Пример #1
0
  /**
   * Logically casts input to UTF32 ints then looks up the output or null if the input is not
   * accepted. FST must be INPUT_TYPE.BYTE4.
   */
  public static <T> T get(FST<T> fst, CharSequence input) throws IOException {
    assert fst.inputType == FST.INPUT_TYPE.BYTE4;

    // TODO: would be nice not to alloc this on every lookup
    final FST.Arc<T> arc = fst.getFirstArc(new FST.Arc<T>());

    int charIdx = 0;
    final int charLimit = input.length();

    // Accumulate output as we go
    final T NO_OUTPUT = fst.outputs.getNoOutput();
    T output = NO_OUTPUT;

    while (charIdx < charLimit) {
      final int utf32 = Character.codePointAt(input, charIdx);
      charIdx += Character.charCount(utf32);

      if (fst.findTargetArc(utf32, arc, arc) == null) {
        return null;
      } else if (arc.output != NO_OUTPUT) {
        output = fst.outputs.add(output, arc.output);
      }
    }

    if (fst.findTargetArc(FST.END_LABEL, arc, arc) == null) {
      return null;
    } else if (arc.output != NO_OUTPUT) {
      return fst.outputs.add(output, arc.output);
    } else {
      return output;
    }
  }
Пример #2
0
  /**
   * Looks up the output for this input, or null if the input is not accepted. FST must be
   * INPUT_TYPE.BYTE4.
   */
  public static <T> T get(FST<T> fst, IntsRef input) throws IOException {
    assert fst.inputType == FST.INPUT_TYPE.BYTE4;

    // TODO: would be nice not to alloc this on every lookup
    final FST.Arc<T> arc = fst.getFirstArc(new FST.Arc<T>());

    // Accumulate output as we go
    final T NO_OUTPUT = fst.outputs.getNoOutput();
    T output = NO_OUTPUT;
    for (int i = 0; i < input.length; i++) {
      if (fst.findTargetArc(input.ints[input.offset + i], arc, arc) == null) {
        return null;
      } else if (arc.output != NO_OUTPUT) {
        output = fst.outputs.add(output, arc.output);
      }
    }

    if (fst.findTargetArc(FST.END_LABEL, arc, arc) == null) {
      return null;
    } else if (arc.output != NO_OUTPUT) {
      return fst.outputs.add(output, arc.output);
    } else {
      return output;
    }
  }