Java TreeFactory.newTreeNode方法代码示例

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override
protected boolean match(ExtractorMatcher matcher,
    TregexMatcher tregexMatcher, Tree tree) {

  boolean result = super.match(matcher, tregexMatcher, tree);
  if (result) {
    List<Tree> children = matcher.getMatch().getChildrenAsList();
    int triggerIndex = children.indexOf(matcher.trigger);
    int argIndex = children.indexOf(matcher.argument);

    matcher.matched = tregexMatcher.getMatch().deepCopy();
    children = matcher.matched.getChildrenAsList();

    matcher.trigger = children.get(triggerIndex);

    TreeFactory tf = matcher.matched.treeFactory();
    matcher.argument = tf.newTreeNode(
        "NP",
        children.subList(argIndex, children.size()));
  }
  return result;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override
protected boolean match(ExtractorMatcher matcher,
    TregexMatcher tregexMatcher, Tree tree) {

  boolean result = super.match(matcher, tregexMatcher, tree);
  if (result) {
    List<Tree> children = matcher.getMatch().getChildrenAsList();

    int argIndex = children.indexOf(matcher.argument);

    matcher.matched = tregexMatcher.getMatch().deepCopy();
    children = matcher.matched.getChildrenAsList();

    matcher.trigger = matcher.matched;

    TreeFactory tf = matcher.matched.treeFactory();
    matcher.argument = tf
        .newTreeNode("NP", children.subList(0, argIndex + 1));
  }
  return result;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
private Tree fixNonUnaryRoot(Tree t, TreeFactory tf) {
  List<Tree> kids = t.getChildrenAsList();
  if(kids.size() == 2 && t.firstChild().isPhrasal() && tlp.isSentenceFinalPunctuationTag(t.lastChild().value())) {
    List<Tree> grandKids = t.firstChild().getChildrenAsList();
    grandKids.add(t.lastChild());
    t.firstChild().setChildren(grandKids);
    kids.remove(kids.size() - 1);
    t.setChildren(kids);
    t.setValue(tlp.startSymbol());
    
  } else {
    t.setValue(nonUnaryRoot);
    t = tf.newTreeNode(tlp.startSymbol(), Collections.singletonList(t));
  }
  return t;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override

  protected boolean match(ExtractorMatcher matcher,
      TregexMatcher tregexMatcher,
      Tree tree) {

    boolean result = super.match(matcher, tregexMatcher, tree);
    if (result) {
      List<Tree> children = matcher.getMatch().getChildrenAsList();
      int triggerIndex = children.indexOf(matcher.trigger);
      int argIndex = children.indexOf(matcher.argument);

      matcher.matched = tregexMatcher.getMatch();
      TreeFactory tf = tregexMatcher.getMatch().treeFactory();

      matcher.matched = tregexMatcher.getMatch().deepCopy(tf);
      children = matcher.matched.getChildrenAsList();
      
      matcher.trigger = tf.newTreeNode(
          "NP",
            children.subList(triggerIndex, children.size()));

      matcher.argument = tf
          .newTreeNode("NP", children.subList(0, argIndex + 1));

      while (matcher.matched.numChildren() != 0) {
        matcher.matched.removeChild(0);
      }
      matcher.matched.addChild(matcher.argument);
      matcher.matched.addChild(matcher.trigger);
    }
    return result;
  }
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override
protected boolean match(ExtractorMatcher matcher,
    TregexMatcher tregexMatcher, Tree tree) {

  Tree leaf = tregexMatcher.getMatch().getChild(0);
  if (!leaf.isLeaf()) {
    return false;
  }

  OffsetLabel label = (OffsetLabel) leaf.label();
  int dash = label.value().lastIndexOf('-');
  if (dash == -1 || dash == 0) {
    return false;
  }

  matcher.matched = leaf.parent(tree).deepCopy();

  TreeFactory tf = matcher.matched.treeFactory();

