[d2rq-dev] Using the assembler to union d2rq to a TDB model [SEC=UNCLASSIFIED]

[Paul M. <pm...@an...>]

I hope this list is still live.

I am attempting to union a d2rq model to some data in TDB (and some static RDF files). The effect that I am trying to achieve is to add our new database to the data available at http://biodiversity.org.au/sparql/ . The TDB data there contains legacy identifiers. Our new database has mappings for these old IDs, so If I can jam the while lot into a single big graph, then it should be possible to span the data with SPARQL. 

To keep things organised, I am putting the different data sets into named graphs. I don't see that the d2r server does this at all.

I have tried a variety of things - I run into class version problems, incompatible interfaces, various other things. I'm at the stage of thinking that I need to hack up the d2r code and  rewrite SystemLoader, rigging it up to return a Graph that is not the d2r graph but that has that d2r graph as a subgraph. I have found that if I change the declaration of getModelD2RQ and getGraphD2RQ to return just the interface type, none of the existing code complains - but the point is that I shouldn't have to do this: there is a d2r jana assembler and it should just work.

So:

* Is there a version of fuseki/joseki that the d2rq assembler is known to work with?
Alternatively:
* can I launch the d2r server with an external assembler file?

---------------
If you are curious about the existing data, try this SPARQL query on http://biodiversity.org.au/sparql/ . (If you explore our data, please add a limit clause to all your sparql)

prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
prefix dcterms: <http://purl.org/dc/terms/>
prefix g: <http://biodiversity.org.au/voc/graph/GRAPH#>

select ?label ?title ?desc ?uri
where {
  graph g:meta {
    ?uri a g:GraphURI .
    OPTIONAL { ?uri rdfs:label ?label  } .
    OPTIONAL { ?uri dcterms:title ?title  } .
    OPTIONAL { ?uri dcterms:description ?desc  } .
  }
}
ORDER BY ?uri
LIMIT 100