SPIN API - inferencing over a Sesame repository

558 views

Skip to first unread message

Alison Callahan

unread,

Jan 31, 2012, 5:17:39 PM1/31/12

to topbrai...@googlegroups.com

Hello all,

I am using the SPIN API 1.2.0 to execute SPIN rules using the SPINInferences.run(com.hp.hpl.jena.rdf.model.Model queryModel, com.hp.hpl.jena.rdf.model.Model newTriples,SPINExplanations explanations, List<SPINStatistics> statistics, boolean singlePass, ProgressMonitor monito) method. However, rather than having to load all of my SPIN rules/linked data/ontologies into a Jena model (quite memory intensive) I would like to instead connect to an OWLIM Sesame repository, as is done by TopBraid Composer.

I am able to connect to an OWLIM Sesame repository using Jena using the Jena DatasetGraph interface (see http://owlim.ontotext.com/display/OWLIMv43/OWLIM-SE+Installation#OWLIM-SEInstallation-InstantiateOWLIMSEadapterusingtheprovidedAssembler), but it would defeat the purpose of connecting to an OWLIM Sesame repository to then load the DatasetGraph into a Jena model to pass into SPINInferences.run().

Can anyone provide suggestions as to how to pass a Jena DatasetGraph, or otherwise connect to a Sesame repository such that I can use it with SPINInferences.run()?

Thanks,

Alison

Peter Ansell

unread,

Jan 31, 2012, 6:19:12 PM1/31/12

to topbrai...@googlegroups.com

Hi Allison,

I am not sure if there is the could try the JenaSesame library from
either my fork at https://github.com/ansell/JenaSesame or Andy
Seaborne https://github.com/afs/JenaSesame . I haven't quite completed
and tested the sesame-jena implementation but when I do I will send a
pull request back to Andys fork, but you shouldn't need to use
sesame-jena from what I can tell.

The jena-sesame module which should be very similar in implementation
between my fork and Andys, (as opposed to the sesame-jena module)
enables the creation and use of either a Jena DatasetGraph or a Jena
Model as a wrapper around an existing Sesame RepositoryConnection. It
is very alpha software though, but all of the tests that exist pass.
Feel free to fork it and send changes back if you find bugs.

Hope that helps,

Peter

> --
> You received this message because you are subscribed to the Google
> Group "TopBraid Suite Users", the topics of which include Enterprise
> Vocabulary Network (EVN), TopBraid Composer,
> TopBraid Live, TopBraid Ensemble, SPARQLMotion and SPIN.
> To post to this group, send email to
> topbrai...@googlegroups.com
> To unsubscribe from this group, send email to
> topbraid-user...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/topbraid-users?hl=en

Alison Callahan

unread,

Feb 2, 2012, 1:05:23 PM2/2/12

to topbrai...@googlegroups.com, ansell...@gmail.com

Hi Peter,

Thanks for your help. I have forked your JenaSesame repository and am trying it out with my code.

I have instantiated a Jena model from a remote OWLIM repository using JenaSesame.createModel(repositoryConnection), and tried to do a simple test using the model.listNameSpaces() method, with the result that after 3200 seconds I got a Java OutOfMemory error. I had given Eclipse 4GB to work with.

All that to say, can you tell me more about how the JenaSesame.createModel(RepositoryConnection connection) method works? Does it populate a model with all of the statements in a OWLIM Sesame repository, or does it access them as needed?

Thanks again,

Alison

Peter Ansell

unread,

Feb 2, 2012, 4:04:50 PM2/2/12

to Alison Callahan, topbrai...@googlegroups.com

HI Alison,

The GraphRepository that is created by JenaSesame in the background is
definitely a streaming wrapper, using iterators to access the results,
so it is less likely that it is the wrapper causing the issue,
although it could be.

I don't think any of the current JenaSesame tests use listNameSpaces,
so we will need to do some testing to see why that is happening. The
nature of that method is such that Jena itself would need to store all
of the results in memory to make sure the results set is distinct
before emitting any results to free up memory. Have you tested the
performance of some of the other Model methods? Does listSubjects
(also distinct-based so may have the same issue) or listStatements
(should stream results cleanly) give you results?

Note that JenaSesame is setup to use jena-2.7.0-incubating, so it will
not work with SPIN-API just yet per the focus on getting a jena-2.6.4
based release out right now. I have done some experiments with getting
SPIN-API to work with jena-2.7.0-incubating [1] to get around this but
they haven't gone far just yet

Cheers,

Peter

[1] https://github.com/ansell/spin/tree/feature/jena27upgrade

Alison Callahan

unread,

Feb 3, 2012, 11:37:35 AM2/3/12

to Peter Ansell, topbrai...@googlegroups.com

Hi again Peter and all,

I did eventually get the JenaSesame code working (I had forgotten to add VM arguments to the JUnit test to increase the heap size), and was able to list the statements in the remote Sesame repository that was connected to a Jena model using JenaSesame.createModel(repositoryConnection).

