Java Code Examples for org.apache.lucene.util.IntsRefBuilder#get()

public static FST<?> getFST(SortedSet<String> values) throws IOException {
    final IntsRefBuilder irBuilder = new IntsRefBuilder();
    // The builder options with defaults
    FST.INPUT_TYPE inputType = FST.INPUT_TYPE.BYTE1;
    int minSuffixCount1 = 0;
    int minSuffixCount2 = 0;
    boolean doShareSuffix = true;
    boolean doShareNonSingletonNodes = true;
    int shareMaxTailLength = Integer.MAX_VALUE;
    
    boolean allowArrayArcs = true;
    int bytesPageBits = 15;
    final Outputs<Object> outputs = NoOutputs.getSingleton();
    
    // create the FST from the values
    org.apache.lucene.util.fst.Builder<Object> fstBuilder = new org.apache.lucene.util.fst.Builder<>(inputType, minSuffixCount1, minSuffixCount2,
                    doShareSuffix, doShareNonSingletonNodes, shareMaxTailLength, outputs, allowArrayArcs, bytesPageBits);
    
    for (String value : values) {
        Util.toUTF16(value, irBuilder);
        final IntsRef scratchInt = irBuilder.get();
        fstBuilder.add(scratchInt, outputs.getNoOutput());
    }
    return fstBuilder.finish();
}
 

public static boolean matchesFst(Object object, FST fst) throws IOException {
    final IntsRefBuilder irBuilder = new IntsRefBuilder();
    Util.toUTF16(object.toString(), irBuilder);
    final IntsRef ints = irBuilder.get();
    synchronized (fst) {
        return Util.get(fst, ints) != null;
    }
}
 

static IntsRef toIntsRefUTF32(String s, IntsRefBuilder ir) {
  final int charLength = s.length();
  int charIdx = 0;
  int intIdx = 0;
  ir.clear();
  while(charIdx < charLength) {
    ir.grow(intIdx+1);
    final int utf32 = s.codePointAt(charIdx);
    ir.append(utf32);
    charIdx += Character.charCount(utf32);
    intIdx++;
  }
  return ir.get();
}
 

static IntsRef toIntsRef(BytesRef br, IntsRefBuilder ir) {
  ir.grow(br.length);
  ir.clear();
  for(int i=0;i<br.length;i++) {
    ir.append(br.bytes[br.offset+i]&0xFF);
  }
  return ir.get();
}
 

/** If this automaton accepts a single input, return it.  Else, return null.
 *  The automaton must be deterministic. */
public static IntsRef getSingleton(Automaton a) {
  if (a.isDeterministic() == false) {
    throw new IllegalArgumentException("input automaton must be deterministic");
  }
  IntsRefBuilder builder = new IntsRefBuilder();
  HashSet<Integer> visited = new HashSet<>();
  int s = 0;
  Transition t = new Transition();
  while (true) {
    visited.add(s);
    if (a.isAccept(s) == false) {
      if (a.getNumTransitions(s) == 1) {
        a.getTransition(s, 0, t);
        if (t.min == t.max && !visited.contains(t.dest)) {
          builder.append(t.min);
          s = t.dest;
          continue;
        }
      }
    } else if (a.getNumTransitions(s) == 0) {
      return builder.get();
    }

    // Automaton accepts more than one string:
    return null;
  }
}
 

/** Just maps each UTF16 unit (char) to the ints in an
 *  IntsRef. */
public static IntsRef toUTF16(CharSequence s, IntsRefBuilder scratch) {
  final int charLimit = s.length();
  scratch.setLength(charLimit);
  scratch.grow(charLimit);
  for (int idx = 0; idx < charLimit; idx++) {
    scratch.setIntAt(idx, (int) s.charAt(idx));
  }
  return scratch.get();
}
 

/** Decodes the Unicode codepoints from the provided
 *  CharSequence and places them in the provided scratch
 *  IntsRef, which must not be null, returning it. */
public static IntsRef toUTF32(CharSequence s, IntsRefBuilder scratch) {
  int charIdx = 0;
  int intIdx = 0;
  final int charLimit = s.length();
  while(charIdx < charLimit) {
    scratch.grow(intIdx+1);
    final int utf32 = Character.codePointAt(s, charIdx);
    scratch.setIntAt(intIdx, utf32);
    charIdx += Character.charCount(utf32);
    intIdx++;
  }
  scratch.setLength(intIdx);
  return scratch.get();
}
 

/** Decodes the Unicode codepoints from the provided
 *  char[] and places them in the provided scratch
 *  IntsRef, which must not be null, returning it. */
public static IntsRef toUTF32(char[] s, int offset, int length, IntsRefBuilder scratch) {
  int charIdx = offset;
  int intIdx = 0;
  final int charLimit = offset + length;
  while(charIdx < charLimit) {
    scratch.grow(intIdx+1);
    final int utf32 = Character.codePointAt(s, charIdx, charLimit);
    scratch.setIntAt(intIdx, utf32);
    charIdx += Character.charCount(utf32);
    intIdx++;
  }
  scratch.setLength(intIdx);
  return scratch.get();
}
 

/** Just takes unsigned byte values from the BytesRef and
 *  converts into an IntsRef. */
public static IntsRef toIntsRef(BytesRef input, IntsRefBuilder scratch) {
  scratch.clear();
  for(int i=0;i<input.length;i++) {
    scratch.append(input.bytes[i+input.offset] & 0xFF);
  }
  return scratch.get();
}
 

private static IntsRef toIntsRef(String s) {
  IntsRefBuilder b = new IntsRefBuilder();
  for (int i = 0, cp = 0; i < s.length(); i += Character.charCount(cp)) {
    cp = s.codePointAt(i);
    b.append(cp);
  }

  return b.get();
}