Examples with Constituent edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent used on opensource projects

Example 46 with Constituent

use of edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent in project cogcomp-nlp by CogComp.

the class TestPOSMikheevFeatureExtractor method test.

@Test
public final void test() throws Exception {
    POSMikheevFeatureExtractor posMikheev = new POSMikheevFeatureExtractor("posMikheev", "test_corpus", TestPosHelper.corpus);
    logger.info("POSMikheev Feature Extractor");
    logger.info("Only print the features with known tags");
    // Using the first TA and a constituent between span of 30-40 as a test
    int i = 0;
    for (TextAnnotation ta : tas) {
        ArrayList<String> outFeatures = new ArrayList<>();
        View TOKENS = ta.getView("TOKENS");
        for (Constituent TOKEN : TOKENS) {
            Set<Feature> feats = posMikheev.getFeatures(TOKEN);
            if (feats.isEmpty()) {
                logger.info("Feats list is returning NULL.");
            }
            for (Feature f : feats) if (!f.getName().contains("UNKNOWN")) {
                outFeatures.add(f.getName());
            }
        }
        if (!outFeatures.isEmpty()) {
            logger.info("-------------------------------------------------------");
            logger.info("Text Annotation: " + i);
            logger.info("Text Features: ");
            for (String out : outFeatures) logger.info(out);
            logger.info("-------------------------------------------------------");
        }
        i++;
    }
    logger.info("GOT FEATURES YES!");
}

Example 47 with Constituent

use of edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent in project cogcomp-nlp by CogComp.

the class TestPOSWindowTwo method test.

@Test
public final void test() throws EdisonException {
    log.debug("POSWindowTwo Feature Extractor");
    // Using the first TA and a constituent between span of 30-40 as a test
    TextAnnotation ta = tas.get(2);
    View TOKENS = ta.getView("TOKENS");
    List<Constituent> testlist = TOKENS.getConstituentsCoveringSpan(0, 20);
    for (Constituent c : testlist) {
        log.debug(c.getSurfaceForm());
    }
    Constituent test = testlist.get(1);
    log.debug("The constituent we are extracting features from in this test is: " + test.getSurfaceForm());
    POSWindowTwo POSW = new POSWindowTwo("POSWindowTwo");
    log.debug("Startspan is " + test.getStartSpan() + " and Endspan is " + test.getEndSpan());
    Set<Feature> feats = POSW.getFeatures(test);
    String[] expected_outputs = { "POSWindowTwo:0(DT)", "POSWindowTwo:1(VBZ)", "POSWindowTwo:2(DT)", "POSWindowTwo:3(NN)" };
    if (feats == null) {
        log.debug("Feats are returning NULL.");
    }
    log.debug("Printing Set of Features");
    for (Feature f : feats) {
        assert (ArrayUtils.contains(expected_outputs, f.getName()));
    }
// System.exit(0);
}

Example 48 with Constituent

use of edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent in project cogcomp-nlp by CogComp.

the class TestPosWordConjunctionSizeTwoWindowSizeTwo method test.

@Test
public final void test() throws EdisonException {
    // / Using the 3rd constituent as a test
    List<Constituent> testList = ta.getView("TOKENS").getConstituents();
    Constituent test = testList.get(3);
    PosWordConjunctionSizeTwoWindowSizeTwo fex = new PosWordConjunctionSizeTwoWindowSizeTwo("PosWordConj2Win2");
    Set<Feature> feats = fex.getFeatures(test);
    String[] expected_outputs = { "PosWordConjunctionSizeTwoWindowSizeTwo:-2_1(NN-construction)", "PosWordConjunctionSizeTwoWindowSizeTwo:-1_1(IN-of)", "PosWordConjunctionSizeTwoWindowSizeTwo:0_1(DT-the)", "PosWordConjunctionSizeTwoWindowSizeTwo:1_1(NNP-John)", "PosWordConjunctionSizeTwoWindowSizeTwo:2_1(NNP-Smith)", "PosWordConjunctionSizeTwoWindowSizeTwo:-2_2(NN-construction_IN-of)", "PosWordConjunctionSizeTwoWindowSizeTwo:-1_2(IN-of_DT-the)", "PosWordConjunctionSizeTwoWindowSizeTwo:0_2(DT-the_NNP-John)", "PosWordConjunctionSizeTwoWindowSizeTwo:1_2(NNP-John_NNP-Smith)", "PosWordConjunctionSizeTwoWindowSizeTwo:2_2(NNP-Smith)" };
    if (feats == null)
        fail("Feats are returning NULL.");
    for (Feature f : feats) {
        assertTrue(ArrayUtils.contains(expected_outputs, f.getName()));
    }
}

Example 49 with Constituent

use of edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent in project cogcomp-nlp by CogComp.

the class TestSrlVerbSimplePredicateFeatures method test.

