Better SPARQL: improve resolution of graphs URI

[vemonet]

At the moment we try to get only the classes of the graph

A few issues when resolving a graph URI:

• Not all classes are always retrieved on big graph: the query needs to be optimized. But this kind of query can be quite expensive to run, so we are using limit at the moment
• If the graph has no classes then the UI shows nothing (add fallback query which get any statement?)

[vemonet]

Would be interesting to have 2 approaches:

• Graphs are datasets (current way to explore)
• Graphs are "statements", smaller unit, so much more. For model like :
  • nanopubs
  • EU Cellar Law dataset
  • LOD SPARQL endpoint

Maybe as options in settings?