Support loading queries from arbitrary HTTP locations

[albertmeronyo]

• Add another route http://grlc.io/api/url
• After that we have parameters ?iri=http://mysparqlquery.rq
• That latter we HTTP GET request with preferred content text/plain (hopefully the result will come with YAML annotations)
• We make sure we take into account modification headers for caching
• Optionally we support parameters ?prefix=http://verylonguri?q1=query_name_1&q2=query_name2&... to allow APIs with multiple operations

[albertmeronyo]

Update: support both

• URI-encoded URLs that contain the query in accept: plan/text or accept: sparql
• and GET/POST supplied URLs for the query locations: e.g. http://grlc.io/api/url and then the query URL comes at GET/POST query-url: https://api.druid.datalegend.net/datasets/IISG/iisg-kg/queries/rosa-luxemburg-gallery/1 and its first parameter at param1: value1

[albertmeronyo]

• Try to push a protocol for query collections, so first there's a URI that represents the collection, and when you de-reference it you get in an HTTP header links to all the queries the collection contains

[albertmeronyo]

We're making this happen through an HTTP deferenceable list (e.g. in plain text) that containts the full list of query URIs to include in a spec