Desired TKG Data Platform Support and Endpoints

[RichardBruskiewich]

Just putting a stake in the ground to acknowledge that the TKG is anticipated to be a common "graph data" resource which may be persisted in many complementary ways and accessed with similar diversity:

1. Persistence: in Neo4j and in BlazeGraph (or other RDF triple store?).
2. Access: via REST API's (e.g. Knowledge Beacon API? SmartAPI?)? Via direct Neo4J Cypher queries? Via a SPARQL endpoint?

[cmungall]

We can make the neo4j canonical, but would be good to have a bg clone. We'll need to work out the standard patterns for edge properties cc @balhoff

access: bolt/cypher + sparql for bg. We can stand up a beacon easily, and SciGraph.

but the main use case will be exchange of large fragments into reasoner local graphs, so we need to define a bulk format and request mechanism. could be rest api or simple script, analogous to git clone

[balhoff]

We could possibly explore Blazegraph's Reification Done Right extension, but that loses the standards-based quality of RDF/SPARQL.

[andrewsu]

As mentioned on a recent call, we're going to investigate whether Wikibase (the software underlying Wikidata) would be suitable here. Still very early days, but we should know more in a few weeks after @stuppie and @andrawaag explore this at an upcoming hackathon...