All Downloads are FREE. Search and download functionalities are using the official Maven repository.

edu.stanford.nlp.parser.nndep.ArcStandard Maven / Gradle / Ivy

Go to download

Stanford CoreNLP provides a set of natural language analysis tools which can take raw English language text input and give the base forms of words, their parts of speech, whether they are names of companies, people, etc., normalize dates, times, and numeric quantities, mark up the structure of sentences in terms of phrases and word dependencies, and indicate which noun phrases refer to the same entities. It provides the foundational building blocks for higher level text understanding applications.

There is a newer version: 4.5.7
Show newest version
package edu.stanford.nlp.parser.nndep;

import edu.stanford.nlp.ling.CoreAnnotations;
import edu.stanford.nlp.trees.TreebankLanguagePack;
import edu.stanford.nlp.util.CoreMap;

import java.util.ArrayList;
import java.util.List;

/**
 * Defines an arc-standard transition-based dependency parsing system
 * (Nivre, 2004).
 *
 * @author Danqi Chen
 */
public class ArcStandard extends ParsingSystem {
  private boolean singleRoot = true;

  public ArcStandard(TreebankLanguagePack tlp, List labels, boolean verbose) {
    super(tlp, labels, verbose);
  }

  @Override
  public boolean isTerminal(Configuration c) {
    return (c.getStackSize() == 1 && c.getBufferSize() == 0);
  }

  @Override
  public void makeTransitions() {
    transitions = new ArrayList<>();

    // TODO store these as objects!
    for (String label : labels)
      transitions.add("L(" + label + ")");
    for (String label : labels)
      transitions.add("R(" + label + ")");

    transitions.add("S");
  }

  @Override
  public Configuration initialConfiguration(CoreMap s) {
    Configuration c = new Configuration(s);
    int length = s.get(CoreAnnotations.TokensAnnotation.class).size();

    // For each token, add dummy elements to the configuration's tree
    // and add the words onto the buffer
    for (int i = 1; i <= length; ++i) {
      c.tree.add(Config.NONEXIST, Config.UNKNOWN);
      c.buffer.add(i);
    }

    // Put the ROOT node on the stack
    c.stack.add(0);

    return c;
  }

  @Override
  public boolean canApply(Configuration c, String t) {
    if (t.startsWith("L") || t.startsWith("R")) {
      String label = t.substring(2, t.length() - 1);
      int h = t.startsWith("L") ? c.getStack(0) : c.getStack(1);
      if (h < 0) return false;
      if (h == 0 && !label.equals(rootLabel)) return false;
      //if (h > 0 && label.equals(rootLabel)) return false;
    }

    int nStack = c.getStackSize();
    int nBuffer = c.getBufferSize();

    if (t.startsWith("L"))
      return nStack > 2;
    else if (t.startsWith("R")) {
      if (singleRoot)
        return (nStack > 2) || (nStack == 2 && nBuffer == 0);
      else
        return nStack >= 2;
    } else
      return nBuffer > 0;
  }

  @Override
  public void apply(Configuration c, String t) {
    int w1 = c.getStack(1);
    int w2 = c.getStack(0);
    if (t.startsWith("L")) {
      c.addArc(w2, w1, t.substring(2, t.length() - 1));
      c.removeSecondTopStack();
    } else if (t.startsWith("R")) {
      c.addArc(w1, w2, t.substring(2, t.length() - 1));
      c.removeTopStack();
    } else c.shift();
  }

  // O(n) implementation
  @Override
  public String getOracle(Configuration c, DependencyTree dTree) {
    int w1 = c.getStack(1);
    int w2 = c.getStack(0);
    if (w1 > 0 && dTree.getHead(w1) == w2)
      return "L(" + dTree.getLabel(w1) + ")";
    else if (w1 >= 0 && dTree.getHead(w2) == w1 && !c.hasOtherChild(w2, dTree))
      return "R(" + dTree.getLabel(w2) + ")";
    else
      return "S";
  }

  // NOTE: unused. need to check the correctness again.
  public boolean canReach(Configuration c, DependencyTree dTree) {
    int n = c.getSentenceSize();
    for (int i = 1; i <= n; ++i)
      if (c.getHead(i) != Config.NONEXIST && c.getHead(i) != dTree.getHead(i))
        return false;

    boolean[] inBuffer = new boolean[n + 1];
    boolean[] depInList = new boolean[n + 1];

    int[] leftL = new int[n + 2];
    int[] rightL = new int[n + 2];

    for (int i = 0; i < c.getBufferSize(); ++i)
      inBuffer[c.buffer.get(i)] = true;

    int nLeft = c.getStackSize();
    for (int i = 0; i < nLeft; ++i) {
      int x = c.stack.get(i);
      leftL[nLeft - i] = x;
      if (x > 0) depInList[dTree.getHead(x)] = true;
    }

    int nRight = 1;
    rightL[nRight] = leftL[1];
    for (int i = 0; i < c.getBufferSize(); ++i) {
      boolean inList = false;
      int x = c.buffer.get(i);
      if (!inBuffer[dTree.getHead(x)] || depInList[x]) {
        rightL[++nRight] = x;
        depInList[dTree.getHead(x)] = true;
      }
    }

    int[][] g = new int[nLeft + 1][nRight + 1];
    for (int i = 1; i <= nLeft; ++i)
      for (int j = 1; j <= nRight; ++j)
        g[i][j] = -1;

    g[1][1] = leftL[1];
    for (int i = 1; i <= nLeft; ++i)
      for (int j = 1; j <= nRight; ++j)
        if (g[i][j] != -1) {
          int x = g[i][j];
          if (j < nRight && dTree.getHead(rightL[j + 1]) == x) g[i][j + 1] = x;
          if (j < nRight && dTree.getHead(x) == rightL[j + 1]) g[i][j + 1] = rightL[j + 1];
          if (i < nLeft && dTree.getHead(leftL[i + 1]) == x) g[i + 1][j] = x;
          if (i < nLeft && dTree.getHead(x) == leftL[i + 1]) g[i + 1][j] = leftL[i + 1];
        }
    return g[nLeft][nRight] != -1;
  }

  @Override
  public boolean isOracle(Configuration c, String t, DependencyTree dTree) {
    if (!canApply(c, t))
      return false;

    if (t.startsWith("L") && !dTree.getLabel(c.getStack(1)).equals(t.substring(2, t.length() - 1)))
      return false;

    if (t.startsWith("R") && !dTree.getLabel(c.getStack(0)).equals(t.substring(2, t.length() - 1)))
      return false;

    Configuration ct = new Configuration(c);
    apply(ct, t);
    return canReach(ct, dTree);
  }
}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy