org.apache.lucene.search.SimpleCollector#org.apache.lucene.index.BinaryDocValues源码实例Demo

public void testIndexingPointsAndDocValues() throws Exception {
  FieldType type = new FieldType();
  type.setDimensions(1, 4);
  type.setDocValuesType(DocValuesType.BINARY);
  type.freeze();
  Document doc = new Document();
  byte[] packedPoint = "term".getBytes(StandardCharsets.UTF_8);
  doc.add(new BinaryPoint("field", packedPoint, type));
  MemoryIndex mi = MemoryIndex.fromDocument(doc, analyzer);
  LeafReader leafReader = mi.createSearcher().getIndexReader().leaves().get(0).reader();

  assertEquals(1, leafReader.getPointValues("field").size());
  assertArrayEquals(packedPoint, leafReader.getPointValues("field").getMinPackedValue());
  assertArrayEquals(packedPoint, leafReader.getPointValues("field").getMaxPackedValue());

  BinaryDocValues dvs = leafReader.getBinaryDocValues("field");
  assertEquals(0, dvs.nextDoc());
  assertEquals("term", dvs.binaryValue().utf8ToString());
}
 

@Override
public ShapeValues getValues(LeafReaderContext readerContext) throws IOException {
  final BinaryDocValues docValues = DocValues.getBinary(readerContext.reader(), fieldName);

  return new ShapeValues() {
    @Override
    public boolean advanceExact(int doc) throws IOException {
      return docValues.advanceExact(doc);
    }

    @Override
    public Shape value() throws IOException {
      BytesRef bytesRef = docValues.binaryValue();
      DataInputStream dataInput
          = new DataInputStream(new ByteArrayInputStream(bytesRef.bytes, bytesRef.offset, bytesRef.length));
      return binaryCodec.readShape(dataInput);
    }

  };
}
 

public BinaryDocValues getTerms(LeafReader reader, String field, float acceptableOverheadRatio) throws IOException {
  BinaryDocValues valuesIn = reader.getBinaryDocValues(field);
  if (valuesIn == null) {
    valuesIn = reader.getSortedDocValues(field);
  }

  if (valuesIn != null) {
    // Not cached here by FieldCacheImpl (cached instead
    // per-thread by SegmentReader):
    return valuesIn;
  }

  final FieldInfo info = reader.getFieldInfos().fieldInfo(field);
  if (info == null) {
    return DocValues.emptyBinary();
  } else if (info.getDocValuesType() != DocValuesType.NONE) {
    throw new IllegalStateException("Type mismatch: " + field + " was indexed as " + info.getDocValuesType());
  } else if (info.getIndexOptions() == IndexOptions.NONE) {
    return DocValues.emptyBinary();
  }

  BinaryDocValuesImpl impl = (BinaryDocValuesImpl) caches.get(BinaryDocValues.class).get(reader, new CacheKey(field, acceptableOverheadRatio));
  return impl.iterator();
}
 

@Override
public BinaryDocValues getBinary(FieldInfo field) throws IOException {
  BinaryDocValues binaryDocValues = _binaryDocValuesCache.get(field.number);
  if (binaryDocValues != null) {
    return binaryDocValues;
  }
  synchronized (_binaryDocValuesCache) {
    binaryDocValues = _binaryDocValuesCache.get(field.number);
    if (binaryDocValues != null) {
      return binaryDocValues;
    }
    binaryDocValues = newBinary(field);
    if (_cache && binaryDocValues != null) {
      _binaryDocValuesCache.put(field.number, binaryDocValues);
    }
    return binaryDocValues;
  }
}
 

@Override
public BinaryDocValues getBinaryDocValues(String field) throws IOException {
  final BinaryDocValues binaryDocValues = in.getBinaryDocValues(field);
  if (binaryDocValues == null) {
    return null;
  }
  return new BinaryDocValues() {

    @Override
    public void get(int docID, BytesRef result) {
      try {
        if (_accessControl.hasAccess(ReadType.BINARY_DOC_VALUE, docID)) {
          binaryDocValues.get(docID, result);
          return;
        }
        // Default missing value.
        result.bytes = MISSING;
        result.length = 0;
        result.offset = 0;
      } catch (IOException e) {
        throw new RuntimeException(e);
      }
    }
  };
}
 

