Examples with TregexPattern edu.stanford.nlp.trees.tregex.TregexPattern used on opensource projects

Example 6 with TregexPattern

use of edu.stanford.nlp.trees.tregex.TregexPattern in project CoreNLP by stanfordnlp.

the class CorefMentionFinder method getPleonasticPatterns.

private static TregexPattern[] getPleonasticPatterns() {
    final String[] patterns = { // in practice, go with this one (best results)
    "@NP < (PRP=m1 < it|IT|It) $.. (@VP < (/^V.*/ < /^(?i:is|was|be|becomes|become|became)$/ $.. (@VP < (VBN $.. @S|SBAR))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (ADJP $.. (/S|SBAR/))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (ADJP < (/S|SBAR/))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (NP < /S|SBAR/)))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (NP $.. ADVP $.. /S|SBAR/)))", // these next 5 had buggy space in "$ ..", which I fixed
    "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (VP < (VBN $.. /S|SBAR/))))))", // extraposed. OK 1/2 correct; need non-adverbial case
    "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (ADJP $.. (/S|SBAR/))))))", // OK: 3/3 good matches on dev; but 3/4 wrong on WSJ
    "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (ADJP < (/S|SBAR/))))))", "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (NP < /S|SBAR/)))))", "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (NP $.. ADVP $.. /S|SBAR/)))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:seems|appears|means|follows)/) $.. /S|SBAR/))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:turns|turned)/) $.. PRT $.. /S|SBAR/))" };
    TregexPattern[] tgrepPatterns = new TregexPattern[patterns.length];
    for (int i = 0; i < tgrepPatterns.length; i++) {
        tgrepPatterns[i] = TregexPattern.compile(patterns[i]);
    }
    return tgrepPatterns;
}

Example 7 with TregexPattern

use of edu.stanford.nlp.trees.tregex.TregexPattern in project CoreNLP by stanfordnlp.

the class HybridCorefMentionFinder method extractNPorPRP.

private static void extractNPorPRP(CoreMap s, List<Mention> mentions, Set<IntPair> mentionSpanSet, Set<IntPair> namedEntitySpanSet) {
    List<CoreLabel> sent = s.get(CoreAnnotations.TokensAnnotation.class);
    Tree tree = s.get(TreeCoreAnnotations.TreeAnnotation.class);
    tree.indexLeaves();
    SemanticGraph basicDependency = s.get(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation.class);
    SemanticGraph enhancedDependency = s.get(SemanticGraphCoreAnnotations.EnhancedDependenciesAnnotation.class);
    if (enhancedDependency == null) {
        enhancedDependency = s.get(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation.class);
    }
    TregexPattern tgrepPattern = npOrPrpMentionPattern;
    TregexMatcher matcher = tgrepPattern.matcher(tree);
    while (matcher.find()) {
        Tree t = matcher.getMatch();
        List<Tree> mLeaves = t.getLeaves();
        int beginIdx = ((CoreLabel) mLeaves.get(0).label()).get(CoreAnnotations.IndexAnnotation.class) - 1;
        int endIdx = ((CoreLabel) mLeaves.get(mLeaves.size() - 1).label()).get(CoreAnnotations.IndexAnnotation.class);
        // try not to have span that ends with ,
        if (",".equals(sent.get(endIdx - 1).word())) {
            endIdx--;
        }
        IntPair mSpan = new IntPair(beginIdx, endIdx);
        //      if(!mentionSpanSet.contains(mSpan) && (!insideNE(mSpan, namedEntitySpanSet)) ) {
        if (!mentionSpanSet.contains(mSpan) && (!insideNE(mSpan, namedEntitySpanSet) || t.value().startsWith("PRP"))) {
            int dummyMentionId = -1;
            Mention m = new Mention(dummyMentionId, beginIdx, endIdx, sent, basicDependency, enhancedDependency, new ArrayList<>(sent.subList(beginIdx, endIdx)), t);
            mentions.add(m);
            mentionSpanSet.add(mSpan);
            if (m.originalSpan.size() > 1) {
                boolean isNE = true;
                for (CoreLabel cl : m.originalSpan) {
                    if (!cl.tag().startsWith("NNP"))
                        isNE = false;
                }
                if (isNE) {
                    namedEntitySpanSet.add(mSpan);
                }
            }
        }
    }
}

Example 8 with TregexPattern

use of edu.stanford.nlp.trees.tregex.TregexPattern in project CoreNLP by stanfordnlp.

the class RuleBasedCorefMentionFinder method extractNPorPRP.

