edu.berkeley.nlp.syntax.RichLabel Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of berkeleyparser Show documentation
The Berkeley parser analyzes the grammatical structure of natural language using probabilistic context-free grammars (PCFGs).
The newest version!
package edu.berkeley.nlp.syntax;

import edu.berkeley.nlp.ling.HeadFinder;
import edu.berkeley.nlp.ling.CollinsHeadFinder;
import edu.berkeley.nlp.util.Pair;

import java.util.List;
import java.util.ArrayList;
import java.io.StringReader;

/**
 * Created by IntelliJ IDEA.
 * User: aria42
 * Date: Oct 25, 2008
 * Time: 4:04:53 PM
 */
public class RichLabel {
	private String headWord;
	private String headTag;
	private int start;
	private int stop;
	private int headIndex;
	private String label;
	private Tree origNode;

	public int getSpanSize() {
		return stop-start;	
	}

	public int getHeadIndex() {
		return headIndex;
	}

	public void setHeadIndex(int headIndex) {
		this.headIndex = headIndex;
	}

	public String getHeadWord() {
		return headWord;
	}

	public void setHeadWord(String headWord) {
		this.headWord = headWord;
	}

	public String getHeadTag() {
		return headTag;
	}

	public void setHeadTag(String headTag) {
		this.headTag = headTag;
	}

	public int getStart() {
		return start;
	}

	public void setStart(int start) {
		this.start = start;
	}

	public int getStop() {
		return stop;
	}

	public void setStop(int stop) {
		this.stop = stop;
	}

	public String getLabel() {
		return label;
	}

	public void setLabel(String label) {
		this.label = label;
	}

	public Tree getOriginalNode() {
		return origNode;
	}

	public void setOriginalNode(Tree origNode) {
		this.origNode = origNode;
	}

	@Override
	public String toString() {
		return String.format("%s(%s[%d]-%s)[%d,%d]",label,headWord,headIndex,headTag,start,stop);	
	}

  private static final CollinsHeadFinder cf = new CollinsHeadFinder();

  public static Tree getRichTree(Tree tree) {
    return getRichTree(tree, cf);    
  }


	public static Tree getRichTree(Tree tree, HeadFinder headFinder) {
		return buildRecursive(tree,headFinder, 0);
	}

	private static Pair getHeadWordTag(Tree tree, HeadFinder headFinder) {
		if (tree.isPreTerminal()) {
			Tree term = tree.getChildren().get(0);
			return Pair.newPair(term.getLabel(),tree.getLabel());
		}
		if (tree.isLeaf()) {
			return Pair.newPair(tree.getLabel(),null);
		}
		Tree head = headFinder.determineHead(tree);
		return getHeadWordTag(head,headFinder);
	}

	private static Tree buildRecursive(Tree tree, HeadFinder headFinder, int start) {
		RichLabel label = new RichLabel();
		label.setStart(start);
		label.setStop(start + tree.getYield().size());
		label.setLabel(tree.getLabel());
		label.setOriginalNode(tree);
		Pair headWordTagPair = getHeadWordTag(tree,headFinder);		
		label.setHeadWord(headWordTagPair.getFirst());
		label.setHeadTag(headWordTagPair.getSecond());
		int offset = start;
		List> richChildren = new ArrayList>();
		for (Tree child : tree.getChildren()) {
			Tree richChild = buildRecursive(child,headFinder,offset);
			richChildren.add(richChild);
			offset += child.getYield().size();
		}
		// Head Index
		if (tree.isPhrasal()) {
			Tree headChild = headFinder.determineHead(tree);
			for (Tree child: richChildren) {
			   if (child.getLabel().origNode == headChild) {
				   label.setHeadIndex(child.getLabel().getHeadIndex());
			   }
			}
		}   else {
			label.setHeadIndex(label.start);
		}
		return new Tree(label,richChildren);
	}

	public static void main(String[] args) {
		String tStr = "((S (NP (DT The) (NN man)) (VP (VBD ran) (PP (IN down) (NP (DT the) (NNS stairs))))))";
		Tree t = new Trees.PennTreeReader(new StringReader(tStr)).next();
		System.out.println("Rich Tree: " + getRichTree(t, new CollinsHeadFinder()));
	}
}