can be found.
Thus incoming mappings like this:
"MGI" : "https://monarchinitiative.org/resolve/MGI:"1) Are there are other use cases like ours? 2) If using a context file, is there any convention for encoding 'incoming' versus 'outgoing'
Our RDF swerves this incoming/outgoing issue, however most of those consumers interested in our data do not have the skill or patience to build a sparql query. We are working toward HCLS dataset descriptions too; however, this effort is stalled due to issues about how to identify certain tricky entities such as institutions. Two @context files (one incoming, one outgoing) would achieve this in a snap. JSON-LD context scoping, as exemplified, can be useful for consuming either of these two contexts as needed. The question is:
1) What's the best way to distinguish incoming from outgoing in the biocontext registry? Separate folders Separate filename conventions Something else?
From a maintenance perspective on Monarch's part, the incoming context file can be deterministically derived from the outgoing one, which will be the one that is hand maintained. For others like idot, this would not be so easy as the concept of a prefix is not yet well established there. Also, it may be useful to put a comment in each mapping regarding the regex we expect at that incoming link. Stuff falls in and out of identifiers and folks should be aware of what we expect.