Package org.apache.lucene.store

Examples of org.apache.lucene.store.ByteArrayDataInput


    return decompress(decompressor, compressed, originalLength);
  }

  static byte[] decompress(Decompressor decompressor, byte[] compressed, int originalLength) throws IOException {
    final BytesRef bytes = new BytesRef();
    decompressor.decompress(new ByteArrayDataInput(compressed), originalLength, 0, originalLength, bytes);
    return Arrays.copyOfRange(bytes.bytes, bytes.offset, bytes.offset + bytes.length);
  }
View Full Code Here


  }

  byte[] decompress(byte[] compressed, int originalLength, int offset, int length) throws IOException {
    Decompressor decompressor = mode.newDecompressor();
    final BytesRef bytes = new BytesRef();
    decompressor.decompress(new ByteArrayDataInput(compressed), originalLength, offset, length, bytes);
    return Arrays.copyOfRange(bytes.bytes, bytes.offset, bytes.offset + bytes.length);
  }
View Full Code Here

    final int numBytes = termsIn.readVInt();

    if (termState.bytes == null) {
      termState.bytes = new byte[ArrayUtil.oversize(numBytes, 1)];
      termState.bytesReader = new ByteArrayDataInput();
    } else if (termState.bytes.length < numBytes) {
      termState.bytes = new byte[ArrayUtil.oversize(numBytes, 1)];
    }

    termsIn.readBytes(termState.bytes, 0, numBytes);
View Full Code Here

              BytesRef output = Util.get(index, prefix);
              if (output == null) {
                out.println("      broken seek state: prefix is not final in index");
                throw new RuntimeException("seek state is broken");
              } else if (isSeekFrame && !f.isFloor) {
                final ByteArrayDataInput reader = new ByteArrayDataInput(output.bytes, output.offset, output.length);
                final long codeOrig = reader.readVLong();
                final long code = (f.fp << BlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS) | (f.hasTerms ? BlockTreeTermsWriter.OUTPUT_FLAG_HAS_TERMS:0) | (f.isFloor ? BlockTreeTermsWriter.OUTPUT_FLAG_IS_FLOOR:0);
                if (codeOrig != code) {
                  out.println("      broken seek state: output code=" + codeOrig + " doesn't match frame code=" + code);
                  throw new RuntimeException("seek state is broken");
                }
View Full Code Here

      } else {
        termsEnum = new TVTermsEnum();
      }
      termsEnum.reset(numTerms, flags, prefixLengths, suffixLengths, termFreqs, positionIndex, positions, startOffsets, lengths,
          payloadIndex, payloadBytes,
          new ByteArrayDataInput(termBytes.bytes, termBytes.offset, termBytes.length));
      return termsEnum;
    }
View Full Code Here

    final int len = termsIn.readVInt();

    // if (DEBUG) System.out.println("  SPR.readTermsBlock bytes=" + len + " ts=" + _termState);
    if (termState.bytes == null) {
      termState.bytes = new byte[ArrayUtil.oversize(len, 1)];
      termState.bytesReader = new ByteArrayDataInput();
    } else if (termState.bytes.length < len) {
      termState.bytes = new byte[ArrayUtil.oversize(len, 1)];
    }

    termsIn.readBytes(termState.bytes, 0, len);
View Full Code Here

      // Build FST:
      BytesRef previousAnalyzed = null;
      BytesRef analyzed = new BytesRef();
      BytesRef surface = new BytesRef();
      IntsRef scratchInts = new IntsRef();
      ByteArrayDataInput input = new ByteArrayDataInput();

      // Used to remove duplicate surface forms (but we
      // still index the hightest-weight one).  We clear
      // this when we see a new analyzed form, so it cannot
      // grow unbounded (at most 256 entries):
      Set<BytesRef> seenSurfaceForms = new HashSet<BytesRef>();

      int dedup = 0;
      while (reader.read(scratch)) {
        input.reset(scratch.bytes, scratch.offset, scratch.length);
        short analyzedLength = input.readShort();
        analyzed.grow(analyzedLength+2);
        input.readBytes(analyzed.bytes, 0, analyzedLength);
        analyzed.length = analyzedLength;

        long cost = input.readInt();

        surface.bytes = scratch.bytes;
        if (hasPayloads) {
          surface.length = input.readShort();
          surface.offset = input.getPosition();
        } else {
          surface.offset = input.getPosition();
          surface.length = scratch.length - surface.offset;
        }
       
        if (previousAnalyzed == null) {
          previousAnalyzed = new BytesRef();
          previousAnalyzed.copyBytes(analyzed);
          seenSurfaceForms.add(BytesRef.deepCopyOf(surface));
        } else if (analyzed.equals(previousAnalyzed)) {
          dedup++;
          if (dedup >= maxSurfaceFormsPerAnalyzedForm) {
            // More than maxSurfaceFormsPerAnalyzedForm
            // dups: skip the rest:
            continue;
          }
          if (seenSurfaceForms.contains(surface)) {
            continue;
          }
          seenSurfaceForms.add(BytesRef.deepCopyOf(surface));
        } else {
          dedup = 0;
          previousAnalyzed.copyBytes(analyzed);
          seenSurfaceForms.clear();
          seenSurfaceForms.add(BytesRef.deepCopyOf(surface));
        }

        // TODO: I think we can avoid the extra 2 bytes when
        // there is no dup (dedup==0), but we'd have to fix
        // the exactFirst logic ... which would be sort of
        // hairy because we'd need to special case the two
        // (dup/not dup)...

