Examples with Feature edu.illinois.cs.cogcomp.edison.features.Feature used on opensource projects

Example 46 with Feature

use of edu.illinois.cs.cogcomp.edison.features.Feature in project cogcomp-nlp by CogComp.

the class TestLabelTwoBefore method test.

@Test
public final void test() throws Exception {
    logger.info("LabelTwoBefore 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");
    logger.info("GOT TOKENS FROM TEXTAnn");
    List<Constituent> testlist = TOKENS.getConstituentsCoveringSpan(0, 20);
    for (Constituent c : testlist) {
        logger.info(c.getSurfaceForm());
    }
    logger.info("Testlist size is " + testlist.size());
    // Constituent test = testlist.get(1);
    // logger.info("The constituent we are extracting features from
    // in this test is: " + test.getSurfaceForm());
    POSBaseLineCounter posBaseLine = new POSBaseLineCounter("posBaseLine");
    posBaseLine.buildTable(TestPosHelper.corpus);
    POSMikheevCounter posMikheev = new POSMikheevCounter("posMikheev");
    posMikheev.buildTable(TestPosHelper.corpus);
    LabelTwoBefore l2bPOS = new LabelTwoBefore("l2bPOS");
    LabelTwoBefore l2bPOSBaseLine = new LabelTwoBefore("l2bPOSBaseLine", posBaseLine);
    LabelTwoBefore l2bPOSMikheev = new LabelTwoBefore("l2bPOSMikheev", posMikheev);
    // Test when using POS View
    ArrayList<Set<Feature>> featslist = new ArrayList<>();
    for (Constituent test : testlist) featslist.add(l2bPOS.getFeatures(test));
    if (featslist.isEmpty()) {
        logger.info("Feats list is returning NULL.");
    }
    logger.info("\n" + "Test when using POS View");
    logger.info("Printing list of Feature set");
    for (Set<Feature> feats : featslist) {
        for (Feature f : feats) logger.info(f.getName());
    }
    // Test when using POS baseline Counting
    featslist.clear();
    for (Constituent test : testlist) featslist.add(l2bPOSBaseLine.getFeatures(test));
    if (featslist.isEmpty()) {
        logger.info("Feats list is returning NULL.");
    }
    logger.info("\n" + "Test when using POS baseline Counting");
    logger.info("Printing list of Feature set");
    for (Set<Feature> feats : featslist) {
        for (Feature f : feats) logger.info(f.getName());
    }
    // Test when using POS Mikheev Counting
    featslist.clear();
    for (Constituent test : testlist) featslist.add(l2bPOSMikheev.getFeatures(test));
    if (featslist.isEmpty()) {
        logger.info("Feats list is returning NULL.");
    }
    logger.info("\n" + "Test when using POS Mikheev Counting");
    logger.info("Printing list of Feature set");
    for (Set<Feature> feats : featslist) {
        for (Feature f : feats) logger.info(f.getName());
    }
    logger.info("GOT FEATURES YES!");
}

Example 47 with Feature

use of edu.illinois.cs.cogcomp.edison.features.Feature in project cogcomp-nlp by CogComp.

the class TestMixedChunkWindowTwoBeforePOSWindowThreeBefore method test.

@Test
public final void test() throws EdisonException {
    log.debug("SOPREVIOUS");
    // Using the first TA and a constituent between span of 0-20 as a test
    TextAnnotation ta = tas.get(3);
    View TOKENS = ta.getView("TOKENS");
    log.debug("GOT TOKENS FROM TEXTAnn");
    List<Constituent> testlist = TOKENS.getConstituentsCoveringSpan(0, 20);
    for (Constituent c : testlist) {
        log.debug(c.getSurfaceForm());
    }
    log.debug("Testlist size is " + testlist.size());
    Constituent test = testlist.get(5);
    log.debug("The constituent we are extracting features from in this test is: " + test.getSurfaceForm());
    MixedChunkWindowTwoBeforePOSWindowThreeBefore SOP = new MixedChunkWindowTwoBeforePOSWindowThreeBefore("MixedChunkWindowTwoBeforePOSWindowThreeBefore");
    Set<Feature> feats = SOP.getFeatures(test);
    String[] expected_outputs = { "MixedChunkWindowTwoBeforePOSWindowThreeBefore:ll(NP_VP)", "MixedChunkWindowTwoBeforePOSWindowThreeBefore:lt1(NP_RB)", "MixedChunkWindowTwoBeforePOSWindowThreeBefore:lt2VP_VBN" };
    if (feats == null) {
        log.debug("Feats are returning NULL.");
    }
    log.debug("Printing Set of Features");
    for (Feature f : feats) {
        log.debug(f.getName());
        assert (ArrayUtils.contains(expected_outputs, f.getName()));
    }
// System.exit(0);
}

