owlet is a query expansion preprocessor for SPARQL. It parses embedded OWL class expressions (in Manchester Syntax) and uses an OWL reasoner to replace them with a VALUES block containing the URIs of subclasses of the given class expression (or superclasses, equivalent classes, or instances).
owlet is written in Scala but can be used in any Java application.
owlet is available from Maven Central. To use it in your project, add the owlet dependency to your pom.xml:
<dependency>
<groupId>org.phenoscape</groupId>
<artifactId>owlet</artifactId>
<version>2.0</version>
</dependency>The OWL class expression to be expanded is a string literal at an appropriate location in a SPARQL query. The query expander recognizes the literal containing the class expression by its datatype, which must be http://purl.org/phenoscape/owlet/syntax#omn. Datatypes of literals in SPARQL are assigned using the ^^TypeURI suffix notation (where TypeURI is the URI of the datatype; URI prefixes are allowed).
The literal containing the class expression must be in one of the following 5 triple patterns (with rdf, rdfs, owl, sesame declared as in prefix.cc, and owlet as <http://purl.org/phenoscape/owlet/syntax#>):
- Subclass triple:
?x rdfs:subClassOf "owl:Thing"^^owlet:omn - Equivalent class triple:
?x owl:equivalentClass "owl:Thing"^^owlet:omn - Instance of triple:
?x rdf:type "owl:Thing"^^owlet:omn - Direct subclass triple:
?x sesame:directSubClassOf "owl:Thing"^^owlet:omn - Direct instance of triple:
?x sesame:directType "owl:Thing"^^owlet:omn
Each of these triples will be rewritten by the expander in the following pattern:
VALUES ?x { <URI1> <URI2> <URI3> ... }
where <URI1>,<URI2>,<URI3>,... is the list of class or instance identifiers that satisfy the OWL class expression, as determined by the reasoner.
Below is an example of loading an ontology from a file into an OWL reasoner, and then using the reasoner to expand a SPARQL query to a triple store. In this example, the triple store is in-memory and holds the same ontology as the reasoner, but you could replace that part with querying a remote SPARQL endpoint that, for example, contains data linked to the ontology. The owlet tests include executable versions in both Java and Scala.
import java.io.IOException;
import java.io.InputStream;
import org.junit.Test;
import org.semanticweb.elk.owlapi.ElkReasonerFactory;
import org.semanticweb.owlapi.apibinding.OWLManager;
import org.semanticweb.owlapi.model.OWLOntology;
import org.semanticweb.owlapi.model.OWLOntologyCreationException;
import org.semanticweb.owlapi.model.OWLOntologyManager;
import org.semanticweb.owlapi.reasoner.InferenceType;
import org.semanticweb.owlapi.reasoner.OWLReasoner;
import com.hp.hpl.jena.query.Query;
import com.hp.hpl.jena.query.QueryExecutionFactory;
import com.hp.hpl.jena.query.QueryFactory;
import com.hp.hpl.jena.query.ResultSet;
import com.hp.hpl.jena.rdf.model.Model;
import com.hp.hpl.jena.rdf.model.ModelFactory;
import org.phenoscape.owlet.QueryExpander;
public class SPARQLQueryRunner {
public void runSPARQLQuery() throws OWLOntologyCreationException, IOException {
final OWLOntologyManager manager = OWLManager.createOWLOntologyManager();
final OWLOntology ontology = manager.loadOntologyFromOntologyDocument(new File("my_ontology.owl"));
final OWLReasoner reasoner = new ElkReasonerFactory().createReasoner(ontology);
reasoner.precomputeInferences(InferenceType.CLASS_HIERARCHY);
final Model rdfModel = ModelFactory.createDefaultModel();
rdfModel.read("my_ontology.owl");
final String queryText = "PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>\n"
+ "PREFIX ow: <http://purl.org/phenoscape/owlet/syntax#>\n"
+ "PREFIX axial_skeleton: <http://purl.obolibrary.org/obo/VSAO_0000056>\n"
+ "PREFIX definition: <http://purl.obolibrary.org/obo/IAO_0000115>\n"
+ "PREFIX part_of: <http://purl.obolibrary.org/obo/BFO_0000050>\n"
+ "SELECT DISTINCT ?structure ?label ?definition\n"
+ "WHERE\n"
+ "{\n"
+ "?structure rdfs:label ?label .\n"
+ "?structure definition: ?definition .\n"
+ "?structure rdfs:subClassOf \"part_of: some axial_skeleton:\"^^ow:omn .\n"
+ "}";
final QueryExpander expander = new QueryExpander(reasoner);
final Query query = QueryFactory.create(queryText);
final Query expandedQuery = expander.expandQuery(query);
final ResultSet results = QueryExecutionFactory.create(expandedQuery, rdfModel).execSelect();
}
}More documentation can be found on the owlet wiki.