Examples with BlankNodeOrIRI org.apache.clerezza.commons.rdf.BlankNodeOrIRI used on opensource projects

Example 41 with BlankNodeOrIRI

use of org.apache.clerezza.commons.rdf.BlankNodeOrIRI in project stanbol by apache.

the class LocationEnhancementEngine method computeEnhancements.

@Override
public void computeEnhancements(ContentItem ci) throws EngineException {
    IRI contentItemId = ci.getUri();
    Graph graph = ci.getMetadata();
    LiteralFactory literalFactory = LiteralFactory.getInstance();
    //get all the textAnnotations
    /*
         * this Map holds the name as key and all the text annotations of
         * dc:type dbpedia:Place that select this name as value
         * this map is used to avoid multiple lookups for text annotations
         * selecting the same name.
         */
    Map<String, Collection<BlankNodeOrIRI>> name2placeEnhancementMap = new HashMap<String, Collection<BlankNodeOrIRI>>();
    Iterator<Triple> iterator = graph.filter(null, DC_TYPE, DBPEDIA_PLACE);
    while (iterator.hasNext()) {
        //the enhancement annotating an place
        BlankNodeOrIRI placeEnhancement = iterator.next().getSubject();
        //this can still be an TextAnnotation of an EntityAnnotation
        //so we need to filter TextAnnotation
        Triple isTextAnnotation = new TripleImpl(placeEnhancement, RDF_TYPE, ENHANCER_TEXTANNOTATION);
        if (graph.contains(isTextAnnotation)) {
            //now get the name
            String name = EnhancementEngineHelper.getString(graph, placeEnhancement, ENHANCER_SELECTED_TEXT);
            if (name == null) {
                log.warn("Unable to process TextAnnotation " + placeEnhancement + " because property" + ENHANCER_SELECTED_TEXT + " is not present");
            } else {
                Collection<BlankNodeOrIRI> placeEnhancements = name2placeEnhancementMap.get(name);
                if (placeEnhancements == null) {
                    placeEnhancements = new ArrayList<BlankNodeOrIRI>();
                    name2placeEnhancementMap.put(name, placeEnhancements);
                }
                placeEnhancements.add(placeEnhancement);
            }
        } else {
        //TODO: if we also ant to process EntityAnnotations with the dc:type dbpedia:Place
        //      than we need to parse the name based on the enhancer:entity-name property
        }
    }
    //Now we do have all the names we need to lookup
    Map<SearchRequestPropertyEnum, Collection<String>> requestParams = new EnumMap<SearchRequestPropertyEnum, Collection<String>>(SearchRequestPropertyEnum.class);
    if (getMaxLocationEnhancements() != null) {
        requestParams.put(SearchRequestPropertyEnum.maxRows, Collections.singleton(getMaxLocationEnhancements().toString()));
    }
    for (Map.Entry<String, Collection<BlankNodeOrIRI>> entry : name2placeEnhancementMap.entrySet()) {
        List<Toponym> results;
        try {
            requestParams.put(SearchRequestPropertyEnum.name, Collections.singleton(entry.getKey()));
            results = geonamesService.searchToponyms(requestParams);
        } catch (Exception e) {
            /*
                     * TODO: Review if it makes sense to catch here for each name, or
                     * to catch the whole loop.
                     * This depends if single requests can result in Exceptions
                     * (e.g. because of encoding problems) or if usually Exceptions
                     * are thrown because of general things like connection issues
                     * or service unavailability.
                     */
            throw new EngineException(this, ci, e);
        }
        if (results != null) {
            Double maxScore = results.isEmpty() ? null : results.get(0).getScore();
            for (Toponym result : results) {
                log.debug("process result {} {}", result.getGeoNameId(), result.getName());
                Double score = getToponymScore(result, maxScore);
                log.debug("  > score {}", score);
                if (score != null) {
                    if (score < minScore) {
                        //if score is lower than the under bound, than stop
                        break;
                    }
                } else {
                    log.warn("NULL returned as Score for " + result.getGeoNameId() + " " + result.getName());
                /*
                         * NOTE: If score is not present all suggestions are
                         * added as enhancements to the metadata of the content
                         * item.
                         */
                }
                //write the enhancement!
                BlankNodeOrIRI locationEnhancement = writeEntityEnhancement(contentItemId, graph, literalFactory, result, entry.getValue(), null, score);
                log.debug("  > {}  >= {}", score, minHierarchyScore);
                if (score != null && score >= minHierarchyScore) {
                    log.debug("  > getHierarchy for {} {}", result.getGeoNameId(), result.getName());
                    //get the hierarchy
                    try {
                        Iterator<Toponym> hierarchy = getHierarchy(result).iterator();
                        for (int level = 0; hierarchy.hasNext(); level++) {
                            Toponym hierarchyEntry = hierarchy.next();
                            //  maybe add an configuration
                            if (level == 0) {
                                //Mother earth -> ignore
                                continue;
                            }
                            //write it as dependent to the locationEnhancement
                            if (result.getGeoNameId() != hierarchyEntry.getGeoNameId()) {
                                //TODO: add additional checks based on possible
                                //      configuration here!
                                log.debug("    - write hierarchy {} {}", hierarchyEntry.getGeoNameId(), hierarchyEntry.getName());
                                /*
                                     * The hierarchy service dose not provide a score, because it would be 1.0
                                     * so we need to set the score to this value.
                                     * Currently is is set to the value of the suggested entry
                                     */
                                writeEntityEnhancement(contentItemId, graph, literalFactory, hierarchyEntry, null, Collections.singletonList(locationEnhancement), 1.0);
                            }
                        }
                    } catch (Exception e) {
                        log.warn("Unable to get Hierarchy for " + result.getGeoNameId() + " " + result.getName(), e);
                    }
                }
            }
        }
    }
}

