Closed stevencox closed 6 years ago
On beacons, implemented one for RENCI's ChemBio SPARQL endpoint. The underlying data (chem2bio2rdf) lacks the hierarchical conceptual semantic structure in the designs of both Wikidata and Monarch data sets. As a result, implementing beacon semantics would involve manually enumerating queries for each "kind" of thing in the data. This was deemed not a helpful activity. I'll be considering a potential role for GraphQL next.
In pursuit of a comprehensive Translator service architecture, I've made progress on the GraphQL implementation of a Green Translator API. The alpha version is simplistic, uses Strings for types, and is deployed in an experimental mode.
I'm currently implementing a new version that will use strong types and a better framework (Graphene). I'm also preparing a blog post to document findings about GraphQL and where it fits in the landscape of smartAPI, SPARQL, and other technologies. In particular, the architecture team is interested in how it might integrate with the effort to build a service registry and systems like TKBio.
In terms of status:
Deployed alpha version of Green GraphQL API and wrote blog post. Shared w/Green architecture for feedback. Next steps are the rest of the points in the previous comment.
Further progress on this item and the related activity of updating CQ notebooks to use this service is largely blocked waiting for service or infrastructure availability:
Working with @mjstealey and @arunacs to move towards a generic exposures oriented interface.
Develop method to make SPARQL and GraphQL APIs first class citizens of the Translator Registry.
Made progress on GraphQL representation at the Hackathon. Next steps - change representation of service in green API description to use postBody:
Supplanted by work on Gamma, now producing a meta Translator API
Participate in translator architecture efforts to find a good architecture for federated services.
Current candidates include: