Esempio n. 1
0
 @Override
 public void writeTo(StreamOutput out) throws IOException {
   super.writeTo(out);
   out.writeInt(numDocs);
   out.writeInt(termsList.size());
   for (TermInfo ti : termsList) ti.writeTo(out);
 }
Esempio n. 2
0
  private static Stack<TermInfo> groupTokenInfos(List<TermInfo> infos) {
    List<TermInfo> tis = new ArrayList<TermInfo>();
    for (TermInfo info : infos) {
      Integer len = info.getEnd() - info.getStart();
      if (len == 1) {
        tis.add(info);
      }
    }

    Stack<TermInfo> tiStack = new Stack<TermInfo>();
    for (Integer index = 0; index < tis.size(); index++) {
      TermInfo info = tis.get(index);
      TermInfo plus = info;
      for (Integer next = index + 1; next < tis.size(); next++) {
        TermInfo nextInfo = tis.get(next);
        if (plus == null) {
          plus = nextInfo;
          continue;
        }
        plus = plus.plus(nextInfo);
        if (plus != null) {
          if (tiStack.isEmpty()) {
            tiStack.push(plus);
          } else if (!tiStack.contains(plus)) {
            tiStack.push(plus);
          }
        }
      }
    }
    return tiStack;
  }
Esempio n. 3
0
 @Override
 public void readFrom(StreamInput in) throws IOException {
   super.readFrom(in);
   numDocs = in.readInt();
   int n = in.readInt();
   termsList = new ArrayList<>();
   for (int i = 0; i < n; i++) {
     TermInfo ti = new TermInfo();
     ti.readFrom(in);
     termsList.add(ti);
   }
 }
Esempio n. 4
0
  private static String[] groupTokens(Analyzer analyzer, String input) throws IOException {
    if (Resources.debug) {
      Resources.LOGGER.debug("TokenParser:" + input);
      Resources.LOGGER.debug("Analyzer:" + analyzer.getClass());
    }
    TokenStream tokenStream = analyzer.tokenStream("input", input);
    OffsetAttribute offsetAttribute = tokenStream.addAttribute(OffsetAttribute.class);
    PositionIncrementAttribute positionIncrementAttribute =
        tokenStream.addAttribute(PositionIncrementAttribute.class);
    CharTermAttribute charTermAttribute = tokenStream.addAttribute(CharTermAttribute.class);
    TypeAttribute typeAttribute = tokenStream.addAttribute(TypeAttribute.class);
    tokenStream.reset();
    int position = 0;

    List<TermInfo> infos = new ArrayList<TermInfo>();
    while (tokenStream.incrementToken()) {
      int increment = positionIncrementAttribute.getPositionIncrement();
      if (increment > 0) {
        position = position + increment;
        if (Resources.debug) {
          Resources.LOGGER.debug(position + ":");
        }
      }

      int startOffset = offsetAttribute.startOffset();
      int endOffset = offsetAttribute.endOffset();
      String term = charTermAttribute.toString();
      TermInfo info = new TermInfo();
      info.setStart(startOffset);
      info.setEnd(endOffset);
      infos.add(info);
      if (Resources.debug) {
        Resources.LOGGER.debug(
            "["
                + term
                + "]"
                + ":("
                + startOffset
                + "-->"
                + endOffset
                + "):"
                + typeAttribute.type());
      }
    }
    tokenStream.end();
    tokenStream.close();

    Stack<TermInfo> tiStack = groupTokenInfos(infos);
    List<String> terms = new ArrayList<String>();
    while (!tiStack.isEmpty()) {
      TermInfo termInfo = tiStack.pop();
      if (termInfo.getEnd() <= input.length() && termInfo.getStart() >= 1) {
        String term = input.substring(termInfo.getStart(), termInfo.getEnd());
        terms.add(term);
      }
    }
    return terms.toArray(new String[] {});
  }