Example 48 with Feature

use of edu.illinois.cs.cogcomp.edison.features.Feature in project cogcomp-nlp by CogComp.

the class WriteSVMLightFormat method writeFeatureExample.

/**
     * Generate a String value corresponding to a SVMLight format example from Edison feature
     * representation (String label and Collection of Feature). Stores a mapping between features
     * and integer ids so that when a new example is passed in, if a feature is the same as an
     * example that has already been processed, that feature will get the same integer id as it did
     * previously. From SVMLight documentation, binary classification labels must be {-1,1} for
     * binary problem, but otherwise labels are just integers -- one integer per class for
     * multi-class problem integer rank for ranking problem Apparently, no problem for label to have
     * same integer id as feature value
     *
     * Assumes sparse boolean feature representation (if a feature is active, it has a unique string
     * identifier in the Collection argument). IMPORTANT: feature ids MUST be written out in
     * ascending order.
     *
     * @param label
     * @param activeFeatures
     * @return
     */
public String writeFeatureExample(String label, Collection<Feature> activeFeatures) {
    int lab = getLabel(label);
    int[] featIds = new int[activeFeatures.size()];
    int index = 0;
    for (Feature f : activeFeatures) featIds[index++] = featureLex.getFeatureId(f.getName());
    Arrays.sort(featIds);
    StringBuilder bldr = new StringBuilder();
    bldr.append(lab);
    for (int featId : featIds) bldr.append(" ").append(featId).append(":1");
    return bldr.toString();
}

Example 49 with Feature

use of edu.illinois.cs.cogcomp.edison.features.Feature in project cogcomp-nlp by CogComp.

the class TestWordFeatureFactory method testFeatureCollection.

@Test
public final void testFeatureCollection() throws Exception {
    FeatureCollection f = new FeatureCollection("features");
    f.addFeatureExtractor(WordFeatureExtractorFactory.conflatedPOS);
    f.addFeatureExtractor(WordFeatureExtractorFactory.gerundMarker);
    f.addFeatureExtractor(WordFeatureExtractorFactory.nominalizationMarker);
    logger.info("\tTesting feature collection");
    Map<Integer, String> map = IOUtils.readObjectAsResource(TestWordFeatureFactory.class, "feature.collection.test");
    for (TextAnnotation ta : tas) {
        for (int tokenId = 0; tokenId < ta.size(); tokenId++) {
            Constituent c = new Constituent("", "", ta, tokenId, tokenId + 1);
            Set<Feature> features = f.getFeatures(c);
            if (features.size() > 0) {
                String id = ta.getTokenizedText() + ":" + tokenId;
                assertEquals(map.get(id.hashCode()), features.toString());
            }
        }
    }
}

Example 50 with Feature

use of edu.illinois.cs.cogcomp.edison.features.Feature in project cogcomp-nlp by CogComp.

the class TextStatistics method consume.

@Override
protected void consume(TextAnnotation ta) {
    for (Constituent c : constituentGenerator.transform(ta)) {
        try {
            Set<Feature> feats = fex.getFeatures(c);
            for (Feature feat : feats) {
                count(feat);
            }
            constituentCounter.incrementAndGet();
        } catch (EdisonException e) {
            e.printStackTrace();
        }
    }
    textCounter.incrementAndGet();
}