POSBracketToToken.java

package LBJ2.nlp.seg;

import LBJ2.nlp.POSBracketToVector;
import LBJ2.nlp.Word;
import LBJ2.parse.LinkedVector;


/**
  * This parser expects labeled data as described in
  * {@link POSBracketToVector}, and it returns {@link Token} objects
  * representing that labeled data.
  *
  * @author Nick Rizzolo
 **/
public class POSBracketToToken extends POSBracketToVector
{
  /**
    * A pointer to the most recently returned word is kept for easy access to
    * the next word.
   **/
  public Token currentWord;


  /**
    * Initializes an instance with the named file.
    *
    * @param file The name of the file containing labeled data.
   **/
  public POSBracketToToken(String file) { super(file); }


  /** Returns the next labeled word in the data. */
  public Object next() {
    if (currentWord == null) {
      LinkedVector vector = (LinkedVector) super.next();
      while (vector != null && vector.size() == 0)
        vector = (LinkedVector) super.next();
      if (vector == null) return null;

      Word w = (Word) vector.get(0);
      Token t = currentWord = new Token(w, null, w.partOfSpeech);
      t.partOfSpeech = null;

      while (w.next != null) {
        w = (Word) w.next;
        t.next = new Token(w, t, w.partOfSpeech);
        t.partOfSpeech = null;
        t = (Token) t.next;
      }
    }

    Token result = currentWord;
    currentWord = (Token) currentWord.next;
    return result;
  }
}