Examples with MentionAnnotator org.cogcomp.md.MentionAnnotator used on opensource projects

Example 1 with MentionAnnotator

use of org.cogcomp.md.MentionAnnotator in project cogcomp-nlp by CogComp.

the class SemEvalMentionReader method initExternalTools.

public void initExternalTools() {
    try {
        _posAnnotator = new POSAnnotator();
        Datastore ds = new Datastore(new ResourceConfigurator().getDefaultConfig());
        File gazetteersResource = ds.getDirectory("org.cogcomp.gazetteers", "gazetteers", 1.3, false);
        _gazetteers = (FlatGazetteers) GazetteersFactory.get(5, gazetteersResource.getPath() + File.separator + "gazetteers", true, Language.English);
        WordNetManager.loadConfigAsClasspathResource(true);
        _wordnet = WordNetManager.getInstance();
        __chunker = new ChunkerAnnotator(true);
        __chunker.initialize(new ChunkerConfigurator().getDefaultConfig());
        Properties stanfordProps = new Properties();
        stanfordProps.put("annotators", "pos, parse");
        stanfordProps.put("parse.originalDependencies", true);
        stanfordProps.put("parse.maxlen", Stanford331Configurator.STFRD_MAX_SENTENCE_LENGTH);
        stanfordProps.put("parse.maxtime", Stanford331Configurator.STFRD_TIME_PER_SENTENCE);
        POSTaggerAnnotator posAnnotator = new POSTaggerAnnotator("pos", stanfordProps);
        ParserAnnotator parseAnnotator = new ParserAnnotator("parse", stanfordProps);
        __stanfordDep = new StanfordDepHandler(posAnnotator, parseAnnotator);
        __mentionAnnotator = new MentionAnnotator("ACE_TYPE");
    } catch (Exception e) {
        e.printStackTrace();
    }
}

Example 2 with MentionAnnotator

use of org.cogcomp.md.MentionAnnotator in project cogcomp-nlp by CogComp.

the class RelationExtractionTest method testAnnotator.

@Test
public void testAnnotator() {
    File modelDir = null;
    try {
        Datastore ds = new Datastore(new ResourceConfigurator().getDefaultConfig());
        modelDir = ds.getDirectory("org.cogcomp.re", "ACE_TEST_DOCS", 1.1, false);
    } catch (Exception e) {
        e.printStackTrace();
    }
    try {
        ACEReaderWithTrueCaseFixer aceReader = new ACEReaderWithTrueCaseFixer(modelDir.getAbsolutePath() + File.separator + "ACE_TEST_DOCS", false);
        POSAnnotator pos_annotator = new POSAnnotator();
        ChunkerAnnotator chunker = new ChunkerAnnotator(true);
        chunker.initialize(new ChunkerConfigurator().getDefaultConfig());
        Properties stanfordProps = new Properties();
        stanfordProps.put("annotators", "pos, parse");
        stanfordProps.put("parse.originalDependencies", true);
        stanfordProps.put("parse.maxlen", Stanford331Configurator.STFRD_MAX_SENTENCE_LENGTH);
        stanfordProps.put("parse.maxtime", Stanford331Configurator.STFRD_TIME_PER_SENTENCE);
        POSTaggerAnnotator posAnnotator = new POSTaggerAnnotator("pos", stanfordProps);
        ParserAnnotator parseAnnotator = new ParserAnnotator("parse", stanfordProps);
        StanfordDepHandler stanfordDepHandler = new StanfordDepHandler(posAnnotator, parseAnnotator);
        MentionAnnotator mentionAnnotator = new MentionAnnotator("ACE_TYPE");
        RelationAnnotator relationAnnotator = new RelationAnnotator();
        for (TextAnnotation ta : aceReader) {
            ta.addView(pos_annotator);
            chunker.addView(ta);
            stanfordDepHandler.addView(ta);
            mentionAnnotator.addView(ta);
            relationAnnotator.addView(ta);
            View mentionView = ta.getView(ViewNames.MENTION);
            assertTrue(mentionView.getConstituents().size() > 0);
            View relationView = ta.getView(ViewNames.RELATION);
            assertTrue(relationView.getRelations().size() > 0);
        }
    } catch (Exception e) {
        e.printStackTrace();
    }
}

Example 3 with MentionAnnotator

use of org.cogcomp.md.MentionAnnotator in project cogcomp-nlp by CogComp.

the class ACERelationTester method testRandomText.

