[Pellet-users] Pellet-backed models not being garbage collected

[Evren Sirin]

Garbage collection will happen only if you call model.close() 
explicitly. Otherwise some references to internal fields stay around 
preventing gc and  ultimately causing OutOfMemory. In your example 
program calling model.close() at the end of each iteration solved the 
problem on my side.

Cheers,
Evren

PS: There were some issues we identified regarding model.close() 
operation in Pellet (and will be fixed in the next version). That bug 
prevented gc even if you explicitly close the model. But that occurs 
when the same Pellet model was reused. This is not the case here so that 
bug should not affect you.

On 6/23/08 1:16 PM, Todor Dimitrov wrote:
> Hi everyone, 
>
> I'm trying to use pellet in a "production" system where different 
> ontologies have to be loaded upon request. The system is comprised of 
> a knowledge base with only ONE loaded pellet-backed model at a time. 
> I've assumed that whenever I create a new model the old one should be 
> garbage collected. This is obviously not the case as I keep getting 
> "OutOfMemory" errors. Here is a sample piece of code to test the behavior:
>
> /**
>  * *@author* Todor Dimitrov (todor.dimitrov at stud.uni-due.de 
> <mailto:dimitrov at stud.uni-due.de>)
>  * *@version* Jun 23, 2008 3:26:02 PM
>  */
> *public* *class* TestJenaPellet {
>
> /** The base URI for the domain model */
> *public* *static* *final* String /MODEL_URI/ = "http://www.owl-ontologies.com/domain-ontology.owl#";
>
> /** The base URI for the instances model */
> *public* *static* *final* String /INSTANCES_URI/ = "http://www.owl-ontologies.com/test-ontology.owl#";
>
> /** Prefix for all SPAQL queries */
> *public* *static* *final* String /SPARQL_PREFIX/ = "PREFIX domain:   
> <" + /MODEL_URI/
> + ">\n"
> + "PREFIX rdf:   <http://www.w3.org/1999/02/22-rdf-syntax-ns#>\n"
> + "PREFIX rdfs:  <http://www.w3.org/2000/01/rdf-schema#>\n"
> + "PREFIX owl:  <http://www.w3.org/2002/07/owl#>\n"
> + "PREFIX xsd:  <http://www.w3.org/2001/XMLSchema#>\n";
>
> *static* {
> *final* OntDocumentManager manager = OntDocumentManager./getInstance/();
> *final* String modelUri = /MODEL_URI/.substring(0, /MODEL_URI/
> .lastIndexOf("#"));
> *try* {
> manager.addAltEntry(modelUri, TestJenaPellet.*class*.getResource(
> "domain-ontology.owl").toString());
> } *catch* (Exception ex) {
> *throw* *new* RuntimeException(ex);
> }
> }
>
> /**
>  * Retrieves the pellet model
>  * 
>  * *@return* the non-null model
>  */
> *private* *static* OntModel getPelletModel() {
> *final* OntModel baseModel = /getBaseModel/();
> *return* ModelFactory./createOntologyModel/(PelletReasonerFactory./THE_SPEC/,
> baseModel);
> }
>
> /**
>  * Retrieves the base model
>  * 
>  * *@return* the non-null base model
>  */
> *private* *static* OntModel getBaseModel() {
> *final* OntModel ontModel = ModelFactory
> ./createOntologyModel/(OntModelSpec./OWL_DL_MEM/);
> *final* String modelUri = /MODEL_URI/.substring(0, /MODEL_URI/
> .lastIndexOf("#"));
> *final* String instanceUri = /INSTANCES_URI/.substring(0, /INSTANCES_URI/
> .lastIndexOf("#"));
> *final* Ontology ontology = ontModel.createOntology(instanceUri);
> ontology.addImport(ontModel.createResource(modelUri));
> ontModel.loadImports();
> *return* ontModel;
> }
>
> /**
>  * Executes SPARQL queries
>  */
> @Test
> *public* *void* testQueries() *throws* Exception {
> *final* String queryString = "select ?room where {?room rdf:type 
> domain:Room.}";
> *final* Set<Individual> rooms = *new* HashSet<Individual>();
> *for* (*int* i = 0; i < 10000; i++) {
> rooms.clear();
> *final* OntModel model = /getPelletModel/();
> // final OntModel model = getBaseModel();
> *final* QueryExecution qe = QueryExecutionFactory./create/(
> /SPARQL_PREFIX/ + queryString, model);
> *try* {
> // Execute the query and obtain results
> *final* ResultSet result = qe.execSelect();
> *while* (result.hasNext()) {
> *final* QuerySolution solution = result.nextSolution();
> *final* RDFNode node = solution.get("room");
> *final* Individual room = model.getIndividual(node.asNode()
> .getURI());
> rooms.add(room);
> }
> } *finally* {
> // Important - free up resources used running the query
> qe.close();
> }
> System./out/.println("Iteration " + (i + 1) + " :: number of rooms: "
> + rooms.size());
> }
> }
> }
>
> and the very simple ontology:
>
> <?xml version="1.0"?>
> <rdf:RDF
>     xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>     xmlns:protege="http://protege.stanford.edu/plugins/owl/protege#"
>     xmlns:xsp="http://www.owl-ontologies.com/2005/08/07/xsp.owl#"
>     xmlns:owl="http://www.w3.org/2002/07/owl#"
>     xmlns:xsd="http://www.w3.org/2001/XMLSchema#"
>     xmlns:swrl="http://www.w3.org/2003/11/swrl#"
>     xmlns:swrlb="http://www.w3.org/2003/11/swrlb#"
>     xmlns="http://www.owl-ontologies.com/domain-ontology.owl#"
>     xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#"
>   xml:base="http://www.owl-ontologies.com/domain-ontology.owl">
>   <owl:Ontology rdf:about=""/>
>   <owl:Class rdf:ID="Room"/>
>   <Room rdf:ID="Room_4"/>
>   <Room rdf:ID="Room_3"/>
>   <Room rdf:ID="Room_5"/>
>   <Room rdf:ID="Room_6"/>
>   <Room rdf:ID="Room_1"/>
>   <Room rdf:ID="Room_10"/>
>   <Room rdf:ID="Room_9"/>
>   <Room rdf:ID="Room_11"/>
>   <Room rdf:ID="Room_7"/>
>   <Room rdf:ID="Room_2"/>
>   <Room rdf:ID="Room_12"/>
>   <Room rdf:ID="Room_8"/>
> </rdf:RDF>
>
> In the "testQueries()" method approximately 700-800 iterations are 
> possible (with the default heap size!). If I only use the base model, 
> the program terminates normally. Does pellet cache any of the models 
> and is there a way to clear the cache?
>
>
> Thanks in advance,
>
> Todor
> ------------------------------------------------------------------------
>
> _______________________________________________
> Pellet-users mailing list
> Pellet-users at lists.owldl.com
> http://lists.owldl.com/mailman/listinfo/pellet-users
> _______________________________________________
>
> Sponsored by Clark & Parsia, LLC http://clarkparsia.com/