public void extractNPorPRP(CoreMap s, List<Mention> mentions, Set<IntPair> mentionSpanSet, Set<IntPair> namedEntitySpanSet) {
    List<CoreLabel> sent = s.get(CoreAnnotations.TokensAnnotation.class);
    Tree tree = s.get(TreeCoreAnnotations.TreeAnnotation.class);
    tree.indexLeaves();
    SemanticGraph basicDependency = s.get(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation.class);
    SemanticGraph enhancedDependency = s.get(SemanticGraphCoreAnnotations.EnhancedDependenciesAnnotation.class);
    if (enhancedDependency == null) {
        enhancedDependency = s.get(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation.class);
    }
    TregexPattern tgrepPattern = npOrPrpMentionPattern;
    TregexMatcher matcher = tgrepPattern.matcher(tree);
    while (matcher.find()) {
        Tree t = matcher.getMatch();
        List<Tree> mLeaves = t.getLeaves();
        int beginIdx = ((CoreLabel) mLeaves.get(0).label()).get(CoreAnnotations.IndexAnnotation.class) - 1;
        int endIdx = ((CoreLabel) mLeaves.get(mLeaves.size() - 1).label()).get(CoreAnnotations.IndexAnnotation.class);
        //if (",".equals(sent.get(endIdx-1).word())) { endIdx--; } // try not to have span that ends with ,
        IntPair mSpan = new IntPair(beginIdx, endIdx);
        if (!mentionSpanSet.contains(mSpan) && (lang == Locale.CHINESE || !insideNE(mSpan, namedEntitySpanSet))) {
            //      if(!mentionSpanSet.contains(mSpan) && (!insideNE(mSpan, namedEntitySpanSet) || t.value().startsWith("PRP")) ) {
            int dummyMentionId = -1;
            Mention m = new Mention(dummyMentionId, beginIdx, endIdx, sent, basicDependency, enhancedDependency, new ArrayList<>(sent.subList(beginIdx, endIdx)), t);
            mentions.add(m);
            mentionSpanSet.add(mSpan);
        //        if(m.originalSpan.size() > 1) {
        //          boolean isNE = true;
        //          for(CoreLabel cl : m.originalSpan) {
        //            if(!cl.tag().startsWith("NNP")) isNE = false;
        //          }
        //          if(isNE) {
        //            namedEntitySpanSet.add(mSpan);
        //          }
        //        }
        }
    }
}

Example 9 with TregexPattern

use of edu.stanford.nlp.trees.tregex.TregexPattern in project CoreNLP by stanfordnlp.

the class RuleBasedCorefMentionFinder method extractEnumerations.

protected static void extractEnumerations(CoreMap s, List<Mention> mentions, Set<IntPair> mentionSpanSet, Set<IntPair> namedEntitySpanSet) {
    List<CoreLabel> sent = s.get(CoreAnnotations.TokensAnnotation.class);
    Tree tree = s.get(TreeCoreAnnotations.TreeAnnotation.class);
    SemanticGraph dependency = s.get(SemanticGraphCoreAnnotations.EnhancedDependenciesAnnotation.class);
    TregexPattern tgrepPattern = enumerationsMentionPattern;
    TregexMatcher matcher = tgrepPattern.matcher(tree);
    Map<IntPair, Tree> spanToMentionSubTree = Generics.newHashMap();
    while (matcher.find()) {
        matcher.getMatch();
        Tree m1 = matcher.getNode("m1");
        Tree m2 = matcher.getNode("m2");
        List<Tree> mLeaves = m1.getLeaves();
        int beginIdx = ((CoreLabel) mLeaves.get(0).label()).get(CoreAnnotations.IndexAnnotation.class) - 1;
        int endIdx = ((CoreLabel) mLeaves.get(mLeaves.size() - 1).label()).get(CoreAnnotations.IndexAnnotation.class);
        spanToMentionSubTree.put(new IntPair(beginIdx, endIdx), m1);
        mLeaves = m2.getLeaves();
        beginIdx = ((CoreLabel) mLeaves.get(0).label()).get(CoreAnnotations.IndexAnnotation.class) - 1;
        endIdx = ((CoreLabel) mLeaves.get(mLeaves.size() - 1).label()).get(CoreAnnotations.IndexAnnotation.class);
        spanToMentionSubTree.put(new IntPair(beginIdx, endIdx), m2);
    }
    for (IntPair mSpan : spanToMentionSubTree.keySet()) {
        if (!mentionSpanSet.contains(mSpan) && !insideNE(mSpan, namedEntitySpanSet)) {
            int dummyMentionId = -1;
            Mention m = new Mention(dummyMentionId, mSpan.get(0), mSpan.get(1), dependency, new ArrayList<>(sent.subList(mSpan.get(0), mSpan.get(1))), spanToMentionSubTree.get(mSpan));
            mentions.add(m);
            mentionSpanSet.add(mSpan);
        }
    }
}

Example 10 with TregexPattern

use of edu.stanford.nlp.trees.tregex.TregexPattern in project CoreNLP by stanfordnlp.

the class RuleBasedCorefMentionFinder method getPleonasticPatterns.

private static TregexPattern[] getPleonasticPatterns() {
    final String[] patterns = { // in practice, go with this one (best results)
    "@NP < (PRP=m1 < it|IT|It) $.. (@VP < (/^V.*/ < /^(?i:is|was|be|becomes|become|became)$/ $.. (@VP < (VBN $.. @S|SBAR))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (ADJP $.. (/S|SBAR/))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (ADJP < (/S|SBAR/))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (NP < /S|SBAR/)))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:is|was|become|became)/) $.. (NP $.. ADVP $.. /S|SBAR/)))", // these next 5 had buggy space in "$ ..", which I fixed
    "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (VP < (VBN $.. /S|SBAR/))))))", // extraposed. OK 1/2 correct; need non-adverbial case
    "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (ADJP $.. (/S|SBAR/))))))", // OK: 3/3 good matches on dev; but 3/4 wrong on WSJ
    "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (ADJP < (/S|SBAR/))))))", "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (NP < /S|SBAR/)))))", "NP < (PRP=m1) $.. (VP < (MD $.. (VP < ((/^V.*/ < /^(?:be|become)/) $.. (NP $.. ADVP $.. /S|SBAR/)))))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:seems|appears|means|follows)/) $.. /S|SBAR/))", "NP < (PRP=m1) $.. (VP < ((/^V.*/ < /^(?:turns|turned)/) $.. PRT $.. /S|SBAR/))" };
    TregexPattern[] tgrepPatterns = new TregexPattern[patterns.length];
    for (int i = 0; i < tgrepPatterns.length; i++) {
        tgrepPatterns[i] = TregexPattern.compile(patterns[i]);
    }
    return tgrepPatterns;
}