Example 42 with BlankNodeOrIRI

use of org.apache.clerezza.commons.rdf.BlankNodeOrIRI in project stanbol by apache.

the class ClerezzaOntologyProvider method toOWLOntology.

/**
     * 
     * @param graphName
     * @param forceMerge
     *            if set to false, the selected import management policy will be applied.
     * @return
     * @throws OWLOntologyCreationException
     */
protected OWLOntology toOWLOntology(IRI graphName, boolean forceMerge) throws OWLOntologyCreationException {
    log.debug("Exporting graph to OWLOntology");
    log.debug(" -- ImmutableGraph name : {}", graphName);
    OWLOntologyManager mgr = OWLManager.createOWLOntologyManager();
    // Never try to import
    mgr.addIRIMapper(new PhonyIRIMapper(Collections.<org.semanticweb.owlapi.model.IRI>emptySet()));
    Set<OWLOntologyID> loaded = new HashSet<OWLOntologyID>();
    Graph graph = store.getGraph(graphName);
    IRI ontologyId = null;
    // Get the id of this ontology.
    Iterator<Triple> itt = graph.filter(null, RDF.type, OWL.Ontology);
    if (itt.hasNext()) {
        BlankNodeOrIRI nl = itt.next().getSubject();
        if (nl instanceof IRI)
            ontologyId = (IRI) nl;
    }
    List<OWLOntologyID> revImps = new Stack<OWLOntologyID>();
    List<OWLOntologyID> lvl1 = new Stack<OWLOntologyID>();
    fillImportsReverse(keymap.getReverseMapping(graphName), revImps, lvl1);
    // If not set to merge (either by policy of by force), adopt the set import policy.
    if (!forceMerge && !ImportManagementPolicy.MERGE.equals(getImportManagementPolicy())) {
        OWLOntology o = OWLAPIToClerezzaConverter.clerezzaGraphToOWLOntology(graph, mgr);
        // TODO make it not flat.
        // Examining the reverse imports stack will flatten all imports.
        List<OWLOntologyChange> changes = new ArrayList<OWLOntologyChange>();
        OWLDataFactory df = OWLManager.getOWLDataFactory();
        List<OWLOntologyID> listToUse;
        switch(getImportManagementPolicy()) {
            case FLATTEN:
                listToUse = revImps;
                break;
            case PRESERVE:
                listToUse = lvl1;
                break;
            default:
                listToUse = lvl1;
                break;
        }
        for (OWLOntologyID ref : listToUse) if (!loaded.contains(ref) && !ref.equals(keymap.getReverseMapping(graphName))) {
            changes.add(new AddImport(o, df.getOWLImportsDeclaration(ref.getOntologyIRI())));
            loaded.add(ref);
        }
        o.getOWLOntologyManager().applyChanges(changes);
        return o;
    } else {
        // If there is just the root ontology, convert it straight away.
        if (revImps.size() == 1 && revImps.contains(graphName)) {
            OWLOntology o = OWLAPIToClerezzaConverter.clerezzaGraphToOWLOntology(graph, mgr);
            return o;
        }
        // FIXME when there's more than one ontology, this way of merging them seems inefficient...
        Graph tempGraph = new IndexedGraph();
        // The set of triples that will be excluded from the merge
        Set<Triple> exclusions = new HashSet<Triple>();
        // Examine all reverse imports
        for (OWLOntologyID ref : revImps) if (!loaded.contains(ref)) {
            // Get the triples
            Graph imported = // store.getTriples(ref);
            getStoredOntology(getKey(ref), Graph.class, false);
            // For each owl:Ontology
            Iterator<Triple> remove = imported.filter(null, RDF.type, OWL.Ontology);
            while (remove.hasNext()) {
                BlankNodeOrIRI subj = remove.next().getSubject();
                /*
                         * If it's not the root ontology, trash all its triples. If the root ontology is
                         * anonymous, all ontology annotations are to be trashed without distinction.
                         */
                if (ontologyId == null || !subj.equals(ontologyId)) {
                    Iterator<Triple> it = imported.filter(subj, null, null);
                    while (it.hasNext()) {
                        Triple t = it.next();
                        exclusions.add(t);
                    }
                }
            }
            Iterator<Triple> it = imported.iterator();
            while (it.hasNext()) {
                Triple t = it.next();
                if (!exclusions.contains(t))
                    tempGraph.add(t);
            }
            loaded.add(ref);
        }
        // online.
        return OWLAPIToClerezzaConverter.clerezzaGraphToOWLOntology(tempGraph, mgr);
    }
}