        // NOTE: must be byte 0 so we sort before whatever
        // is next
        analyzed.bytes[analyzed.offset+analyzed.length] = 0;
        analyzed.bytes[analyzed.offset+analyzed.length+1] = (byte) dedup;
        analyzed.length += 2;

        Util.toIntsRef(analyzed, scratchInts);
        //System.out.println("ADD: " + scratchInts + " -> " + cost + ": " + surface.utf8ToString());
        if (!hasPayloads) {
          builder.add(scratchInts, outputs.newPair(cost, BytesRef.deepCopyOf(surface)));
        } else {
          int payloadOffset = input.getPosition() + surface.length;
          int payloadLength = scratch.length - payloadOffset;
          BytesRef br = new BytesRef(surface.length + 1 + payloadLength);
          System.arraycopy(surface.bytes, surface.offset, br.bytes, 0, surface.length);
          br.bytes[surface.length] = PAYLOAD_SEP;
          System.arraycopy(scratch.bytes, payloadOffset, br.bytes, surface.length+1, payloadLength);
View Full Code Here

      this.rootCode = rootCode;
      // if (DEBUG) {
      //   System.out.println("BTTR: seg=" + segment + " field=" + fieldInfo.name + " rootBlockCode=" + rootCode + " divisor=" + indexDivisor);
      // }

      rootBlockFP = (new ByteArrayDataInput(rootCode.bytes, rootCode.offset, rootCode.length)).readVLong() >>> BlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS;

      if (indexIn != null) {
        final IndexInput clone = indexIn.clone();
        //System.out.println("start=" + indexStartFP + " field=" + fieldInfo.name);
        clone.seek(indexStartFP);
View Full Code Here

    }

    decompressor.decompress(fieldsStream, totalLength, offset, length, bytes);
    assert bytes.length == length;

    final ByteArrayDataInput documentInput = new ByteArrayDataInput(bytes.bytes, bytes.offset, bytes.length);
    for (int fieldIDX = 0; fieldIDX < numStoredFields; fieldIDX++) {
      final long infoAndBits = documentInput.readVLong();
      final int fieldNumber = (int) (infoAndBits >>> TYPE_BITS);
      final FieldInfo fieldInfo = fieldInfos.fieldInfo(fieldNumber);

      final int bits = (int) (infoAndBits & TYPE_MASK);
      assert bits <= NUMERIC_DOUBLE: "bits=" + Integer.toHexString(bits);

      switch(visitor.needsField(fieldInfo)) {
        case YES:
          readField(documentInput, visitor, fieldInfo, bits);
          assert documentInput.getPosition() <= bytes.offset + bytes.length : documentInput.getPosition() + " " + bytes.offset + bytes.length;
          break;
        case NO:
          skipField(documentInput, bits);
          assert documentInput.getPosition() <= bytes.offset + bytes.length : documentInput.getPosition() + " " + bytes.offset + bytes.length;
          break;
        case STOP:
          return;
      }
    }
    assert documentInput.getPosition() == bytes.offset + bytes.length : documentInput.getPosition() + " " + bytes.offset + " " + bytes.length;
  }
View Full Code Here

    final Arc<Long> firstArc = new Arc<Long>();
    final Arc<Long> scratchArc = new Arc<Long>();
    final IntsRef scratchInts = new IntsRef();
    final BytesRefFSTEnum<Long> fstEnum = new BytesRefFSTEnum<Long>(fst);
    final BytesRef ref = new BytesRef();
    final ByteArrayDataInput input = new ByteArrayDataInput();
    return new SortedSetDocValues() {
      long currentOrd;

      @Override
      public long nextOrd() {
        if (input.eof()) {
          return NO_MORE_ORDS;
        } else {
          currentOrd += input.readVLong();
          return currentOrd;
        }
      }
     
      @Override
      public void setDocument(int docID) {
        docToOrds.get(docID, ref);
        input.reset(ref.bytes, ref.offset, ref.length);
        currentOrd = 0;
      }

      @Override
      public void lookupOrd(long ord, BytesRef result) {
View Full Code Here

TOP

Related Classes of org.apache.lucene.store.ByteArrayDataInput

Copyright © 2018 www.massapicom. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.