@Test
public void testBinaryDocValues() throws IOException {
  SecureAtomicReader secureReader = getSecureReader();
  BinaryDocValues binaryDocValues = secureReader.getBinaryDocValues("bin");
  BytesRef result = new BytesRef();
  binaryDocValues.get(0, result);
  assertEquals(new BytesRef("0".getBytes()), result);

  binaryDocValues.get(1, result);
  assertEquals(new BytesRef(), result);

  binaryDocValues.get(2, result);
  assertEquals(new BytesRef("2".getBytes()), result);

  binaryDocValues.get(3, result);
  assertEquals(new BytesRef(), result);
}
 

@Override
public SortedBinaryDocValues getBytesValues() {
    try {
        final BinaryDocValues values = DocValues.getBinary(reader, field);
        final Bits docsWithField = DocValues.getDocsWithField(reader, field);
        return FieldData.singleton(values, docsWithField);
    } catch (IOException e) {
        throw new IllegalStateException("Cannot load doc values", e);
    }
}
 

@Override
protected Accountable createValue(final AtomicReader reader, CacheKey key, boolean setDocsWithField)
    throws IOException {
  int maxDoc = reader.maxDoc();

  final int[][] matrix = new int[maxDoc][];
  BinaryDocValues valuesIn = reader.getBinaryDocValues(key.field);
  if (valuesIn == null) {
    for (int i = 0; i < maxDoc; ++i) {
      matrix[i] = new int[0];
    }
    return new IntList(matrix);
  }
  for (int i = 0; i < maxDoc; ++i) {
    String str = valuesIn.get(i).utf8ToString();
    if (StringUtils.isEmpty(str)) {
      matrix[i] = new int[0];
      continue;
    }
    JSONArray array = JSON.parseArray(str);
    matrix[i] = new int[array.size()];
    for (int j = 0; j < array.size(); ++j) {
      matrix[i][j] = array.getInteger(j);
    }
  }
  return new IntList(matrix);
}
 

@Override
protected Accountable createValue(final AtomicReader reader, CacheKey key, boolean setDocsWithField)
    throws IOException {
  int maxDoc = reader.maxDoc();

  final long[][] matrix = new long[maxDoc][];
  BinaryDocValues valuesIn = reader.getBinaryDocValues(key.field);
  if (valuesIn == null) {
    for (int i = 0; i < maxDoc; ++i) {
      matrix[i] = new long[0];
    }
    return new LongList(matrix);
  }
  for (int i = 0; i < maxDoc; ++i) {
    String str = valuesIn.get(i).utf8ToString();
    if (StringUtils.isEmpty(str)) {
      matrix[i] = new long[0];
      continue;
    }
    JSONArray array = JSON.parseArray(str);
    matrix[i] = new long[array.size()];
    for (int j = 0; j < array.size(); ++j) {
      matrix[i][j] = array.getInteger(j);
    }
  }
  return new LongList(matrix);
}
 

@Override
protected Accountable createValue(final AtomicReader reader, CacheKey key, boolean setDocsWithField)
    throws IOException {
  int maxDoc = reader.maxDoc();

  final float[][] matrix = new float[maxDoc][];
  BinaryDocValues valuesIn = reader.getBinaryDocValues(key.field);
  if (valuesIn == null) {
    for (int i = 0; i < maxDoc; ++i) {
      matrix[i] = new float[0];
    }
    return new FloatList(matrix);
  }
  for (int i = 0; i < maxDoc; ++i) {
    String str = valuesIn.get(i).utf8ToString();
    if (StringUtils.isEmpty(str)) {
      matrix[i] = new float[0];
      continue;
    }
    JSONArray array = JSON.parseArray(str);
    matrix[i] = new float[array.size()];
    for (int j = 0; j < array.size(); ++j) {
      matrix[i][j] = array.getFloat(j);
    }
  }
  return new FloatList(matrix);
}
 