public final void test() throws Exception {
    logger.info("PredicateFeatures Feature Extractor");
    String[] viewsToAdd = { ViewNames.POS, ViewNames.LEMMA, ViewNames.SHALLOW_PARSE, ViewNames.PARSE_GOLD, ViewNames.SRL_VERB, ViewNames.PARSE_STANFORD, ViewNames.NER_CONLL };
    TextAnnotation ta = DummyTextAnnotationGenerator.generateAnnotatedTextAnnotation(viewsToAdd, true, 3);
    int i = 0;
    ta.addView(ClauseViewGenerator.STANFORD);
    ta.addView(PseudoParse.STANFORD);
    logger.info("This textannotation annotates the text: \n" + ta.getText());
    View TOKENS = ta.getView("TOKENS");
    List<Constituent> testlist = TOKENS.getConstituentsCoveringSpan(10, 13);
    testlist.addAll(TOKENS.getConstituentsCoveringSpan(26, 27));
    FeatureManifest featureManifest;
    FeatureExtractor fex;
    String fileName = Constant.prefix + "/Verb/Predicate/predicate-features.fex";
    featureManifest = new FeatureManifest(new FileInputStream(fileName));
    FeatureManifest.setFeatureExtractor("hyphen-argument-feature", FeatureGenerators.hyphenTagFeature);
    FeatureManifest.setTransformer("parse-left-sibling", FeatureGenerators.getParseLeftSibling(ViewNames.PARSE_STANFORD));
    FeatureManifest.setTransformer("parse-right-sibling", FeatureGenerators.getParseRightSibling(ViewNames.PARSE_STANFORD));
    FeatureManifest.setFeatureExtractor("pp-features", FeatureGenerators.ppFeatures(ViewNames.PARSE_STANFORD));
    FeatureManifest.setFeatureExtractor("projected-path", new ProjectedPath(ViewNames.PARSE_STANFORD));
    featureManifest.useCompressedName();
    featureManifest.setVariable("*default-parser*", ViewNames.PARSE_STANFORD);
    fex = featureManifest.createFex();
    SrlVerbSimplePredicateFeatures pf = new SrlVerbSimplePredicateFeatures();
    for (Constituent test : testlist) {
        assertTrue(SRLFeaturesComparator.isEqual(test, fex, pf));
    }
}

Example 50 with Constituent

use of edu.illinois.cs.cogcomp.core.datastructures.textannotation.Constituent in project cogcomp-nlp by CogComp.

the class TestSrlVerbPredicateFeatures method test.

public final void test() throws Exception {
    logger.info("PredicateFeatures Feature Extractor");
    String[] viewsToAdd = { ViewNames.POS, ViewNames.LEMMA, ViewNames.SHALLOW_PARSE, ViewNames.PARSE_GOLD, ViewNames.SRL_VERB, ViewNames.PARSE_STANFORD, ViewNames.NER_CONLL };
    TextAnnotation ta = DummyTextAnnotationGenerator.generateAnnotatedTextAnnotation(viewsToAdd, true, 3);
    int i = 0;
    ta.addView(ClauseViewGenerator.STANFORD);
    ta.addView(PseudoParse.STANFORD);
    logger.info("This textannotation annotates the text: \n" + ta.getText());
    View TOKENS = ta.getView("TOKENS");
    List<Constituent> testlist = TOKENS.getConstituentsCoveringSpan(10, 13);
    testlist.addAll(TOKENS.getConstituentsCoveringSpan(26, 27));
    FeatureManifest featureManifest;
    FeatureExtractor fex;
    String fileName = Constant.prefix + "/Verb/Identifier/predicate-features.fex";
    featureManifest = new FeatureManifest(new FileInputStream(fileName));
    FeatureManifest.setFeatureExtractor("hyphen-argument-feature", FeatureGenerators.hyphenTagFeature);
    FeatureManifest.setTransformer("parse-left-sibling", FeatureGenerators.getParseLeftSibling(ViewNames.PARSE_STANFORD));
    FeatureManifest.setTransformer("parse-right-sibling", FeatureGenerators.getParseRightSibling(ViewNames.PARSE_STANFORD));
    FeatureManifest.setFeatureExtractor("pp-features", FeatureGenerators.ppFeatures(ViewNames.PARSE_STANFORD));
    FeatureManifest.setFeatureExtractor("projected-path", new ProjectedPath(ViewNames.PARSE_STANFORD));
    featureManifest.useCompressedName();
    featureManifest.setVariable("*default-parser*", ViewNames.PARSE_STANFORD);
    fex = featureManifest.createFex();
    SrlVerbPredicateFeatures af = new SrlVerbPredicateFeatures();
    for (Constituent test : testlist) {
        assertTrue(SRLFeaturesComparator.isEqual(test, fex, af));
    }
}