Examples of org.apache.clerezza.rdf.core.LiteralFactory

org.apache.clerezza.rdf.core.LiteralFactory
This class provides methods to convert java objects to typed literals and vice versa. While the default implementation will provide literal objects storing the data's lexical form in memory, other implementations may create literal optimized for processing within the store. Note: this class uses the notion of "Convertor" (in the Exception naming), but does not currently provide a mechanism to register such Convertors. An implementation is said to provide Convertors for the types it supports. @since 0.3 @author reto

     *
     * @return the URI of the new enhancement instance
     */
    protected static UriRef createEnhancement(MGraph metadata,
            EnhancementEngine engine, UriRef contentItemId){
        LiteralFactory literalFactory = LiteralFactory.getInstance();


        UriRef enhancement = new UriRef("urn:enhancement-"
                + EnhancementEngineHelper.randomUUID());
        //add the Enhancement Type
        metadata.add(new TripleImpl(enhancement, Properties.RDF_TYPE,
                TechnicalClasses.ENHANCER_ENHANCEMENT));
        //add the extracted from content item
        metadata.add(new TripleImpl(enhancement,
                Properties.ENHANCER_EXTRACTED_FROM, contentItemId));
        // creation date
        metadata.add(new TripleImpl(enhancement, Properties.DC_CREATED,
                literalFactory.createTypedLiteral(new Date())));


        // the engines that extracted the data
        // TODO: add some kind of versioning info for the extractor?
        // TODO: use a public dereferencing URI instead? that would allow for
        // explicit versioning too
        /* NOTE (Rupert Westenthaler 2010-05-26):
         * The Idea is to use the  ComponentContext in the activate() method of
         * an Enhancer to get the bundle name/version and use that as an
         * URI for the creator.
         * We would need to add getEnhancerID() method to the enhancer interface
         * to access this information
          */
        metadata.add(new TripleImpl(enhancement, Properties.DC_CREATOR,
                literalFactory.createTypedLiteral(engine.getClass().getName())));
        return enhancement;
    }

View Full Code Here

     * @param enhancement the enhancement
     * @param engine the engine
     */
    public static void addContributingEngine(MGraph metadata, UriRef enhancement,
                                             EnhancementEngine engine){
        LiteralFactory literalFactory = LiteralFactory.getInstance();
        // TODO: use a public dereferencing URI instead?
        metadata.add(new TripleImpl(enhancement, Properties.DC_CONTRIBUTOR,
            literalFactory.createTypedLiteral(engine.getClass().getName())));
        //set the modification date to the current date.
        set(metadata,enhancement,Properties.DC_MODIFIED,new Date(),literalFactory);
    }

View Full Code Here

     * @see EnhancementEngineHelper#createTextEnhancement(ContentItem, EnhancementEngine)
     */
    @Deprecated
    public static UriRef createNewExtraction(ContentItem ci,
            EnhancementEngine engine) {
        LiteralFactory literalFactory = LiteralFactory.getInstance();


        MGraph metadata = ci.getMetadata();
        UriRef extraction = new UriRef("urn:extraction-"
                + EnhancementEngineHelper.randomUUID());


        metadata.add(new TripleImpl(extraction, Properties.RDF_TYPE,
                TechnicalClasses.ENHANCER_EXTRACTION));


        // relate the extraction to the content item
        metadata.add(new TripleImpl(extraction,
                Properties.ENHANCER_RELATED_CONTENT_ITEM, new UriRef(ci.getUri().getUnicodeString())));


        // creation date
        metadata.add(new TripleImpl(extraction, Properties.DC_CREATED,
                literalFactory.createTypedLiteral(new Date())));


        // the engines that extracted the data
        // TODO: add some kind of versioning info for the extractor?
        // TODO: use a public dereferencing URI instead? that would allow for
        // explicit versioning too
        metadata.add(new TripleImpl(extraction, Properties.DC_CREATOR,
                literalFactory.createTypedLiteral(engine.getClass().getName())));


        return extraction;
    }

View Full Code Here

        confidence
    }


    private void initOccurrences() {
        MGraph graph = contentItem.getMetadata();
        LiteralFactory lf = LiteralFactory.getInstance();
        Map<UriRef,Collection<NonLiteral>> suggestionMap = new HashMap<UriRef,Collection<NonLiteral>>();
        // 1) get Entity Annotations
        Map<NonLiteral,Map<EAProps,Object>> entitySuggestionMap = new HashMap<NonLiteral,Map<EAProps,Object>>();
        Iterator<Triple> entityAnnotations = graph.filter(null, RDF.type, ENHANCER_ENTITYANNOTATION);
        while(entityAnnotations.hasNext()){