public static void testRandomText(String text) {
    String corpus = "";
    String textId = "";
    TextAnnotationBuilder stab = new TokenizerTextAnnotationBuilder(new StatefulTokenizer());
    TextAnnotation ta = stab.createTextAnnotation(corpus, textId, text);
    try {
        POSAnnotator pos_annotator = new POSAnnotator();
        ChunkerAnnotator chunker = new ChunkerAnnotator(true);
        chunker.initialize(new ChunkerConfigurator().getDefaultConfig());
        Properties stanfordProps = new Properties();
        stanfordProps.put("annotators", "pos, parse");
        stanfordProps.put("parse.originalDependencies", true);
        stanfordProps.put("parse.maxlen", Stanford331Configurator.STFRD_MAX_SENTENCE_LENGTH);
        stanfordProps.put("parse.maxtime", Stanford331Configurator.STFRD_TIME_PER_SENTENCE);
        POSTaggerAnnotator posAnnotator = new POSTaggerAnnotator("pos", stanfordProps);
        ParserAnnotator parseAnnotator = new ParserAnnotator("parse", stanfordProps);
        StanfordDepHandler stanfordDepHandler = new StanfordDepHandler(posAnnotator, parseAnnotator);
        MentionAnnotator mentionAnnotator = new MentionAnnotator("ACE_TYPE");
        RelationAnnotator relationAnnotator = new RelationAnnotator();
        ta.addView(pos_annotator);
        stanfordDepHandler.addView(ta);
        chunker.addView(ta);
        mentionAnnotator.addView(ta);
        relationAnnotator.addView(ta);
        for (Relation r : ta.getView(ViewNames.RELATION).getRelations()) {
            IOHelper.printRelation(r);
        }
    } catch (Exception e) {
        e.printStackTrace();
    }
}

Example 4 with MentionAnnotator

use of org.cogcomp.md.MentionAnnotator in project cogcomp-nlp by CogComp.

the class ACERelationTester method testAnnotator.

public static void testAnnotator() {
    int total_correct = 0;
    int total_labeled = 0;
    int total_predicted = 0;
    int total_coarse_correct = 0;
    try {
        POSAnnotator pos_annotator = new POSAnnotator();
        ChunkerAnnotator chunker = new ChunkerAnnotator(true);
        chunker.initialize(new ChunkerConfigurator().getDefaultConfig());
        Properties stanfordProps = new Properties();
        stanfordProps.put("annotators", "pos, parse");
        stanfordProps.put("parse.originalDependencies", true);
        stanfordProps.put("parse.maxlen", Stanford331Configurator.STFRD_MAX_SENTENCE_LENGTH);
        stanfordProps.put("parse.maxtime", Stanford331Configurator.STFRD_TIME_PER_SENTENCE);
        POSTaggerAnnotator posAnnotator = new POSTaggerAnnotator("pos", stanfordProps);
        ParserAnnotator parseAnnotator = new ParserAnnotator("parse", stanfordProps);
        StanfordDepHandler stanfordDepHandler = new StanfordDepHandler(posAnnotator, parseAnnotator);
        ACEReader aceReader = new ACEReader("data/partition_with_dev/dev", false);
        MentionAnnotator mentionAnnotator = new MentionAnnotator("ACE_TYPE");
        RelationAnnotator relationAnnotator = new RelationAnnotator();
        for (TextAnnotation ta : aceReader) {
            ta.addView(pos_annotator);
            stanfordDepHandler.addView(ta);
            chunker.addView(ta);
            mentionAnnotator.addView(ta);
            relationAnnotator.addView(ta);
            total_labeled += ta.getView(ViewNames.MENTION_ACE).getRelations().size();
            total_predicted += ta.getView(ViewNames.RELATION).getRelations().size();
            for (Relation pr : ta.getView(ViewNames.RELATION).getRelations()) {
                for (Relation gr : ta.getView(ViewNames.MENTION_ACE).getRelations()) {
                    Constituent prSourceHead = RelationFeatureExtractor.getEntityHeadForConstituent(pr.getSource(), ta, "");
                    Constituent grSourceHead = RelationFeatureExtractor.getEntityHeadForConstituent(gr.getSource(), ta, "");
                    Constituent prTargetHead = RelationFeatureExtractor.getEntityHeadForConstituent(pr.getTarget(), ta, "");
                    Constituent grTargetHead = RelationFeatureExtractor.getEntityHeadForConstituent(gr.getTarget(), ta, "");
                    if (prSourceHead.getStartSpan() == grSourceHead.getStartSpan() && prSourceHead.getEndSpan() == grSourceHead.getEndSpan() && prTargetHead.getEndSpan() == grTargetHead.getEndSpan() && prTargetHead.getStartSpan() == grTargetHead.getStartSpan()) {
                        if (pr.getAttribute("RelationType").equals(gr.getAttribute("RelationType"))) {
                            total_coarse_correct++;
                        }
                        if (pr.getAttribute("RelationSubtype").equals(gr.getAttribute("RelationSubtype"))) {
                            total_correct++;
                        }
                    }
                }
            }
        }
    } catch (Exception e) {
        e.printStackTrace();
    }
    System.out.println("Total labeled: " + total_labeled);
    System.out.println("Total predicted: " + total_predicted);
    System.out.println("Total correct: " + total_correct);
    System.out.println("Total coarse correct: " + total_coarse_correct);
    double p = (double) total_correct * 100.0 / (double) total_predicted;
    double r = (double) total_correct * 100.0 / (double) total_labeled;
    double f = 2 * p * r / (p + r);
    System.out.println("Precision: " + p);
    System.out.println("Recall: " + r);
    System.out.println("Fine Type F1: " + f);
    System.out.println("Coarse Type F1: " + f * (double) total_coarse_correct / (double) total_correct);
}