@Override
protected Accountable createValue(final AtomicReader reader, CacheKey key, boolean setDocsWithField)
    throws IOException {
  int maxDoc = reader.maxDoc();

  final double[][] matrix = new double[maxDoc][];
  BinaryDocValues valuesIn = reader.getBinaryDocValues(key.field);
  if (valuesIn == null) {
    for (int i = 0; i < maxDoc; ++i) {
      matrix[i] = new double[0];
    }
    return new DoubleList(matrix);
  }
  for (int i = 0; i < maxDoc; ++i) {
    String str = valuesIn.get(i).utf8ToString();
    if (StringUtils.isEmpty(str)) {
      matrix[i] = new double[0];
      continue;
    }
    JSONArray array = JSON.parseArray(str);
    matrix[i] = new double[array.size()];
    for (int j = 0; j < array.size(); ++j) {
      matrix[i][j] = array.getFloat(j);
    }
  }
  return new DoubleList(matrix);
}
 

@Override
protected Accountable createValue(final AtomicReader reader, CacheKey key, boolean setDocsWithField)
    throws IOException {
  int maxDoc = reader.maxDoc();

  final String[][] matrix = new String[maxDoc][];
  BinaryDocValues valuesIn = reader.getBinaryDocValues(key.field);
  if (valuesIn == null) {
    for (int i = 0; i < maxDoc; ++i) {
      matrix[i] = new String[0];
    }
    return new StringList(matrix);
  }
  for (int i = 0; i < maxDoc; ++i) {
    String str = valuesIn.get(i).utf8ToString();
    if (StringUtils.isEmpty(str)) {
      matrix[i] = new String[0];
      continue;
    }
    JSONArray array = JSON.parseArray(str);
    matrix[i] = new String[array.size()];
    for (int j = 0; j < array.size(); ++j) {
      matrix[i][j] = array.getString(j);
    }
  }
  return new StringList(matrix);
}
 

public void testDocValuesDoNotAffectBoostPositionsOrOffset() throws Exception {
  Document doc = new Document();
  doc.add(new BinaryDocValuesField("text", new BytesRef("quick brown fox")));
  doc.add(new TextField("text", "quick brown fox", Field.Store.NO));
  MemoryIndex mi = MemoryIndex.fromDocument(doc, analyzer, true, true);
  LeafReader leafReader = mi.createSearcher().getIndexReader().leaves().get(0).reader();
  TermsEnum tenum = leafReader.terms("text").iterator();

  assertEquals("brown", tenum.next().utf8ToString());
  PostingsEnum penum = tenum.postings(null, PostingsEnum.OFFSETS);
  assertEquals(0, penum.nextDoc());
  assertEquals(1, penum.freq());
  assertEquals(1, penum.nextPosition());
  assertEquals(6, penum.startOffset());
  assertEquals(11, penum.endOffset());

  assertEquals("fox", tenum.next().utf8ToString());
  penum = tenum.postings(penum, PostingsEnum.OFFSETS);
  assertEquals(0, penum.nextDoc());
  assertEquals(1, penum.freq());
  assertEquals(2, penum.nextPosition());
  assertEquals(12, penum.startOffset());
  assertEquals(15, penum.endOffset());

  assertEquals("quick", tenum.next().utf8ToString());
  penum = tenum.postings(penum, PostingsEnum.OFFSETS);
  assertEquals(0, penum.nextDoc());
  assertEquals(1, penum.freq());
  assertEquals(0, penum.nextPosition());
  assertEquals(0, penum.startOffset());
  assertEquals(5, penum.endOffset());

  BinaryDocValues binaryDocValues = leafReader.getBinaryDocValues("text");
  assertEquals(0, binaryDocValues.nextDoc());
  assertEquals("quick brown fox", binaryDocValues.binaryValue().utf8ToString());
}
 

private Optional<DocValues> createBinaryDocValues(int docid, String field, DocValuesType dvType)
    throws IOException {
  BinaryDocValues bvalues = IndexUtils.getBinaryDocValues(reader, field);

  if (bvalues.advanceExact(docid)) {
    DocValues dv = DocValues.of(
        dvType,
        Collections.singletonList(BytesRef.deepCopyOf(bvalues.binaryValue())),
        Collections.emptyList());
    return Optional.of(dv);
  }

  return Optional.empty();
}
 

@Override
public FunctionValues getValues(Map<Object, Object> context, LeafReaderContext readerContext) throws IOException
{
  final BinaryDocValues terms = DocValues.getBinary(readerContext.reader(), field);
  final IndexReader top = ReaderUtil.getTopLevelContext(readerContext).reader();
  Terms t = MultiTerms.getTerms(top, qfield);
  final TermsEnum termsEnum = t == null ? TermsEnum.EMPTY : t.iterator();
  
  return new IntDocValues(this) {

    int lastDocID = -1;

