PortugueseExtPOSContextGenerator.java

/**
 * Copyright (C) 2012 cogroo <cogroo@cogroo.org>
 *
 * Licensed 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 org.cogroo.tools.postag;

import java.io.IOException;
import java.util.Arrays;
import java.util.HashSet;
import java.util.List;
import java.util.Set;

import opennlp.tools.dictionary.Dictionary;
import opennlp.tools.util.featuregen.DictionaryFeatureGenerator;
import opennlp.tools.util.featuregen.WindowFeatureGenerator;

import org.cogroo.dictionary.FeatureDictionary;
import org.cogroo.dictionary.impl.FSADictionary;
import org.cogroo.dictionary.impl.FSAFeatureDictionary;
import org.cogroo.util.PairWordPOSTag;

public class PortugueseExtPOSContextGenerator extends
    PortuguesePOSContextGenerator {

  private FSADictionary trans;
  private FeatureDictionary feat;
  private WindowFeatureGenerator dfg;

  public PortugueseExtPOSContextGenerator(Dictionary dict) {
    this(0, dict);
  }

  public PortugueseExtPOSContextGenerator(int cacheSize, Dictionary dict) {
    super(cacheSize, dict);

    // load transitivity dic
    try {
      this.trans = FSADictionary
          .createFromResources("/fsa_dictionaries/pos/pt_br_trans.dict");
      this.feat = FSAFeatureDictionary
          .createFromResources("/fsa_dictionaries/featurizer/pt_br_feats.dict");
      
      dfg = new WindowFeatureGenerator(new DictionaryFeatureGenerator("loc_prep=",
          new Dictionary(getClass().getResourceAsStream(
              "/dictionaries/pt_br/locucoes_prepositivas.xml"))), 2, 2);
          
      
    } catch (IllegalArgumentException e) {
      e.printStackTrace();
    } catch (IOException e) {
      e.printStackTrace();
    }

  }

  @Override
  protected void getContext(final int index, String[] sequence,
      String[] priorDecisions, Object[] additionalContext,
      List<String> modContext) {
    super.getContext(index, sequence, priorDecisions, additionalContext,
        modContext);
    
    // locucoes
//    this.dfg.createFeatures(modContext, sequence, index, null);
//    this.dfg.clearAdaptiveData();

    // Check transitivity
    if (index > 0)
      addTransitivity("prev_", sequence[index - 1], modContext);

    addTransitivity("", sequence[index], modContext);

    if (index < sequence.length - 1)
      addTransitivity("nxt_", sequence[index + 1], modContext);

    // specific features:
    // 'a'
    if ("a".equals(sequence[index].toLowerCase())) {

      if (index > 0 && isVerbTag(priorDecisions[index - 1])) {
        modContext.add("spec_a=v+a");
        addVerbTransitivityAndLemma("spec_a", sequence[index - 1], modContext);
      } else if (index > 1 && isVerbTag(priorDecisions[index - 2])) {
        addVerbTransitivityAndLemma2("spec_pa", sequence[index - 2],
            priorDecisions[index - 1], modContext);
      }

      if (index < sequence.length - 1) {
        String next = sequence[index + 1];
        if (next.length() > 2) {
          if (matchFeature(next, "n", "F")) {
            modContext.add("spec_af");
            if (index > 0)
              modContext.add("spec_af|" + priorDecisions[index - 1]);
          } else if (matchFeature(next, "n", "M")) {
            modContext.add("spec_am");
            if (index > 0)
              modContext.add("spec_am|" + priorDecisions[index - 1]);
          }
        }
      }
    }

    // 'que'
    if ("que".equals(sequence[index].toLowerCase())) {
      if (index < sequence.length - 1) { // at least one before and one after
        boolean nextIsVerb = false;
        boolean prevIsVerb = false;
        boolean prevprevIsVerb = false;
        
        nextIsVerb = isCanBeAVerb(sequence[index + 1]);
        if (index > 0) {
          modContext.add("spec_que_prev=" + priorDecisions[index - 1]);
          prevIsVerb = isCanBeAVerb(sequence[index - 1]);
          if (index > 1) {
            modContext.add("spec_que_pprev=" + priorDecisions[index - 2]);
            prevprevIsVerb = isCanBeAVerb(sequence[index - 2]);
          }
        }
        
        if (nextIsVerb) {
          modContext.add("spec_que=nv");
        }
        
        if(prevIsVerb) {
          modContext.add("spec_que=pv");
          modContext.add("spec_que_pv=" + priorDecisions[index - 1]);
        }
        
        if(prevprevIsVerb) {
          modContext.add("spec_que=ppv");
          modContext.add("spec_que_ppv=" + priorDecisions[index - 2]);
        }
      }
    }
  }

  private boolean matchFeature(String next, String tag, String feature) {
    String[] features = feat.getFeatures(next, null);
    if (features == null)
      features = feat.getFeatures(next.toLowerCase(), null);
    if (features != null) {
      for (String f : features) {
        if (f.contains(feature)) {
          return true;
        }
      }
    }

    return false;
  }

  private boolean isCanBeAVerb(String candidate) {
    return trans.getTags(candidate) != null;
  }

  private void addVerbTransitivityAndLemma2(String prefix, String ppVerb,
      String prevTag, List<String> modContext) {

    List<PairWordPOSTag> tagsAndLemmas = trans.getTagsAndLemms(ppVerb);
    if (tagsAndLemmas != null && tagsAndLemmas.size() > 0) {
      for (PairWordPOSTag pairWordPOSTag : tagsAndLemmas) {
        modContext.add(prefix + "_lm=" + prevTag + "|"
            + pairWordPOSTag.getWord()); // adds the lemma
        modContext.add(prefix + "_tr=" + prevTag + "|"
            + pairWordPOSTag.getPosTag()); // adds the transitivity
        modContext.add(prefix + "_lmtr=" + prevTag + "|"
            + pairWordPOSTag.getWord() + "|" + pairWordPOSTag.getPosTag()); // adds
                                                                            // the
                                                                            // transitivity
      }
    }

  }

  private void addVerbTransitivityAndLemma(String prefix, String verb,
      List<String> modContext) {
    List<PairWordPOSTag> tagsAndLemmas = trans.getTagsAndLemms(verb);
    if (tagsAndLemmas != null && tagsAndLemmas.size() > 0) {
      for (PairWordPOSTag pairWordPOSTag : tagsAndLemmas) {
        modContext.add(prefix + "_lm=" + pairWordPOSTag.getWord()); // adds the
                                                                    // lemma
        modContext.add(prefix + "_tr=" + pairWordPOSTag.getPosTag()); // adds
                                                                      // the
                                                                      // transitivity
        modContext.add(prefix + "_lmtr=" + pairWordPOSTag.getWord() + "|"
            + pairWordPOSTag.getPosTag()); // adds the transitivity
      }
    }
  }

  private boolean isVerbTag(String string) {
    return string.startsWith("v-");
  }

  private void addTransitivity(String prefix, String tok,
      List<String> modContext) {
    if (trans != null) {
      String[] tags = trans.getTags(tok);
      if (tags != null) {
        if (tags.length > 1) {
          Set<String> trans = new HashSet<String>(Arrays.asList(tags));
          for (String t : trans) {
            modContext.add(prefix + "trans=" + t);
          }
        } else {
          modContext.add(prefix + "trans=" + tags[0]);
        }
      }
    }
  }
}