View Full Code Here

    /**
     * @return an RDF/JSON descriptions of places for the word map widget
     */
    public String getPlacesAsJSON() throws ParseException, UnsupportedEncodingException {
        MGraph g = new IndexedMGraph();
        LiteralFactory lf = LiteralFactory.getInstance();
        MGraph metadata = contentItem.getMetadata();
        for (EntityExtractionSummary p : getPlaceOccurrences()) {
            EntitySuggestion bestGuess = p.getBestGuess();
            if (bestGuess == null) {
                continue;
            }
            UriRef uri = new UriRef(bestGuess.getUri());
            Iterator<Triple> latitudes = metadata.filter(uri, GEO_LAT, null);
            if (latitudes.hasNext()) {
                g.add(latitudes.next());
            }
            Iterator<Triple> longitutes = metadata.filter(uri, GEO_LONG, null);
            if (longitutes.hasNext()) {
                g.add(longitutes.next());
                g.add(new TripleImpl(uri, Properties.RDFS_LABEL, lf.createTypedLiteral(bestGuess.getLabel())));
            }
        }
        ByteArrayOutputStream out = new ByteArrayOutputStream();
        serializer.serialize(out, g, SupportedFormat.RDF_JSON);

View Full Code Here

     *
     * @param occs a Collection of entity information
     * @param ci the content item
     */
    public void createEnhancements(Collection<CalaisEntityOccurrence> occs, ContentItem ci) {
        LiteralFactory literalFactory = LiteralFactory.getInstance();
        final Language language; // used for plain literals representing parts fo the content
        String langString = EnhancementEngineHelper.getLanguage(ci);
        if(langString != null && !langString.isEmpty()){
            language = new Language(langString);
        } else {
            language = null;
        }
        //TODO create TextEnhancement (form, start, end, type?) and EntityAnnotation (id, name, type)
        HashMap<Resource, UriRef> entityAnnotationMap = new HashMap<Resource, UriRef>();
        for (CalaisEntityOccurrence occ : occs) {
            UriRef textAnnotation = EnhancementEngineHelper.createTextEnhancement(
                    ci, this);
            MGraph model = ci.getMetadata();
            model.add(new TripleImpl(textAnnotation, DC_TYPE, occ.type));
            // for autotagger use the name instead of the matched term (that might be a pronoun!)
            if (onlyNERMode) {
                model.add(new TripleImpl(textAnnotation, ENHANCER_SELECTED_TEXT,new PlainLiteralImpl(occ.name,language)));
            }
            else {
                model.add(new TripleImpl(textAnnotation, ENHANCER_SELECTED_TEXT, new PlainLiteralImpl(occ.exact,language)));
            }
            model.add(new TripleImpl(textAnnotation, ENHANCER_START, literalFactory.createTypedLiteral(occ.offset)));
            model.add(new TripleImpl(textAnnotation, ENHANCER_END, literalFactory.createTypedLiteral(occ.offset + occ.length)));
            model.add(new TripleImpl(textAnnotation, ENHANCER_SELECTION_CONTEXT, new PlainLiteralImpl(occ.context,language)));
            //use the relevance as confidence
            if(occ.relevance != null && Double.valueOf(0).compareTo(occ.relevance) <= 0 ){
                //we do not know if the relevance is available (may be NULL)
                //or the relevance feature is activated (may be -1)
                model.add(new TripleImpl(textAnnotation, ENHANCER_CONFIDENCE, literalFactory.createTypedLiteral(occ.relevance)));
            }
            //create EntityAnnotation only once but add a reference to the textAnnotation
            if (entityAnnotationMap.containsKey(occ.id)) {
                model.add(new TripleImpl(entityAnnotationMap.get(occ.id), DC_RELATION, textAnnotation));
            } else {

View Full Code Here

        }
        Iterator<Span> spans = at.getEnclosed(EnumSet.of(SpanTypeEnum.Sentence,SpanTypeEnum.Chunk));
        Sentence context = null;
        MGraph metadata = ci.getMetadata();
        Language lang = new Language(language);
        LiteralFactory lf = LiteralFactory.getInstance();
        ci.getLock().writeLock().lock();
        try { //write TextAnnotations for Named Entities
            while(spans.hasNext()){
                Span span = spans.next();
                switch (span.getType()) {
                    case Sentence:
                        context = (Sentence)context;
                        break;
                    default:
                        Value<NerTag> nerAnno = span.getAnnotation(NER_ANNOTATION);
                        if(nerAnno != null){
                            UriRef ta = EnhancementEngineHelper.createTextEnhancement(ci, this);
                            //add span related data
                            metadata.add(new TripleImpl(ta, ENHANCER_SELECTED_TEXT, 
                                new PlainLiteralImpl(span.getSpan(), lang)));
                            metadata.add(new TripleImpl(ta, ENHANCER_START, 
                                lf.createTypedLiteral(span.getStart())));
                            metadata.add(new TripleImpl(ta, ENHANCER_END, 
                                lf.createTypedLiteral(span.getEnd())));
                            metadata.add(new TripleImpl(ta, ENHANCER_SELECTION_CONTEXT, 
                                new PlainLiteralImpl(context == null ?
                                        getDefaultSelectionContext(at.getSpan(), span.getSpan(), span.getStart()) :
                                            context.getSpan(), lang)));
                            //add the NER type
                            if(nerAnno.value().getType() != null){
                                metadata.add(new TripleImpl(ta,DC_TYPE,nerAnno.value().getType()));
                            }
                            if(nerAnno.probability() >= 0) {
                                metadata.add(new TripleImpl(ta, ENHANCER_CONFIDENCE, 
                                    lf.createTypedLiteral(nerAnno.probability())));
                            }
                        }
                        break;
                }
            }

