AttachContextGenerator.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License. You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */


package opennlp.tools.parser.treeinsert;

import java.util.ArrayList;
import java.util.Collection;
import java.util.List;
import java.util.Set;

import opennlp.tools.parser.AbstractContextGenerator;
import opennlp.tools.parser.Cons;
import opennlp.tools.parser.Parse;

public class AttachContextGenerator extends AbstractContextGenerator {

  public AttachContextGenerator(Set<String> punctSet) {
    this.punctSet = punctSet;
  }

  public String[] getContext(Object o) {
    Object[] parts = (Object[]) o;
    return getContext((Parse[]) parts[0], (Integer) parts[1],(List<Parse>) parts[2], (Integer) parts[3]);
  }

  private boolean containsPunct(Collection<Parse> puncts, String punct) {
    if (puncts != null) {
      for (Parse p : puncts) {
        if (p.getType().equals(punct)) {
          return true;
        }
      }
    }
    return false;
  }

  /**
   *
   * @param constituents The constituents as they have been constructed so far.
   * @param index The constituent index of the node being attached.
   * @param rightFrontier The nodes which have been not attach to so far.
   * @return A set of contextual features about this attachment.
   */
  public String[] getContext(Parse[] constituents, int index, List<Parse> rightFrontier, int rfi) {
    List<String> features = new ArrayList<>(100);
    Parse fn = rightFrontier.get(rfi);
    Parse fp = null;
    if (rfi + 1 < rightFrontier.size()) {
      fp = rightFrontier.get(rfi + 1);
    }
    Parse p_1 = null;
    if (rightFrontier.size() > 0) {
      p_1 = rightFrontier.get(0);
    }
    Parse p0 = constituents[index];
    Parse p1 = null;
    if (index + 1 < constituents.length) {
      p1 = constituents[index + 1];
    }

    Collection<Parse> punct_1fs = fn.getPreviousPunctuationSet();
    Collection<Parse> punct_1s = p0.getPreviousPunctuationSet();
    Collection<Parse> punct1s = p0.getNextPunctuationSet();

    String consfp = cons(fp, -3);
    String consf = cons(fn, -2);
    String consp_1 = cons(p_1, -1);
    String consp0 = cons(p0, 0);
    String consp1 = cons(p1, 1);

    String consbofp = consbo(fp, -3);
    String consbof = consbo(fn, -2);
    String consbop_1 = consbo(p_1, -1);
    String consbop0 = consbo(p0, 0);
    String consbop1 = consbo(p1, 1);

    Cons cfp = new Cons(consfp,consbofp,-3,true);
    Cons cf = new Cons(consf,consbof,-2,true);
    Cons c_1 = new Cons(consp_1,consbop_1,-1,true);
    Cons c0 = new Cons(consp0,consbop0,0,true);
    Cons c1 = new Cons(consp1,consbop1,1,true);

    //default
    features.add("default");

    //unigrams
    features.add(consfp);
    features.add(consbofp);
    features.add(consf);
    features.add(consbof);
    features.add(consp_1);
    features.add(consbop_1);
    features.add(consp0);
    features.add(consbop0);
    features.add(consp1);
    features.add(consbop1);

    //productions
    String prod = production(fn,false);
    //String punctProd = production(fn,true,punctSet);
    features.add("pn=" + prod);
    features.add("pd=" + prod + "," + p0.getType());
    features.add("ps=" + fn.getType() + "->" + fn.getType() + "," + p0.getType());
    if (punct_1s != null) {
      StringBuilder punctBuf = new StringBuilder(5);
      for (Parse punct : punct_1s) {
        punctBuf.append(punct.getType()).append(",");
      }
      //features.add("ppd="+punctProd+","+punctBuf.toString()+p0.getType());
      //features.add("pps="+fn.getType()+"->"+fn.getType()+","+punctBuf.toString()+p0.getType());
    }

    //bi-grams
    //cons(fn),cons(0)
    cons2(features,cfp,c0,punct_1s,true);
    cons2(features,cf,c0,punct_1s,true);
    cons2(features,c_1,c0,punct_1s,true);
    cons2(features,c0,c1,punct1s,true);
    cons3(features,cf,c_1,c0,null,punct_1s,true,true,true);
    cons3(features,cf,c0,c1,punct_1s,punct1s,true,true,true);
    cons3(features,cfp,cf,c0,null,punct_1s,true,true,true);
    /*
    for (int ri=0;ri<rfi;ri++) {
      Parse jn = (Parse) rightFrontier.get(ri);
      features.add("jn="+jn.getType());
    }
    */
    int headDistance = (p0.getHeadIndex() - fn.getHeadIndex());
    features.add("hd=" + headDistance);
    features.add("nd=" + rfi);

    features.add("nd=" + p0.getType() + "." + rfi);
    features.add("hd=" + p0.getType() + "." + headDistance);
    //features.add("fs="+rightFrontier.size());
    //paired punct features
    if (containsPunct(punct_1s,"''")) {
      if (containsPunct(punct_1fs,"``")) {
        features.add("quotematch");//? not generating feature correctly

      }
    }
    return features.toArray(new String[features.size()]);
  }
}