Ejemplo n.º 1
0
  private static Map<String, List<String>> generate_result(Directory directory) {
    Map<String, List<String>> result_map = new HashMap<String, List<String>>();

    try {
      IndexReader reader = IndexReader.open(directory);
      TermEnum termEnum = reader.terms();
      while (termEnum.next()) {
        String termEnumString = termEnum.term().toString();
        if (termEnumString.startsWith("content:")) {
          String term = termEnumString.substring(termEnumString.lastIndexOf(":") + 1);
          TermDocs termDocs = reader.termDocs(termEnum.term());
          while (termDocs.next()) {
            Document doc = reader.document(termDocs.doc());
            String relative_path = doc.get("relative_path");

            if (result_map.containsKey(relative_path)) {
              result_map.get(relative_path).add(term + termDocs.freq());
            } else {
              result_map.put(relative_path, new ArrayList<String>());
            }
          }
        }
      }
    } catch (IOException e) {
      e.printStackTrace();
    } finally {
    }

    return result_map;
  }
Ejemplo n.º 2
0
  public void getIndexInfo(String indexdir, int freqThreshold) {
    IndexReader reader = null;

    try {
      Directory dir = FSDirectory.open(new File(indexdir));
      System.out.println(dir);
      reader = IndexReader.open(dir);

      System.out.println("document num:" + reader.numDocs());
      System.out.println("======================");

      TermEnum terms = reader.terms();
      sortedTermQueue.clear();
      maxDocNum = reader.maxDoc();
      linkMap.clear();
      termList.clear();
      while (terms.next()) {
        // System.out.print(terms.term() + "\tDocFreq:" +
        TermDocs termDocs = reader.termDocs(terms.term());
        MyTerm temp = new MyTerm(terms.term(), termDocs, maxDocNum);
        if (temp.totalFreq < freqThreshold) {
          continue;
        } /*
           * if(temp.originTrem.text().length()==1){ continue; }
           */
        linkMap.put(temp.originTrem.text(), temp);
        sortedTermQueue.add(temp);
        termList.add(temp);
      }
      System.out.println("total Size:" + sortedTermQueue.size());
      System.out.println("mapsize:" + linkMap.keySet().size());
      // System.exit(0);
      int num = 0;
      this.maxFreq = sortedTermQueue.peek().totalFreq;
      while (!sortedTermQueue.isEmpty()) {
        num++;
        System.out.println(num + ":" + sortedTermQueue.poll());
      }
      System.out.println("read index info done");
    } catch (IOException e) {
      e.printStackTrace();
    } finally {
      try {
        reader.close();

      } catch (IOException e) {
        e.printStackTrace();
      }
    }
  }