    @Override
    public int intVal(int doc) throws IOException {
      if (doc < lastDocID) {
        throw new IllegalArgumentException("docs were sent out-of-order: lastDocID=" + lastDocID + " vs docID=" + doc);
      }
      lastDocID = doc;
      int curDocID = terms.docID();
      if (doc > curDocID) {
        curDocID = terms.advance(doc);
      }
      if (doc == curDocID) {
        BytesRef term = terms.binaryValue();
        if (termsEnum.seekExact(term)) {
          return termsEnum.docFreq();
        }
      }
      return 0;
    }
  };
}
 

@Override
public BinaryDocValues getBinaryDocValues(String field) throws IOException {
  if (facetFields.contains(field)) {
    final OrdinalsReader ordsReader = getOrdinalsReader(field);
    return new OrdinalMappingBinaryDocValues(ordsReader.getReader(in.getContext()), in.getBinaryDocValues(field));
  } else {
    return in.getBinaryDocValues(field);
  }
}
 

private void verifyResults(Directory indexDir, Directory taxoDir) throws IOException {
  DirectoryReader indexReader = DirectoryReader.open(indexDir);
  DirectoryTaxonomyReader taxoReader = new DirectoryTaxonomyReader(taxoDir);
  IndexSearcher searcher = newSearcher(indexReader);
  
  FacetsCollector collector = new FacetsCollector();
  FacetsCollector.search(searcher, new MatchAllDocsQuery(), 10, collector);

  // tag facets
  Facets tagFacets = new FastTaxonomyFacetCounts("$tags", taxoReader, facetConfig, collector);
  FacetResult result = tagFacets.getTopChildren(10, "tag");
  for (LabelAndValue lv: result.labelValues) {
    if (VERBOSE) {
      System.out.println(lv);
    }
    assertEquals(NUM_DOCS, lv.value.intValue());
  }
  
  // id facets
  Facets idFacets = new FastTaxonomyFacetCounts(taxoReader, facetConfig, collector);
  FacetResult idResult = idFacets.getTopChildren(10, "id");
  assertEquals(NUM_DOCS, idResult.childCount);
  assertEquals(NUM_DOCS * 2, idResult.value); // each "id" appears twice
  
  BinaryDocValues bdv = MultiDocValues.getBinaryValues(indexReader, "bdv");
  BinaryDocValues cbdv = MultiDocValues.getBinaryValues(indexReader, "cbdv");
  for (int i = 0; i < indexReader.maxDoc(); i++) {
    assertEquals(i, bdv.nextDoc());
    assertEquals(i, cbdv.nextDoc());
    assertEquals(Integer.parseInt(cbdv.binaryValue().utf8ToString()), Integer.parseInt(bdv.binaryValue().utf8ToString())*2);
  }
  IOUtils.close(indexReader, taxoReader);
}
 

static GenericTermsCollector createCollectorSV(Function<BinaryDocValues> svFunction,
    ScoreMode mode) {
  
  switch (mode) {
    case None:
      return wrap(new TermsCollector.SV(svFunction));
    case Avg:
      return new SV.Avg(svFunction);
    default:
      return new SV(svFunction, mode);  
  }
}
 

@Override
public BinaryDocValues getBinary(FieldInfo field) throws IOException {
  if (merging) {
    AssertingCodec.assertThread("DocValuesProducer", creationThread);
  }
  assert field.getDocValuesType() == DocValuesType.BINARY;
  BinaryDocValues values = in.getBinary(field);
  assert values != null;
  return new AssertingLeafReader.AssertingBinaryDocValues(values, maxDoc);
}
 

private synchronized void init() {
  caches = new HashMap<>(6);
  caches.put(Long.TYPE, new LongCache(this));
  caches.put(BinaryDocValues.class, new BinaryDocValuesCache(this));
  caches.put(SortedDocValues.class, new SortedDocValuesCache(this));
  caches.put(DocTermOrds.class, new DocTermOrdsCache(this));
  caches.put(DocsWithFieldCache.class, new DocsWithFieldCache(this));
}
 