View Full Code Here

        if(log.isDebugEnabled()){
            log.debug("findNamedEntities model={},  language={}, text=", 
                    new Object[]{ nameFinderModel, language, 
                                  StringUtils.abbreviate(at != null ? at.getSpan() : text, 100) });
        }
        LiteralFactory literalFactory = LiteralFactory.getInstance();
        MGraph g = ci.getMetadata();
        Map<String,List<NameOccurrence>> entityNames;
        if(at != null){
            entityNames = extractNameOccurrences(nameFinderModel, at, lang);
        } else {
            entityNames = extractNameOccurrences(nameFinderModel, text,lang);
        }
        //lock the ContentItem while writing the RDF data for found Named Entities
        ci.getLock().writeLock().lock();
        try {
            Map<String,UriRef> previousAnnotations = new LinkedHashMap<String,UriRef>();
            for (Map.Entry<String,List<NameOccurrence>> nameInContext : entityNames.entrySet()) {
    
                String name = nameInContext.getKey();
                List<NameOccurrence> occurrences = nameInContext.getValue();
    
                UriRef firstOccurrenceAnnotation = null;
    
                for (NameOccurrence occurrence : occurrences) {
                    UriRef textAnnotation = EnhancementEngineHelper.createTextEnhancement(ci, this);
                    g.add(new TripleImpl(textAnnotation, ENHANCER_SELECTED_TEXT, 
                        new PlainLiteralImpl(name, language)));
                    g.add(new TripleImpl(textAnnotation, ENHANCER_SELECTION_CONTEXT, 
                        new PlainLiteralImpl(occurrence.context, language)));
                    if(occurrence.type != null){
                        g.add(new TripleImpl(textAnnotation, DC_TYPE, occurrence.type));
                    }
                    g.add(new TripleImpl(textAnnotation, ENHANCER_CONFIDENCE, literalFactory
                            .createTypedLiteral(occurrence.confidence)));
                    if (occurrence.start != null && occurrence.end != null) {
                        g.add(new TripleImpl(textAnnotation, ENHANCER_START, literalFactory
                                .createTypedLiteral(occurrence.start)));
                        g.add(new TripleImpl(textAnnotation, ENHANCER_END, literalFactory
                                .createTypedLiteral(occurrence.end)));
                    }
    
                    // add the subsumption relationship among occurrences of the same
                    // name

View Full Code Here

            }
        } else { // null indicates to use the Entityhub to lookup Entities
            site = null;
        }
        MGraph graph = ci.getMetadata();
        LiteralFactory literalFactory = LiteralFactory.getInstance();
        // Retrieve the existing text annotations (requires read lock)
        Map<NamedEntity,List<UriRef>> textAnnotations = new HashMap<NamedEntity,List<UriRef>>();
        // the language extracted for the parsed content or NULL if not
        // available
        String contentLangauge;

View Full Code Here


    @Override
    public void computeEnhancements(ContentItem ci) throws EngineException {
        UriRef contentItemId = ci.getUri();
        MGraph 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

View Full Code Here

0 1 2 3 4 5 6

TOP

Related Classes of org.apache.clerezza.rdf.core.LiteralFactory

org.apache.clerezza.platform.content.WebDavUtils

org.apache.clerezza.rdf.jena.tdb.storage.TdbMGraphTest

org.apache.stanbol.cmsadapter.cmis.mapping.CMISRDFMapper

org.apache.stanbol.cmsadapter.core.mapping.DefaultRDFBridgeImpl

org.apache.stanbol.cmsadapter.core.mapping.RDFBridgeHelper

org.apache.stanbol.cmsadapter.jcr.mapping.JCRRDFMapper

org.apache.stanbol.commons.indexedgraph.IndexedGraphTest

org.apache.stanbol.contenthub.web.resources.ContentItemResource

org.apache.stanbol.enhancer.engine.topic.TopicClassificationEngine

org.apache.stanbol.enhancer.engines.celi.lemmatizer.impl.CeliLemmatizerEnhancementEngine

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.