Examples of edu.stanford.nlp.trees.tregex.TregexMatcher

edu.stanford.nlp.trees.tregex.TregexMatcher
A TregexMatcher can be used to match a {@link TregexPattern} against a {@link edu.stanford.nlp.trees.Tree}. Usage should be similar to a {@link java.util.regex.Matcher}. @author Galen Andrew

    // Accumulate all annotations in this string
    StringBuilder annotationStr = new StringBuilder();


    for (String featureName : features) {
      Pair<TregexPattern, Function<TregexMatcher, String>> behavior = annotationPatterns.get(featureName);
      TregexMatcher m = behavior.first().matcher(root);
      if (m.matchesAt(t))
        annotationStr.append(behavior.second().apply(m));
    }


    return annotationStr.toString();
  }

View Full Code Here

      TreeReader tr = trf.newTreeReader(br);


      final TregexPattern pMWE = TregexPattern.compile("/^MW/");
      for(Tree t; (t = tr.readTree()) != null;) {
        //Count MWE statistics
        TregexMatcher m = pMWE.matcher(t);
        while(m.findNextMatchingNode()) {
          Tree match = m.getMatch();
          String label = match.value();
          List<CoreLabel> yield = match.taggedLabeledYield();
          StringBuilder termYield = new StringBuilder();
          StringBuilder posYield = new StringBuilder();
          for(CoreLabel cl : yield) {

View Full Code Here

    Tree tree = s.get(TreeCoreAnnotations.TreeAnnotation.class);
    tree.indexLeaves();
    SemanticGraph dependency = s.get(SemanticGraphCoreAnnotations.CollapsedDependenciesAnnotation.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);

View Full Code Here

    List<CoreLabel> sent = s.get(CoreAnnotations.TokensAnnotation.class);
    Tree tree = s.get(TreeCoreAnnotations.TreeAnnotation.class);
    SemanticGraph dependency = s.get(SemanticGraphCoreAnnotations.CollapsedDependenciesAnnotation.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);

View Full Code Here

    return tgrepPatterns;
  }


  private static boolean checkPleonastic(Mention m, Tree tree, TregexPattern tgrepPattern) {
    try {
      TregexMatcher matcher = tgrepPattern.matcher(tree);
      while (matcher.find()) {
        Tree np1 = matcher.getNode("m1");
        if (((CoreLabel)np1.label()).get(CoreAnnotations.BeginIndexAnnotation.class)+1 == m.headWord.get(CoreAnnotations.IndexAnnotation.class)) {
          return true;
        }
      }
    } catch (Exception e) {

View Full Code Here

    }
  }


  private void findTreePattern(Tree tree, TregexPattern tgrepPattern, Set<Pair<Integer, Integer>> foundPairs) {
    try {
      TregexMatcher m = tgrepPattern.matcher(tree);
      while (m.find()) {
        Tree t = m.getMatch();
        Tree np1 = m.getNode("m1");
        Tree np2 = m.getNode("m2");
        Tree np3 = null;
        if(tgrepPattern.pattern().contains("m3")) np3 = m.getNode("m3");
        addFoundPair(np1, np2, t, foundPairs);
        if(np3!=null) addFoundPair(np2, np3, t, foundPairs);
      }
    } catch (Exception e) {
      // shouldn't happen....

View Full Code Here

      TwoDimensionalCounter<String, String> termLabel)
  {
    updateTagger(unigramTagger,t);


    //Count MWE statistics
    TregexMatcher m = pMWE.matcher(t);
    while (m.findNextMatchingNode()) {
      Tree match = m.getMatch();
      String label = match.value();
      if(RESOLVE_DUMMY_TAGS && label.equals(FrenchXMLTreeReader.MISSING_PHRASAL))
        continue;


      String preterm = Sentence.listToString(match.preTerminalYield());

View Full Code Here

    Set<TreeGraphNode> nodeList = new ArraySet<TreeGraphNode>();
    for (TregexPattern p : targetPatterns) {    // cdm: I deleted: && nodeList.isEmpty()
      // Initialize the TregexMatcher with the HeadFinder so that we
      // can use the same HeadFinder through the entire process of
      // building the dependencies
      TregexMatcher m = p.matcher(root, headFinder);
      while (m.findAt(t)) {
        TreeGraphNode target = (TreeGraphNode) m.getNode("target");
        if (target == null) {
          throw new AssertionError("Expression has no target: " + p);
        }
        nodeList.add(target);
        if (DEBUG) {
          System.err.println("found " + this + "(" + t + "-" + t.headWordNode() + ", " + m.getNode("target") + "-" + ((TreeGraphNode) m.getNode("target")).headWordNode() + ") using pattern " + p);
          for (String nodeName : m.getNodeNames()) {
            if (nodeName.equals("target"))
              continue;
            System.err.println("  node " + nodeName + ": " + m.getNode(nodeName));
          }
        }
      }
    }
    return nodeList;

View Full Code Here


  /**
   * Expand clitic pronouns on verbs matching the given pattern.
   */
  private static Tree expandCliticPronounsInner(Tree t, TregexPattern pattern) {
    TregexMatcher matcher = pattern.matcher(t);
    while (matcher.find()) {
      Tree verbNode = matcher.getNode("vb");
      String verb = verbNode.value();


      if (!SpanishVerbStripper.isStrippable(verb))
        continue;


      Pair<String, List<String>> split = verbStripper.separatePronouns(verb);
      if (split == null)
        continue;


      // Retrieve some context for the pronoun disambiguator: take the
      // matched clause and walk (at most) two constituents up
      StringBuilder clauseYieldBuilder = new StringBuilder();
      for (Label label : matcher.getNode("clause").yield())
        clauseYieldBuilder.append(label.value()).append(" ");
      String clauseYield = clauseYieldBuilder.toString();
      clauseYield = clauseYield.substring(0, clauseYield.length() - 1);


      // Insert clitic pronouns as leaves of pronominal phrases which are

View Full Code Here


  /**
   * ¡Venga, expand conmigo!
   */
  private static Tree expandConmigo(Tree t) {
    TregexMatcher matcher = conmigoPattern.matcher(t);


    while (matcher.find()) {
      Tree conmigoNode = matcher.getNode("conmigo");
      String word = conmigoNode.value();


      String newPronoun = null;
      if (word.equalsIgnoreCase("conmigo"))
        newPronoun = "mí";

View Full Code Here

0 1 2

TOP

Related Classes of edu.stanford.nlp.trees.tregex.TregexMatcher

edu.stanford.nlp.dcoref.Mention

edu.stanford.nlp.dcoref.MentionExtractor

edu.stanford.nlp.dcoref.RuleBasedCorefMentionFinder

edu.stanford.nlp.international.french.pipeline.FTBCorrector

edu.stanford.nlp.international.french.pipeline.MWEPreprocessor

edu.stanford.nlp.international.french.scripts.MWEFrequencyDist

edu.stanford.nlp.international.french.scripts.SplitCanditoTrees

edu.stanford.nlp.international.spanish.pipeline.AnCoraProcessor

edu.stanford.nlp.parser.lexparser.TregexPoweredTreebankParserParams

edu.stanford.nlp.parser.tools.RHSFrequency

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.