  OffsetLabel newLeafLabel = new OffsetLabel(label.value()
      .substring(dash + 1));
  newLeafLabel.setBeginPosition(dash + 1 + label.beginPosition());
  newLeafLabel.setEndPosition(label.endPosition());
  
  leaf = tf.newLeaf(newLeafLabel);
  matcher.trigger = tf.newTreeNode("NP", Collections.singletonList(leaf));
  
  newLeafLabel = new OffsetLabel(label.value()
      .substring(0, dash));
  newLeafLabel.setBeginPosition(label.beginPosition());
  newLeafLabel.setEndPosition(label.beginPosition() + dash);
  
  leaf = tf.newLeaf(newLeafLabel);
  matcher.argument = tf.newTreeNode("NP", Collections.singletonList(leaf));

  matcher.matched.removeChild(0);
  matcher.matched.addChild(matcher.argument);
  matcher.matched.addChild(matcher.trigger);
  return true;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
Tree transformRoot(Tree tree, TreeFactory tf) {
  // XXXX TODO: use tlp and don't assume 1 daughter of ROOT!
  // leave the root intact
  // if (tlp.isStartSymbol(tlp.basicCategory(tree.label().value())))
  if (tree.label().toString().startsWith("ROOT")) {
    return tf.newTreeNode(tree.label(), Collections.singletonList(transformNode(tree.children()[0], tf)));
  }
  return transformNode(tree, tf);
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
/**
 * Construct a fall through tree in case we can't parse this sentence
 * @param words
 * @return a tree with X for all the internal nodes
 */
public static Tree xTree(List<? extends HasWord> words) {
  TreeFactory lstf = new LabeledScoredTreeFactory();
  List<Tree> lst2 = new ArrayList<Tree>();
  for (HasWord obj : words) {
    String s = obj.word();
    Tree t = lstf.newLeaf(s);
    Tree t2 = lstf.newTreeNode("X", Collections.singletonList(t));
    lst2.add(t2);
  }
  return lstf.newTreeNode("X", lst2);
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override
public Tree normalizeWholeTree(Tree tree, TreeFactory tf) {
  tree = tree.prune(hebrewEmptyFilter, tf).spliceOut(aOverAFilter, tf);

  //Add start symbol so that the root has only one sub-state. Escape any enclosing brackets.
  //If the "tree" consists entirely of enclosing brackets e.g. ((())) then this method
  //will return null. In this case, readers e.g. PennTreeReader will try to read the next tree.
  while(tree != null && (tree.value() == null || tree.value().equals("")) && tree.numChildren() <= 1)
    tree = tree.firstChild();

  if(tree != null && !tree.value().equals(tlp.startSymbol()))
    tree = tf.newTreeNode(tlp.startSymbol(), Collections.singletonList(tree));

  return tree;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
/**
 * Normalize a whole tree -- one can assume that this is the
 * root.  This implementation deletes empty elements (ones with
 * nonterminal tag label starting with '*T') from the tree.  It
 * does work for a null tree.
 */
@Override
public Tree normalizeWholeTree(Tree tree, TreeFactory tf) {
  // add an extra root to non-unary roots
  if(tree.value() == null)
    tree = fixNonUnaryRoot(tree, tf);
  else if(!tree.value().equals(tlp.startSymbol()))
    tree = tf.newTreeNode(tlp.startSymbol(), Collections.singletonList(tree));

  tree = tree.prune(emptyFilter, tf).spliceOut(aOverAFilter, tf);

  // insert NPs in PPs if you're supposed to do that
  if (insertNPinPP) {
    insertNPinPPall(tree);
  }

  for(Tree t : tree) {
    if(t.isLeaf() || t.isPreTerminal()) continue;
    if(t.value() == null || t.value().equals("")) t.setValue("DUMMY");

    // there's also a '--' category
    if(t.value().matches("--.*")) continue;