@Override
public BinaryDocValues getBinaryDocValues(String field) throws IOException {
  BinaryDocValues values = in.getBinaryDocValues(field);
  if (values != null) {
    return values;
  }
  Type v = getType(field);
  if (v == Type.BINARY) {
    return FieldCache.DEFAULT.getTerms(in, field);
  } else {
    return null;
  }
}
 

@Override
public BinaryDocValues getBinaryDocValues(String field) throws IOException {
  if (insaneField.equals(field)) {
    return null;
  } else {
    return in.getBinaryDocValues(field);
  }
}
 

private BytesRef getBytesRef(BinaryDocValues bdv, int docId)
            throws IOException {
        if (bdv != null && bdv.advance(docId) == docId) {
//        if (bdv != null && bdv.docID() < docId && bdv.advance(docId) == docId) {
//        if (bdv != null && bdv.advanceExact(docId)) {
            return bdv.binaryValue();
        }
        return new BytesRef(BytesRef.EMPTY_BYTES);
    }
 

private TreeSet<CachingSimpleResult> getReRankedResults(
            Iterator<Integer> docIterator, BinaryDocValues binaryValues,
            GlobalFeature queryFeature, GlobalFeature tmpFeature,
            int maximumHits, IndexSearcher searcher) throws IOException {

        TreeSet<CachingSimpleResult> resultScoreDocs = new TreeSet<>();
        double maxDistance = -1f;
        double tmpScore;
        BytesRef bytesRef;
        CachingSimpleResult tmpResult;
        while (docIterator.hasNext()) {
            // using DocValues to retrieve the field values ...
            int doc = docIterator.next();

//            bytesRef = binaryValues.get(doc);
            bytesRef = getBytesRef(binaryValues, doc);
            tmpFeature.setByteArrayRepresentation(bytesRef.bytes, bytesRef.offset, bytesRef.length);
            // Getting the document from the index.
            // This is the slow step based on the field compression of stored fields.
//            tmpFeature.setByteArrayRepresentation(d.getBinaryValue(name).bytes, d.getBinaryValue(name).offset, d.getBinaryValue(name).length);
            tmpScore = queryFeature.getDistance(tmpFeature);
            if (resultScoreDocs.size() < maximumHits) {
                resultScoreDocs.add(new CachingSimpleResult(tmpScore, searcher.doc(doc), doc));
                maxDistance = resultScoreDocs.last().getDistance();
            } else if (tmpScore < maxDistance) {
                // if it is nearer to the sample than at least one of the current set:
                // remove the last one ...
                tmpResult = resultScoreDocs.last();
                resultScoreDocs.remove(tmpResult);
                // set it with new values and re-insert.
                tmpResult.set(tmpScore, searcher.doc(doc), doc);
                resultScoreDocs.add(tmpResult);
                // and set our new distance border ...
                maxDistance = resultScoreDocs.last().getDistance();
            }
        }
        return resultScoreDocs;
    }
 

private BytesRef getBytesRef(BinaryDocValues bdv, int docId)
            throws IOException {
        if (bdv != null && bdv.advance(docId) == docId) {
//        if (bdv != null && bdv.docID() < docId && bdv.advance(docId) == docId) {
//        if (bdv != null && bdv.advanceExact(docId)) {
            return bdv.binaryValue();
        }
        return new BytesRef(BytesRef.EMPTY_BYTES);
    }
 