Example 43 with BlankNodeOrIRI

use of org.apache.clerezza.commons.rdf.BlankNodeOrIRI in project stanbol by apache.

the class ClerezzaOntologyProvider method getOntologyNetworkConfiguration.

public OntologyNetworkConfiguration getOntologyNetworkConfiguration() {
    Map<String, Collection<OWLOntologyID>> coreOntologies = new HashMap<String, Collection<OWLOntologyID>>(), customOntologies = new HashMap<String, Collection<OWLOntologyID>>();
    Map<String, Collection<String>> attachedScopes = new HashMap<String, Collection<String>>();
    final Graph meta = store.getGraph(new IRI(metaGraphId));
    // Scopes first
    for (Iterator<Triple> it = meta.filter(null, RDF.type, SCOPE_URIREF); it.hasNext(); ) {
        // for each
        // scope
        Triple ta = it.next();
        BlankNodeOrIRI sub = ta.getSubject();
        if (sub instanceof IRI) {
            String s = ((IRI) sub).getUnicodeString(), prefix = _NS_STANBOL_INTERNAL + Scope.shortName + "/";
            if (s.startsWith(prefix)) {
                String scopeId = s.substring(prefix.length());
                log.info("Rebuilding scope \"{}\".", scopeId);
                coreOntologies.put(scopeId, new TreeSet<OWLOntologyID>());
                customOntologies.put(scopeId, new TreeSet<OWLOntologyID>());
                IRI core_ur = null, custom_ur = null;
                RDFTerm r;
                // Check core space
                Iterator<Triple> it2 = meta.filter(sub, HAS_SPACE_CORE_URIREF, null);
                if (it2.hasNext()) {
                    r = it2.next().getObject();
                    if (r instanceof IRI)
                        core_ur = (IRI) r;
                } else {
                    it2 = meta.filter(null, IS_SPACE_CORE_OF_URIREF, sub);
                    if (it2.hasNext()) {
                        r = it2.next().getSubject();
                        if (r instanceof IRI)
                            core_ur = (IRI) r;
                    }
                }
                // Check custom space
                it2 = meta.filter(sub, HAS_SPACE_CUSTOM_URIREF, null);
                if (it2.hasNext()) {
                    r = it2.next().getObject();
                    if (r instanceof IRI)
                        custom_ur = (IRI) r;
                } else {
                    it2 = meta.filter(null, IS_SPACE_CUSTOM_OF_URIREF, sub);
                    if (it2.hasNext()) {
                        r = it2.next().getSubject();
                        if (r instanceof IRI)
                            custom_ur = (IRI) r;
                    }
                }
                // retrieve the ontologies
                if (core_ur != null) {
                    for (it2 = meta.filter(core_ur, null, null); it2.hasNext(); ) {
                        Triple t = it2.next();
                        IRI predicate = t.getPredicate();
                        if (predicate.equals(MANAGES_URIREF)) {
                            if (t.getObject() instanceof IRI)
                                coreOntologies.get(scopeId).add(// FIXME must be very
                                keymap.buildPublicKey((IRI) t.getObject()));
                        }
                    }
                    for (it2 = meta.filter(null, null, core_ur); it2.hasNext(); ) {
                        Triple t = it2.next();
                        IRI predicate = t.getPredicate();
                        if (predicate.equals(IS_MANAGED_BY_URIREF)) {
                            if (t.getSubject() instanceof IRI)
                                coreOntologies.get(scopeId).add(// FIXME must be very
                                keymap.buildPublicKey((IRI) t.getSubject()));
                        }
                    }
                }
                if (custom_ur != null) {
                    for (it2 = meta.filter(custom_ur, null, null); it2.hasNext(); ) {
                        Triple t = it2.next();
                        IRI predicate = t.getPredicate();
                        if (predicate.equals(MANAGES_URIREF)) {
                            if (t.getObject() instanceof IRI)