Example 5 with MentionAnnotator

use of org.cogcomp.md.MentionAnnotator in project cogcomp-nlp by CogComp.

the class PipelineFactory method buildAnnotators.

/**
 * instantiate a set of annotators for use in an AnnotatorService object by default, will use
 * lazy initialization where possible -- change this behavior with the
 * {@link PipelineConfigurator#USE_LAZY_INITIALIZATION} property.
 *
 * @param nonDefaultRm ResourceManager with all non-default values for Annotators
 * @return a Map from annotator view name to annotator
 */
private static Map<String, Annotator> buildAnnotators(ResourceManager nonDefaultRm) throws IOException {
    ResourceManager rm = new PipelineConfigurator().getConfig(new Stanford331Configurator().getConfig(nonDefaultRm));
    String timePerSentence = rm.getString(Stanford331Configurator.STFRD_TIME_PER_SENTENCE);
    String maxParseSentenceLength = rm.getString(Stanford331Configurator.STFRD_MAX_SENTENCE_LENGTH);
    boolean useLazyInitialization = rm.getBoolean(PipelineConfigurator.USE_LAZY_INITIALIZATION.key, PipelineConfigurator.TRUE);
    Map<String, Annotator> viewGenerators = new HashMap<>();
    if (rm.getBoolean(PipelineConfigurator.USE_POS)) {
        POSAnnotator pos = new POSAnnotator();
        viewGenerators.put(pos.getViewName(), pos);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_LEMMA)) {
        IllinoisLemmatizer lem = new IllinoisLemmatizer(rm);
        viewGenerators.put(lem.getViewName(), lem);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_SHALLOW_PARSE)) {
        viewGenerators.put(ViewNames.SHALLOW_PARSE, new ChunkerAnnotator());
    }
    if (rm.getBoolean(PipelineConfigurator.USE_NER_CONLL)) {
        NERAnnotator nerConll = NerAnnotatorManager.buildNerAnnotator(rm, ViewNames.NER_CONLL);
        viewGenerators.put(nerConll.getViewName(), nerConll);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_NER_ONTONOTES)) {
        NERAnnotator nerOntonotes = NerAnnotatorManager.buildNerAnnotator(rm, ViewNames.NER_ONTONOTES);
        viewGenerators.put(nerOntonotes.getViewName(), nerOntonotes);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_DEP)) {
        DepAnnotator dep = new DepAnnotator();
        viewGenerators.put(dep.getViewName(), dep);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_STANFORD_DEP) || rm.getBoolean(PipelineConfigurator.USE_STANFORD_PARSE)) {
        Properties stanfordProps = new Properties();
        stanfordProps.put("annotators", "pos, parse");
        stanfordProps.put("parse.originalDependencies", true);
        stanfordProps.put("parse.maxlen", maxParseSentenceLength);
        // per sentence? could be per
        stanfordProps.put("parse.maxtime", timePerSentence);
        // document but no idea from
        // stanford javadoc
        POSTaggerAnnotator posAnnotator = new POSTaggerAnnotator("pos", stanfordProps);
        ParserAnnotator parseAnnotator = new ParserAnnotator("parse", stanfordProps);
        int maxLength = Integer.parseInt(maxParseSentenceLength);
        boolean throwExceptionOnSentenceLengthCheck = rm.getBoolean(Stanford331Configurator.THROW_EXCEPTION_ON_FAILED_LENGTH_CHECK.key);
        if (rm.getBoolean(PipelineConfigurator.USE_STANFORD_DEP)) {
            StanfordDepHandler depParser = new StanfordDepHandler(posAnnotator, parseAnnotator, maxLength, throwExceptionOnSentenceLengthCheck);
            viewGenerators.put(depParser.getViewName(), depParser);
        }
        if (rm.getBoolean(PipelineConfigurator.USE_STANFORD_PARSE)) {
            StanfordParseHandler parser = new StanfordParseHandler(posAnnotator, parseAnnotator, maxLength, throwExceptionOnSentenceLengthCheck);
            viewGenerators.put(parser.getViewName(), parser);
        }
    }
    if (rm.getBoolean(PipelineConfigurator.USE_SRL_VERB)) {
        Properties verbProps = new Properties();
        String verbType = SRLType.Verb.name();
        verbProps.setProperty(SrlConfigurator.SRL_TYPE.key, verbType);
