Presentation : Konstantinos Kanaris. What is Jena? Usage of Jena Main Concepts Main Components Storage Models OWL API RDF API Reasoning.

Presentation : Konstantinos Kanaris

What is Jena? Usage of Jena Main Concepts Main Components Storage Models OWL API RDF API Reasoning Application-oriented issues

Open source Semantic Web framework for Java based on W3C recommendations for RDF and OWL

Provides a suitable programming environment for RDF, RDFS, OWL, SPARQL and an integrated rule-based inference mechanism

Import of owl ontologies, RDF, DAML into suitable models (persistent or in-memory)

Processing of these models Export of these models to owl, rdf, daml

files

Almost all ontology concepts: Classes: A class is a definition of a group of

individuals that belong together because they share common properties. For example John and Mary both belong to the class “person”.

Instances: Instances are what the classes consist of; their members, they inherit class characteristics and properties may be used to relate to one another

Properties: binary relations Object Properties: relations between instances

of two classes Datatype Properties: relations between

instances of classes and RDF literals and XML Schema datatypes

com.hp.hpl.jena.rdf.model: package for creating and manipulating RDF graphs Model: RDF model ModelMaker: ModelMaker contains a collection of

named models, methods for creating new models [both named and anonymous] and opening previously-named models, removing models, and accessing a single "default" Model for this Maker.

Literal: RDF Literal Statement: RDF Statement

com.hp.hpl.jena.ontology: package that provides a set of abstractions and convenience classes for accessing and manipluating ontologies represented in RDF OntModel: An enhanced view of a Jena model

that is known to contain ontology data, under a given ontology vocabulary (such as OWL)

OntClass: Interface that represents an ontology node characterising a class description ComplementClass IntersectionClass UnionClass

OntResource: Provides a common super-type for all of the abstractions in this ontology representation package

OntModelSpec: Encapsulates a description of the components of an ontology model, including the storage scheme, reasoner and language profile

ObjectProperty: Interface encapsulating properties whose range values are restricted to individuals (as distinct from datatype valued properties)

DatatypeProperties: Interface that encapsulates the class of properties whose range values are datatype values (as distinct from ObjectProperty whose values are individuals)

com.hp.hpl.jena.reasoner.Reasoner: reasoner subsystem designed to allow a range of inference engines to be plugged into Jena

com.hp.hpl.jena.reasoner.ReasonerRegistry: global registry of known reasoner modules

Persistent Storage Example Source Code:

OntModelSpec spec = new OntModelSpec(OntModelSpec.OWL_DL_MEM); IDBConnection conn = new DBConnection(dbURL, username, password, dbType); ModelMaker maker = ModelFactory.createModelRDBMaker(conn); Model m = maker.createModel("http://www.icsd.aegean.gr/ai-lab/projects/grid4all/ontology/grid4allOnto", false); m.read("file:G4Aonto-example.owl"); conn.close();

In Memory Source Code Example:

ModelMaker modelMaker = ModelFactory.createMemModelMaker(); Model model 1 = modelMaker.createModel("myBase"); java.io.InputStream in=FileManager.get().open(fc.getCurrentDirectory().getAbsolutePath()+"\\"+fc.getSelectedFile().getName()); model1.read(in, ""); OntModelSpec spec = new OntModelSpec(OntModelSpec.OWL_DL_MEM ); com.hp.hpl.jena.reasoner.Reasoner reasoner1 = PelletReasonerFactory.theInstance().create(); spec.setReasoner( reasoner1 ); OntModel ontTestModel = ModelFactory.createOntologyModel(spec, myModel);

Source Code Example: private ObjectProperty[]

getAllPropertiesWithSameDomainAs(ObjectProperty oP){ ObjectProperty[] oPz=new ObjectProperty[0];OntClass[] domainClasses=listDomainClasses(oP);OntModel oM=oP.getOntModel();ExtendedIterator it=oM.listObjectProperties();

while(it.hasNext()){ObjectProperty oP2=(ObjectProperty)it.next();

if(oP2.getLocalName().compareTo(oP.getLocalName())!=0){ OntClass[] domainClasses2=listDomainClasses(oP2); if(isSubclass(domainClasses,domainClasses2)){ oPz=extendArray(oPz,oP2); } } }

it.close(); return oPz; }

private String[] getAllSiblingClasses(OntClass Class2){ OntClass supClass=Class2.getSuperClass(); int iNoOfSubClasses=supClass.listSubClasses().toList().size(); String[] siblings = new String[iNoOfSubClasses-1]; int iCounter=0; for(int i=0; i<iNoOfSubClasses; i++){ OntClass oC=(OntClass)supClass.listSubClasses().toList().get(i); if(!oC.equals(Class2)){ siblings[iCounter]=oC.getLocalName(); iCounter++; } } return siblings; }

ExtendedIterator iter = ontModel1.listDatatypeProperties(); while(iter.hasNext()){ DatatypeProperty dP= (DatatypeProperty)iter.next(); if(dP.getRange().getLocalName().compareTo("int")==0 || dP.getRange().getLocalName().compareTo("double")==0){ String[] sProt=new String[4]; ExtendedIterator iter3= ontModel1.listIndividuals(); while(iter3.hasNext()){ OntResource oRes =(OntResource)iter3.next(); NodeIterator sit=oRes.listPropertyValues(dP); while(sit.hasNext()){ com.hp.hpl.jena.rdf.model.impl.LiteralImpl pr=(com.hp.hpl.jena.rdf.model.impl.LiteralImpl)sit.next();

String sSubject=oRes.getLocalName(); String sObject=pr.getString(); String sPredicate=dP.getLocalName();

Internal Reasoning com.hp.hpl.jena.reasoner.ReasonerRegistry ,mediocre

reasoning ability Source Code Example:

OntModelSpec spec = new OntModelSpec( OntModelSpec.OWL_DL_MEM ); com.hp.hpl.jena.reasoner.Reasoner reasoner1 = ReasonerRegistry.getOWLReasoner(); spec.setReasoner( reasoner1 );

ontTestModel = ModelFactory.createOntologyModel(spec, myModel);

External Reasoning E.g.:Pellet Reasoner, complete reasoning ability Source Code Example

OntModelSpec spec = new OntModelSpec( OntModelSpec.OWL_DL_MEM ); com.hp.hpl.jena.reasoner.Reasoner reasoner1 = PelletReasonerFactory.theInstance().create();spec.setReasoner( reasoner1 ); ontTestModel = ModelFactory.createOntologyModel(spec, myModel);

Jena 2.5.4 used One in memory RDF model used A persistent model approach was initially

introduced but was rejected Two in memory ontology models used

One with a reasoner attached One without reasoner

Pellet 1.5.1 reasoner used is some occasions only to determine subsumptions between subclasses and superclasses and their individuals respectively

Only ontology files imported (OWL-DL) No alterations on the imported ontology Only Knowledge Extraction Use of this knowledge to create Multiple

Choice Questions No SPARQL Queries developed

Limited use of the RDF API Used only to create ontology models and in

some occasions to detect certain statements. Extended use of the Ontology API

Knowledge extracted concerning: Classes Individuals Object Properties Datatype Properties

?????????????????????????????????????????????????????????????????__________???????????????????????????????????/__________\????????????????????????????????/_/?????????\_\??????????????????????????????\_\?????????/_/???????????????????????????????\_\??????/_/?????????????????????????????????????????/_/?????????????????????????????????????????/_/??????????????????????????????????????????|_|?????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????|_|?????????????????????