                                customOntologies.get(scopeId).add(// FIXME must be very
                                keymap.buildPublicKey((IRI) t.getObject()));
                        }
                    }
                    for (it2 = meta.filter(null, null, custom_ur); it2.hasNext(); ) {
                        Triple t = it2.next();
                        IRI predicate = t.getPredicate();
                        if (predicate.equals(IS_MANAGED_BY_URIREF)) {
                            if (t.getSubject() instanceof IRI)
                                customOntologies.get(scopeId).add(// FIXME must be very
                                keymap.buildPublicKey((IRI) t.getSubject()));
                        }
                    }
                }
            }
        }
    }
    // Sessions next
    Map<String, Collection<OWLOntologyID>> sessionOntologies = new HashMap<String, Collection<OWLOntologyID>>();
    for (Iterator<Triple> it = meta.filter(null, RDF.type, SESSION_URIREF); it.hasNext(); ) {
        // for each
        // scope
        Triple ta = it.next();
        BlankNodeOrIRI sub = ta.getSubject();
        if (sub instanceof IRI) {
            IRI ses_ur = (IRI) sub;
            String s = ((IRI) sub).getUnicodeString();
            String prefix = _NS_STANBOL_INTERNAL + Session.shortName + "/";
            if (s.startsWith(prefix)) {
                String sessionId = s.substring(prefix.length());
                log.info("Rebuilding session \"{}\".", sessionId);
                sessionOntologies.put(sessionId, new TreeSet<OWLOntologyID>());
                attachedScopes.put(sessionId, new TreeSet<String>());
                // retrieve the ontologies
                if (ses_ur != null) {
                    for (Iterator<Triple> it2 = meta.filter(ses_ur, MANAGES_URIREF, null); it2.hasNext(); ) {
                        RDFTerm obj = it2.next().getObject();
                        if (obj instanceof IRI)
                            sessionOntologies.get(sessionId).add(// FIXME must be very temporary!
                            keymap.buildPublicKey((IRI) obj));
                    }
                    for (Iterator<Triple> it2 = meta.filter(null, IS_MANAGED_BY_URIREF, ses_ur); it2.hasNext(); ) {
                        RDFTerm subj = it2.next().getSubject();
                        if (subj instanceof IRI)
                            sessionOntologies.get(sessionId).add(// FIXME must be very temporary!
                            keymap.buildPublicKey((IRI) subj));
                    }
                    for (Iterator<Triple> it2 = meta.filter(null, APPENDED_TO_URIREF, ses_ur); it2.hasNext(); ) {
                        RDFTerm subj = it2.next().getSubject();
                        if (subj instanceof IRI) {
                            String s1 = ((IRI) subj).getUnicodeString();
                            String prefix1 = _NS_STANBOL_INTERNAL + Scope.shortName + "/";
                            if (s1.startsWith(prefix1)) {
                                String scopeId = s1.substring(prefix1.length());
                                attachedScopes.get(sessionId).add(scopeId);
                            }
                        }
                    }
                    for (Iterator<Triple> it2 = meta.filter(ses_ur, HAS_APPENDED_URIREF, null); it2.hasNext(); ) {
                        RDFTerm obj = it2.next().getObject();
                        if (obj instanceof IRI) {
                            String s1 = ((IRI) obj).getUnicodeString();
                            String prefix1 = _NS_STANBOL_INTERNAL + Scope.shortName + "/";
                            if (s1.startsWith(prefix1)) {
                                String scopeId = s1.substring(prefix1.length());
                                attachedScopes.get(sessionId).add(scopeId);
                            }
                        }
                    }
                }
            }
        }
    }
    return new OntologyNetworkConfiguration(coreOntologies, customOntologies, sessionOntologies, attachedScopes);
}

