Results of the Discovery API as RDF

[enridaga]

Implement an RDF version of the results of the discovery API.
Possibly using:

• SPIN http://www.w3.org/Submission/spin-modeling/
• LSQ https://github.com/AKSW/LSQ/blob/gh-pages/LSQ_Vocab.rdf

[LaurensRietveld]

+1. How does this relate to #50 ?
Would this RDF data via e.g. a SPARQL endpoint replace the search API?

[enridaga]

When a request asks for an RDF format, the response will be an RDF graph using SPIN+LSQ instead of a tabular list of APIs and hits. This might be incoherent with the actual response format, that is a plain JSON including the list of APIs and details about the hits in the text.
I don't think this is a problem, as the plain JSON will be used to display the search results, while the RDF formats for other use cases.

Adding a SPARQL endpoint to query the query catalogue would be a nice feature, but it would require a bit of work, so I won't put this requirement as a dependency right now, and we might reuse the code needed to generate the RDF descriptions to populate it. I think we should go incrementally.

Similarly, we might use these descriptions to provide a /meta subresource for each query, as figured on #13 .