However, when I added the Jena model connected to the OWLIM Sesame repository as a submodel to the model passed into the SPINInferences.run() method, and ran SPIN inferencing, there were no inferencing results.

Here is the relevant code:

****************************************

SPINModuleRegistry.get().init();

Model inputModel = ModelFactory.createDefaultModel();

OntModel spinModel = ModelFactory.createOntologyModel( OntModelSpec.OWL_MEM );

//add system triples

spinModel.add(SystemTriples.getVocabularyModel());

//add SPIN rules

spinModel.read("http://localhost/tmp/rules.spin.rdf");

//register locally defined functions

SPINModuleRegistry.get().registerAll(spinModel, null);

//connect to OWLIM repository

repositoryManager = new RemoteRepositoryManager("http://localhost:8080/openrdf-sesame");

try {

// Initialise the repository manager

repositoryManager.initialize();

// Get repository

repository = repositoryManager.getRepository("test");

// Open a connection to this repository

repositoryConnection = repository.getConnection();

//create the DatasetGraph instance

//dataset = new SesameDataset(repositoryConnection);

owlimModel = JenaSesame.createModel(repositoryConnection);

} catch (RepositoryException e) {

e.printStackTrace();

} catch (RepositoryConfigException e) {

e.printStackTrace();

}

//add owlim model

spinModel.addSubModel(owlimModel);

//add input model

spinModel.addSubModel(inputModel);

//create output model to contain inferred statements

Model outputModel = ModelFactory.createDefaultModel();

//add output model to SPIN model

spinModel.addSubModel(outputModel);

//run SPIN inferences

SPINInferences.run(spinModel, outputModel, null, null, false, null);

//get number of inferred statements

//System.out.println("Inferred triples: " + outputModel.size());

//get inferred statements

StmtIterator itr = outputModel.listStatements();

System.out.println("Iterating over statements");

while(itr.hasNext()){

Statement s = itr.next();

System.out.println(s.toString());

}

****************************************

So, I need another solution ... are there any working examples of using the SPIN API to execute SPIN rules over a Jena model that is connected to a remote Sesame repository? Presumably TopBraid Composer does something similar in order to manage inferencing over large sets of triples, but is there a way to reproduce this behaviour using the SPIN API?

Any suggestions are much appreciated!

Alison

Jerven Bolleman

unread,

Feb 3, 2012, 11:41:20 AM2/3/12

to topbrai...@googlegroups.com

Hi Allison,

I see that you are actually using OWLIM. Have you looked at

http://owlim.ontotext.com/display/OWLIMv43/OWLIM-SE+Usage+Scenarios#OWLIM-SEUsageScenarios-UsingOWLIMSEwithJena

Regards,
Jerven

> <mailto:alison....@gmail.com>> wrote:
> > Hi Peter,
> >
> > Thanks for your help. I have forked your JenaSesame repository
> and am trying
> > it out with my code.
> >
> > I have instantiated a Jena model from a remote OWLIM repository using
> > JenaSesame.createModel(repositoryConnection), and tried to do a
> simple test
> > using the model.listNameSpaces() method, with the result that
> after 3200
> > seconds I got a Java OutOfMemory error. I had given Eclipse 4GB
> to work
> > with.
> >
> > All that to say, can you tell me more about how the
> > JenaSesame.createModel(RepositoryConnection connection) method
> works? Does
> > it populate a model with all of the statements in a OWLIM Sesame
> repository,
> > or does it access them as needed?
> >
> > Thanks again,
> >
> > Alison
> >
> >
> > On Tue, Jan 31, 2012 at 6:19 PM, Peter Ansell

> <ansell...@gmail.com <mailto:ansell...@gmail.com>>

> > wrote:
> >>
> >> Hi Allison,
> >>
> >> I am not sure if there is the could try the JenaSesame library from
> >> either my fork at https://github.com/ansell/JenaSesame or Andy
> >> Seaborne https://github.com/afs/JenaSesame . I haven't quite
> completed
> >> and tested the sesame-jena implementation but when I do I will
> send a
> >> pull request back to Andys fork, but you shouldn't need to use
> >> sesame-jena from what I can tell.
> >>
> >> The jena-sesame module which should be very similar in
> implementation
> >> between my fork and Andys, (as opposed to the sesame-jena module)
> >> enables the creation and use of either a Jena DatasetGraph or a Jena
> >> Model as a wrapper around an existing Sesame
> RepositoryConnection. It
> >> is very alpha software though, but all of the tests that exist pass.
> >> Feel free to fork it and send changes back if you find bugs.
> >>
> >> Hope that helps,
> >>
> >> Peter
> >>
> >> On 1 February 2012 08:17, Alison Callahan