Example 44 with BlankNodeOrIRI

use of org.apache.clerezza.commons.rdf.BlankNodeOrIRI in project stanbol by apache.

the class ResourceMapping method apply.

@Override
public boolean apply(Graph graph, BlankNodeOrIRI subject, Metadata metadata) {
    boolean added = false;
    BlankNodeOrIRI s = new BlankNode();
    mappingLogger.log(subject, ontProperty, null, s);
    if (!required.isEmpty()) {
        Graph g = new SimpleGraph();
        for (Mapping m : required) {
            if (!m.apply(g, s, metadata)) {
                return false;
            }
        }
        graph.addAll(g);
        added = true;
    }
    for (Mapping m : optional) {
        if (m.apply(graph, s, metadata)) {
            added = true;
        }
    }
    if (added) {
        for (Mapping m : additional) {
            m.apply(graph, s, metadata);
        }
        graph.add(new TripleImpl(subject, ontProperty, s));
    }
    return added;
}

Example 45 with BlankNodeOrIRI

use of org.apache.clerezza.commons.rdf.BlankNodeOrIRI in project stanbol by apache.

the class TikaEngineTest method testGEOMetadata.

@Test
public void testGEOMetadata() throws EngineException, IOException, ParseException {
    log.info(">>> testGEOMetadata <<<");
    //first validate Media RDFTerm Ontology
    IRI hasLocation = new IRI(NamespaceEnum.media + "hasLocation");
    IRI locationLatitude = new IRI(NamespaceEnum.media + "locationLatitude");
    IRI locationLongitude = new IRI(NamespaceEnum.media + "locationLongitude");
    //IRI locationAltitude = new IRI(NamespaceEnum.media+"locationAltitude");
    //"video/x-ms-asf");
    ContentItem ci = createContentItem("testJPEG_GEO.jpg", OCTET_STREAM.toString());
    assertFalse(engine.canEnhance(ci) == CANNOT_ENHANCE);
    engine.computeEnhancements(ci);
    Iterator<Triple> it = ci.getMetadata().filter(ci.getUri(), hasLocation, null);
    assertTrue(it.hasNext());
    RDFTerm r = it.next().getObject();
    assertFalse(it.hasNext());
    assertTrue(r instanceof BlankNodeOrIRI);
    BlankNodeOrIRI location = verifyBlankNodeOrIRI(ci, hasLocation);
    //lat
    verifyValue(ci, location, locationLatitude, XSD.double_, "12.54321");
    //long
    verifyValue(ci, location, locationLongitude, XSD.double_, "-54.1234");
    //second the GEO ont
    IRI lat = new IRI(NamespaceEnum.geo + "lat");
    IRI lon = new IRI(NamespaceEnum.geo + "long");
    //lat
    verifyValue(ci, lat, XSD.double_, "12.54321");
    //long
    verifyValue(ci, lon, XSD.double_, "-54.1234");
}