    // fix a bug in the ACL08 German tiger treebank
    String cat = t.value();
    if(cat == null || cat.equals("")) {
      if (t.numChildren() == 3 && t.firstChild().label().value().equals("NN") && t.getChild(1).label().value().equals("$.")) {
        System.err.println("Correcting treebank error: giving phrase label DL to " + t);
        t.label().setValue("DL");
      }
    }
  }

  return tree;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override
protected boolean match(ExtractorMatcher matcher,
    TregexMatcher tregexMatcher, Tree tree) {

  Tree leaf = tregexMatcher.getMatch().getChild(0);
  if (!leaf.isLeaf()) {
    return false;
  }

  OffsetLabel label = (OffsetLabel) leaf.label();
  if (label.value().contains("-RRB-")) {
    return false;
  }
  Matcher m = p.matcher(label.value());
  if (!m.find()) {
    return false;
  }

  matcher.matched = leaf.parent(tree).deepCopy();
  
  TreeFactory tf = matcher.matched.treeFactory();

  OffsetLabel newLeafLabel = new OffsetLabel(m.group(2));
  newLeafLabel.setBeginPosition(m.start(2) + label.beginPosition());
  newLeafLabel.setEndPosition(m.end(2) + label.beginPosition());
  
  leaf = tf.newLeaf(newLeafLabel);
  matcher.trigger = tf.newTreeNode(matcher.matched.value(), Collections.singletonList(leaf));
  
  newLeafLabel = new OffsetLabel(m.group(1));
  newLeafLabel.setBeginPosition(m.start(1) + label.beginPosition());
  newLeafLabel.setEndPosition(m.end(1) + label.beginPosition());
  
  leaf = tf.newLeaf(newLeafLabel);
  matcher.argument = tf.newTreeNode(matcher.matched.value(), Collections.singletonList(leaf));

  matcher.matched.removeChild(0);
  matcher.matched.addChild(matcher.argument);
  matcher.matched.addChild(matcher.trigger);
  return true;
}
 

import edu.stanford.nlp.trees.TreeFactory; //导入方法依赖的package包/类
@Override
public Tree normalizeWholeTree(Tree tree, TreeFactory tf) {
  tree = tree.prune(emptyFilter, tf).spliceOut(aOverAFilter, tf);

  for(Tree t : tree) {
    //Map punctuation tags back like the PTB
    if(t.isPreTerminal()) {
      String posStr = normalizePreterminal(t);
      t.setValue(posStr);
      if(t.label() instanceof HasTag) ((HasTag) t.label()).setTag(posStr);

    } else if(t.isLeaf()) {
      //Strip off morphological analyses and place them in the OriginalTextAnnotation, which is
      //specified by HasContext.
      if(t.value().contains(MorphoFeatureSpecification.MORPHO_MARK)) {
        String[] toks = t.value().split(MorphoFeatureSpecification.MORPHO_MARK);
        if(toks.length != 2)
          System.err.printf("%s: Word contains malformed morph annotation: %s%n",this.getClass().getName(),t.value());

        else if(t.label() instanceof CoreLabel) {
          ((CoreLabel) t.label()).setValue(toks[0].trim().intern());
          ((CoreLabel) t.label()).setWord(toks[0].trim().intern());
          ((CoreLabel) t.label()).setOriginalText(toks[1].trim().intern());
        } else {
          System.err.printf("%s: Cannot store morph analysis in non-CoreLabel: %s%n",this.getClass().getName(),t.label().getClass().getName());
        }
      }
    }
  }

  //Add start symbol so that the root has only one sub-state. Escape any enclosing brackets.
  //If the "tree" consists entirely of enclosing brackets e.g. ((())) then this method
  //will return null. In this case, readers e.g. PennTreeReader will try to read the next tree.
  while(tree != null && (tree.value() == null || tree.value().equals("")) && tree.numChildren() <= 1)
    tree = tree.firstChild();

  //Ensure that the tree has a top-level unary rewrite
  if(tree != null && !tree.value().equals(rootLabel))
    tree = tf.newTreeNode(rootLabel, Collections.singletonList(tree));

  return tree;
}