DiskDocValuesProducer(SegmentReadState state, String dataCodec, String dataExtension, String metaCodec,
    String metaExtension) throws IOException {
  String metaName = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, metaExtension);
  // read in the entries from the metadata file.
  IndexInput in = state.directory.openInput(metaName, state.context);
  boolean success = false;
  try {
    CodecUtil.checkHeader(in, metaCodec, DiskDocValuesFormat.VERSION_START, DiskDocValuesFormat.VERSION_START);
    numerics = new ConcurrentHashMap<Integer, NumericEntry>();
    ords = new ConcurrentHashMap<Integer, NumericEntry>();
    ordIndexes = new ConcurrentHashMap<Integer, NumericEntry>();
    binaries = new ConcurrentHashMap<Integer, BinaryEntry>();
    _binaryDocValuesCache = new ConcurrentHashMap<Integer, BinaryDocValues>();
    _numericDocValuesCache = new ConcurrentHashMap<Integer, NumericDocValues>();
    _sortedDocValuesCache = new ConcurrentHashMap<Integer, SortedDocValues>();
    _sortedSetDocValuesCache = new ConcurrentHashMap<Integer, SortedSetDocValues>();
    readFields(in, state.fieldInfos);
    success = true;
  } finally {
    if (success) {
      IOUtils.close(in);
    } else {
      IOUtils.closeWhileHandlingException(in);
    }
  }

  String dataName = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, dataExtension);
  data = state.directory.openInput(dataName, state.context);
  CodecUtil.checkHeader(data, dataCodec, DiskDocValuesFormat.VERSION_START, DiskDocValuesFormat.VERSION_START);
}
 

private BinaryDocValues newBinary(FieldInfo field) throws IOException {
  BinaryEntry bytes = binaries.get(field.number);
  if (bytes.minLength == bytes.maxLength) {
    return getFixedBinary(field, bytes);
  } else {
    return getVariableBinary(field, bytes);
  }
}
 

private BinaryDocValues getFixedBinary(FieldInfo field, final BinaryEntry bytes) {
  final IndexInput data = this.data.clone();

  return new LongBinaryDocValues() {

    private final ThreadValue<IndexInput> in = new ThreadValue<IndexInput>() {
      @Override
      protected IndexInput initialValue() {
        return data.clone();
      }
    };

    @Override
    public void get(long id, BytesRef result) {
      long address = bytes.offset + id * bytes.maxLength;
      try {
        IndexInput indexInput = in.get();
        indexInput.seek(address);
        // NOTE: we could have one buffer, but various consumers (e.g.
        // FieldComparatorSource)
        // assume "they" own the bytes after calling this!
        final byte[] buffer = new byte[bytes.maxLength];
        indexInput.readBytes(buffer, 0, buffer.length);
        result.bytes = buffer;
        result.offset = 0;
        result.length = buffer.length;
      } catch (IOException e) {
        throw new RuntimeException(e);
      }
    }
  };
}
 

private SortedDocValues newSortedDocValues(FieldInfo field) throws IOException {
  final int valueCount = (int) binaries.get(field.number).count;
  final BinaryDocValues binary = getBinary(field);
  Tracer trace = Trace.trace("getSorted - BlockPackedReader - create");
  final BlockPackedReader ordinals;
  try {
    NumericEntry entry = ords.get(field.number);
    IndexInput data = this.data.clone();
    data.seek(entry.offset);
    ordinals = new BlockPackedReader(data, entry.packedIntsVersion, entry.blockSize, entry.count, true);
  } finally {
    trace.done();
  }
  return new SortedDocValues() {

    @Override
    public int getOrd(int docID) {
      return (int) ordinals.get(docID);
    }

    @Override
    public void lookupOrd(int ord, BytesRef result) {
      binary.get(ord, result);
    }

    @Override
    public int getValueCount() {
      return valueCount;
    }
  };
}
 

@Override
@SuppressWarnings("rawtypes")
public FunctionValues getValues(Map context, AtomicReaderContext readerContext) throws IOException {
  final BinaryDocValues joinValues = FieldCache.DEFAULT.getTerms(readerContext.reader(), joinField, false);

  return new DoubleDocValues(this) {

    @Override
    public double doubleVal(int doc) {
      BytesRef joinValue = joinValues.get(doc);
      if (joinValue == null) {
        throw new RuntimeException("No such doc: " + doc);
      }
      Object result = results.getResult(joinValue.utf8ToString());
      if (result == null) {
        return defaultValue;
      }
      if (result instanceof Iterable) {
        Double max = null;
        for (Object object : (Iterable)result) {
          if (object != null) {
            double value = getValue(object);
            if (max == null || value > max) {
              max = value;
            }
          }
        }
        return max != null ? max : defaultValue;
      } else {
        return getValue(result);
      }
    }
    
  };
}