Sparql Query With Inferencing

4
votes

i have some rdf & rdfs files and i want to use jena sparql implementation to query it and my code look like :

 //model of my rdf file 
 Model model = ModelFactory.createMemModelMaker().createDefaultModel();
 model.read(inputStream1, null);
//model of my ontology (word net) file
 Model onto = ModelFactory.createOntologyModel( OntModelSpec.RDFS_MEM_RDFS_INF);
 onto.read( inputStream2,null);
    String queryString =
                        "PREFIX rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#> "
                        + "PREFIX wn:<http://www.webkb.org/theKB_terms.rdf/wn#> "
                        + "SELECT ?person "
                        + "WHERE {"
                        + "  ?person    rdf:type   wn:Person  . "
                        + "      }";

    Query query = QueryFactory.create(queryString);
    QueryExecution qe = QueryExecutionFactory.create(query, ????);
    ResultSet results = qe.execSelect();
    ResultSetFormatter.out(System.out, results, query);
    qe.close();

and i have a wordNet Ontology in rdf file and i want to use this ontology in my query to do Inferencing automaticly (when i query for person the query should return eg. Man ,Woman) so how to link the ontology to my query? please help me.

update: now i have tow models : from which i should run my query ?

 QueryExecution qe = QueryExecutionFactory.create(query, ????);

thanks in advance.

1
sparqlontologyjena

1 Answers

7
votes

The key is to recognise that, in Jena, Model is the one of the central abstractions. An inferencing model is just a Model, in which some of the triples are present because they are entailed by inference rules rather than read in from the source document. Thus you only need to change the first line of your example, where you create the model initially.

While you can create inference models directly, it's often easiest just to create an OntModel with the required degree of inference support:

Model model = ModelFactory.createOntologyModel( OntModelSpec.RDFS_MEM_RDFS_INF );

If you want a different reasoner, or OWL support, you can select a different OntModelSpec constant. Be aware that large and/or complex models can make for slow queries.

Update (following edit of original question)

To reason over two models, you want the union. You can do this through OntModel's sub-model factility. I would change your example as follows (note: I haven't tested this code, but it should work):

String rdfFile = "... your RDF file location ...";
Model source = FileManager.get().loadModel( rdfFile );

String ontFile = "... your ontology file location ...";
Model ont = FileManager.get().loadModel( ontFile );

Model m = ModelFactory.createOntologyModel( OntModelSpec.RDFS_MEM_RDFS_INF, ont );
m.addSubModel( source );

String queryString =
                    "PREFIX rdf:<http://www.w3.org/1999/02/22-rdf-syntax-ns#> "
                    + "PREFIX wn:<http://www.webkb.org/theKB_terms.rdf/wn#> "
                    + "SELECT ?person "
                    + "WHERE {"
                    + "  ?person    rdf:type   wn:Person  . "
                    + "      }";

Query query = QueryFactory.create(queryString);
QueryExecution qe = QueryExecutionFactory.create(query, m);
ResultSet results = qe.execSelect();
ResultSetFormatter.out(System.out, results, query);
qe.close();