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

    // Additional processing for specific phrasal annotations
    //


    // special global coding for moving PRD annotation from constituent to verb tag.
    if (markPRDverb) {
      TregexMatcher m = prdVerbPattern.matcher(tree);
      Tree match = null;
      while (m.find()) {
        if (m.getMatch() != match) {
          match = m.getMatch();
          match.label().setValue(match.label().value() + "-PRDverb");
          Tree prd = m.getNode("prd");
          prd.label().setValue(super.normalizeNonterminal(prd.label().value()));
        }
      }
    }


    //Mark *only* subjects in verb-initial clauses
    if(retainNPSbj) {
      TregexMatcher m = npSbjPattern.matcher(tree);
      while (m.find()) {
        Tree match = m.getMatch();
        match.label().setValue("NP");
      }
    }


    if (tree.isPreTerminal()) {

View Full Code Here

   * @return The endpoint of a subtree which should be extracted, or
   *         {@code null} if there are no subtrees which need to be
   *         extracted.
   */
  static Tree findSplitPoint(Tree t) {
    TregexMatcher m = pSplitPoint.matcher(t);
    if (m.find())
      return m.getMatch();
    return null;
  }

View Full Code Here

      TreeReaderFactory trf = new FrenchTreeReaderFactory();
      TreeReader tr = trf.newTreeReader(br);
   
      int nTrees = 0;
      for(Tree t; (t = tr.readTree()) != null;nTrees++) {
        TregexMatcher m = pBadTree.matcher(t);
        TregexMatcher m2 = pBadTree2.matcher(t);
        if(m.find() || m2.find()) {
          System.err.println("Discarding tree: " + t.toString());
        } else {
          Tree fixedT = tt.transformTree(t);
          System.out.println(fixedT.toString());
        }

View Full Code Here

        System.err.println("Missing id: " + id);
        continue;
      }


      Tree tree = treeMap.get(id);
      TregexMatcher m = pBadTree.matcher(tree);
      TregexMatcher m2 = pBadTree2.matcher(tree);
      if(m.find() || m2.find()) {
        System.err.println("Discarding tree: " + tree.toString());
        continue;
      }
      
      // Punctuation normalization, etc.

View Full Code Here

    // other key words as the head.  For example, "and not" means
    // something completely different than "and".  Furthermore,
    // downstream code was written assuming "not" would be the head...
    if (motherCat.equals("CONJP")) {
      for (TregexPattern pattern : headOfConjpTregex) {
        TregexMatcher matcher = pattern.matcher(t);
        if (matcher.matchesAt(t)) {
          return matcher.getNode("head");
        }
      }
      // if none of the above patterns match, use the standard method
    }


    if (motherCat.equals("SBARQ") || motherCat.equals("SINV")) {
      if (!makeCopulaHead) {
        for (TregexPattern pattern : headOfCopulaTregex) {
          TregexMatcher matcher = pattern.matcher(t);
          if (matcher.matchesAt(t)) {
            return matcher.getNode("head");
          }
        }
      }
      // if none of the above patterns match, use the standard method
    }

View Full Code Here

  static final String MONTH_REGEX = "January|February|March|April|May|June|July|August|September|October|November|December|Jan\\.|Feb\\.|Mar\\.|Apr\\.|Aug\\.|Sep\\.|Sept\\.|Oct\\.|Nov\\.|Dec\\.";
  static final TregexPattern tregexMonthYear = TregexPatternCompiler.defaultCompiler.compile("NP=root <1 (NP <: (NNP=month <: /" + MONTH_REGEX + "/)) <2 (NP=yearnp <: (CD=year <: __)) : =root <- =yearnp");
  static final TregexPattern tregexMonthDayYear = TregexPatternCompiler.defaultCompiler.compile("NP=root <1 (NP=monthdayroot <1 (NNP=month <: /" + MONTH_REGEX +"/) <2 (CD=day <: __)) <2 (/^,$/=comma <: /^,$/) <3 (NP=yearroot <: (CD=year <: __)) : (=root <- =yearroot) : (=monthdayroot <- =day)");


  public Tree transformTree(Tree t) {
    TregexMatcher matcher = tregexMonthYear.matcher(t);
    while (matcher.find()) {
      Tree root = matcher.getNode("root");
      Tree month = matcher.getNode("month");
      Tree year = matcher.getNode("year");
      Tree[] children = new Tree[] {month, year};
      root.setChildren(children);
      matcher = tregexMonthYear.matcher(t);
    }
    matcher = tregexMonthDayYear.matcher(t);
    while (matcher.find()) {
      Tree root = matcher.getNode("root");
      Tree month = matcher.getNode("month");
      Tree day = matcher.getNode("day");
      Tree comma = matcher.getNode("comma");
      Tree year = matcher.getNode("year");
      Tree[] children = new Tree[] {month, day, comma, year};
      root.setChildren(children);
      matcher = tregexMonthDayYear.matcher(t);
    }
    return t;

View Full Code Here

      }
    }


    Counter<String> rhsCounter = new ClassicCounter<String>();
    for(Tree t : tb) {
      TregexMatcher m = rootMatch.matcher(t);
      while(m.findNextMatchingNode()) {
        Tree match = m.getMatch();
        StringBuilder sb = new StringBuilder();
        for(Tree kid : match.children())
          sb.append(kid.value()).append(" ");
        rhsCounter.incrementCount(sb.toString().trim());
      }

View Full Code Here

      }


      final String enumerationPattern = "NP < (NP=tmp $.. (/,|CC/ $.. NP))";


      TregexPattern tgrepPattern = TregexPattern.compile(enumerationPattern);
      TregexMatcher m = tgrepPattern.matcher(this.mentionSubTree);
      while (m.find()) {
        //        Tree t = m.getMatch();
        if(this.mentionSubTree==m.getNode("tmp")
           && this.spanToString().toLowerCase().contains(" and ")) {
          number = Number.PLURAL;
        }
      }
    }

View Full Code Here

   * @param p A {@link TsurgeonPattern} to apply.
   * @param t the {@link Tree} to match against and perform surgery on.
   * @return t, which has been surgically modified.
   */
  public static Tree processPattern(TregexPattern matchPattern, TsurgeonPattern p, Tree t) {
    TregexMatcher m = matchPattern.matcher(t);
    TsurgeonMatcher tsm = p.matcher();
    while (m.find()) {
      t = tsm.evaluate(t, m);
      if (t==null) {
        break;
      }
      m = matchPattern.matcher(t);

View Full Code Here

    for (Pair<TregexPattern,TsurgeonPattern> op : ops) {
      try {
        if (DEBUG) {
          System.err.println("Running pattern " + op.first());
        }
        TregexMatcher m = op.first().matcher(t);
        TsurgeonMatcher tsm = op.second().matcher();
        while (m.find()) {
          matchedOnTree = true;
          t = tsm.evaluate(t,m);
          if (t == null) {
            return null;
          }

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.