java/galago/src/galago/parse/PostingsReducer.java

   1 /*
   2  * PostingsReducer
   3  *
   4  * May 5, 2007 -- Trevor Strohman
   5  *
   6  * BSD License (http://www.galagosearch.org/license)
   7  */
   8
   9 package galago.parse;
  10
  11 import galago.Sorts;
  12 import galago.tupleflow.Reducer;
  13 import galago.types.DocumentLengthWordCount;
  14 import java.io.IOException;
  15 import java.util.ArrayList;
  16 import java.util.Collections;
  17 import java.util.List;
  18
  19 /**
  20  *
  21  * @author trevor
  22  */
  23 public class PostingsReducer implements Reducer<DocumentLengthWordCount> {
  24     public ArrayList<DocumentLengthWordCount> reduce(List<DocumentLengthWordCount> input) throws IOException {
  25         Collections.sort(input, new DocumentLengthWordCount().getOrder( "+document", "+word" ).lessThan() );
  26         List<DocumentLengthWordCount> sorted = input;
  27
  28         DocumentLengthWordCount o = null;
  29         ArrayList<DocumentLengthWordCount> newList = new ArrayList<DocumentLengthWordCount>();
  30
  31         for( DocumentLengthWordCount wc : input ) {
  32             if( o != null && wc.word.equals(o.word) && wc.document.equals(o.document) ) {
  33                 o.count += wc.count;
  34             } else {
  35                 if( o != null )
  36                     newList.add(o);
  37                 o = wc;
  38             }
  39         }
  40
  41         return newList;
  42     }
  43 }