> <alison....@gmail.com <mailto:alison....@gmail.com>>

> <mailto:topbrai...@googlegroups.com>

> >> > To unsubscribe from this group, send email to
> >> > topbraid-user...@googlegroups.com

> <mailto:topbraid-users%2Bunsu...@googlegroups.com>

> >> > For more options, visit this group at
> >> > http://groups.google.com/group/topbraid-users?hl=en
> >>
> >> --
> >> You received this message because you are subscribed to the Google
> >> Group "TopBraid Suite Users", the topics of which include Enterprise
> >> Vocabulary Network (EVN), TopBraid Composer,
> >> TopBraid Live, TopBraid Ensemble, SPARQLMotion and SPIN.
> >> To post to this group, send email to
> >> topbrai...@googlegroups.com

> <mailto:topbrai...@googlegroups.com>

> >> To unsubscribe from this group, send email to
> >> topbraid-user...@googlegroups.com

> <mailto:topbraid-users%2Bunsu...@googlegroups.com>

> >> For more options, visit this group at
> >> http://groups.google.com/group/topbraid-users?hl=en
> >
> >
>
>
> --
> You received this message because you are subscribed to the Google
> Group "TopBraid Suite Users", the topics of which include Enterprise
> Vocabulary Network (EVN), TopBraid Composer,
> TopBraid Live, TopBraid Ensemble, SPARQLMotion and SPIN.
> To post to this group, send email to
> topbrai...@googlegroups.com
> To unsubscribe from this group, send email to
> topbraid-user...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/topbraid-users?hl=en

--
-------------------------------------------------------------------
Jerven Bolleman Jerven....@isb-sib.ch
Swiss Institute of Bioinformatics Tel: +41 (0)22 379 58 85
CMU, rue Michel Servet 1 Fax: +41 (0)22 379 58 58
1211 Geneve 4,
Switzerland www.isb-sib.ch - www.uniprot.org
Follow us at https://twitter.com/#!/uniprot
-------------------------------------------------------------------

Alison Callahan

unread,

Feb 3, 2012, 2:02:58 PM2/3/12

to topbrai...@googlegroups.com

Hi Jerven,

Yes I have tried this approach, with the result that I get a java.lang.StackOverflowError when I try to inference using SPINInferences.run() with a Jena model created from an OWLIM repository as shown in the example at http://owlim.ontotext.com/display/OWLIMv43/OWLIM-SE+Usage+Scenarios#OWLIM-SEUsageScenarios-UsingOWLIMSEwithJena.

The relevant code and resulting stack trace are below. Any help is much appreciated!

Alison

****************************************

SPINModuleRegistry.get().init();

Model inputModel = ModelFactory.createDefaultModel();

inputModel.read("http://locahost/tmp/data.rdf");

OntModel spinModel = ModelFactory.createOntologyModel( OntModelSpec.OWL_MEM );

//add system triples

spinModel.add(SystemTriples.getVocabularyModel());

//add rules

spinModel.read("http://localhost/tmp/rules.spin.rdf");

//register locally defined functions

SPINModuleRegistry.get().registerAll(spinModel, null);

//connect to OWLIM repository

repositoryManager = new RemoteRepositoryManager("http://localhost:8080/openrdf-sesame");

try {

// Initialise the repository manager

repositoryManager.initialize();

// Get repository

repository = repositoryManager.getRepository("test");

// Open a connection to this repository

repositoryConnection = repository.getConnection();

//create the DatasetGraph instance

dataset = new SesameDataset(repositoryConnection);

//get model from dataset

owlimModel = ModelFactory.createModelForGraph(dataset.getDefaultGraph());

} catch (RepositoryException e) {

e.printStackTrace();

} catch (RepositoryConfigException e) {

e.printStackTrace();

}

//add owlim model

spinModel.addSubModel(owlimModel);

//add input model

spinModel.addSubModel(inputModel);

//create output model to contain inferred statements

Model outputModel = ModelFactory.createDefaultModel();

//add output model to SPIN model

spinModel.addSubModel(outputModel);

//run SPIN inferences

SPINInferences.run(spinModel, outputModel, null, null, false, null);

//get number of inferred statements

//System.out.println("Inferred triples: " + outputModel.size());

//get inferred statements

StmtIterator itr = outputModel.listStatements();

System.out.println("Iterating over statements");

while(itr.hasNext()){

Statement s = itr.next();

System.out.println(s.toString());

}

****************************************

And the stack trace:

java.lang.StackOverflowError

at com.hp.hpl.jena.graph.impl.GraphBase.contains(GraphBase.java:269)

at com.ontotext.jena.SesameDataset$AnyGraph$3.contains(Unknown Source)

at com.hp.hpl.jena.mem.GraphMem.graphBaseContains(GraphMem.java:75)