        ResourceManager verbRm = new ResourceManager(verbProps);
        rm = Configurator.mergeProperties(rm, verbRm);
        try {
            SemanticRoleLabeler verbSrl = new SemanticRoleLabeler(rm, useLazyInitialization);
            viewGenerators.put(ViewNames.SRL_VERB, verbSrl);
        } catch (Exception e) {
            throw new IOException("SRL verb cannot init: " + e.getMessage());
        }
    }
    if (rm.getBoolean(PipelineConfigurator.USE_SRL_NOM)) {
        Properties nomProps = new Properties();
        String nomType = SRLType.Nom.name();
        nomProps.setProperty(SrlConfigurator.SRL_TYPE.key, nomType);
        ResourceManager nomRm = new ResourceManager(nomProps);
        rm = Configurator.mergeProperties(rm, nomRm);
        try {
            SemanticRoleLabeler nomSrl = new SemanticRoleLabeler(rm, useLazyInitialization);
            // note that you can't call nomSrl (or verbSrl).getViewName() as it may not be
            // initialized yet
            viewGenerators.put(ViewNames.SRL_NOM, nomSrl);
        // viewGenerators.put(ViewNames.SRL_NOM,new SrlHandler("NomSRL", "5.1.9", nomType,
        // ViewNames.SRL_NOM,
        // useLazyInitialization, rm));
        } catch (Exception e) {
            throw new IOException("SRL nom cannot init .." + e.getMessage());
        }
    }
    if (rm.getBoolean(PipelineConfigurator.USE_QUANTIFIER)) {
        Quantifier quantifierAnnotator = new Quantifier();
        viewGenerators.put(ViewNames.QUANTITIES, quantifierAnnotator);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_TRANSLITERATION)) {
        for (Language lang : TransliterationAnnotator.supportedLanguages) {
            TransliterationAnnotator transliterationAnnotator = new TransliterationAnnotator(true, lang);
            viewGenerators.put(ViewNames.TRANSLITERATION + "_" + lang.getCode(), transliterationAnnotator);
        }
    }
    if (rm.getBoolean(PipelineConfigurator.USE_SRL_PREP)) {
        PrepSRLAnnotator prepSRLAnnotator = new PrepSRLAnnotator();
        viewGenerators.put(ViewNames.SRL_PREP, prepSRLAnnotator);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_SRL_COMMA)) {
        CommaLabeler commaLabeler = new CommaLabeler();
        viewGenerators.put(ViewNames.SRL_COMMA, commaLabeler);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_VERB_SENSE)) {
        VerbSenseAnnotator verbSense = new VerbSenseAnnotator();
        viewGenerators.put(ViewNames.VERB_SENSE, verbSense);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_MENTION)) {
        MentionAnnotator mentionAnnotator = new MentionAnnotator("ACE_TYPE");
        viewGenerators.put(ViewNames.MENTION, mentionAnnotator);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_RELATION)) {
        viewGenerators.put(ViewNames.RELATION, new RelationAnnotator(true));
    }
    if (rm.getBoolean(PipelineConfigurator.USE_TIMEX3)) {
        Properties rmProps = new TemporalChunkerConfigurator().getDefaultConfig().getProperties();
        TemporalChunkerAnnotator tca = new TemporalChunkerAnnotator(new ResourceManager(rmProps));
        viewGenerators.put(ViewNames.TIMEX3, tca);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_DATALESS_ESA)) {
        rm = new ESADatalessConfigurator().getConfig(nonDefaultRm);
        ESADatalessAnnotator esaDataless = new ESADatalessAnnotator(rm);
        viewGenerators.put(ViewNames.DATALESS_ESA, esaDataless);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_DATALESS_W2V)) {
        rm = new W2VDatalessConfigurator().getConfig(nonDefaultRm);
        W2VDatalessAnnotator w2vDataless = new W2VDatalessAnnotator(rm);
        viewGenerators.put(ViewNames.DATALESS_W2V, w2vDataless);
    }
    if (rm.getBoolean(PipelineConfigurator.USE_QUESTION_TYPER)) {
        QuestionTypeAnnotator questionTyper = new QuestionTypeAnnotator();
        viewGenerators.put(ViewNames.QUESTION_TYPE, questionTyper);
    }
    return viewGenerators;
}