Java AtomicReader.terms方法代码示例

import org.apache.lucene.index.AtomicReader; //导入方法依赖的package包/类
public void testReuseDocsEnumNoReuse() throws IOException {
  Directory dir = newDirectory();
  Codec cp = TestUtil.alwaysPostingsFormat(new Lucene40RWPostingsFormat());
  RandomIndexWriter writer = new RandomIndexWriter(random(), dir,
      newIndexWriterConfig(new MockAnalyzer(random())).setCodec(cp));
  int numdocs = atLeast(20);
  createRandomIndex(numdocs, writer, random());
  writer.commit();

  DirectoryReader open = DirectoryReader.open(dir);
  for (AtomicReaderContext ctx : open.leaves()) {
    AtomicReader indexReader = ctx.reader();
    Terms terms = indexReader.terms("body");
    TermsEnum iterator = terms.iterator(null);
    IdentityHashMap<DocsEnum, Boolean> enums = new IdentityHashMap<>();
    MatchNoBits bits = new Bits.MatchNoBits(indexReader.maxDoc());
    while ((iterator.next()) != null) {
      DocsEnum docs = iterator.docs(random().nextBoolean() ? bits : new Bits.MatchNoBits(indexReader.maxDoc()), null, random().nextBoolean() ? DocsEnum.FLAG_FREQS : DocsEnum.FLAG_NONE);
      enums.put(docs, true);
    }
    
    assertEquals(terms.size(), enums.size());
  }
  IOUtils.close(writer, open, dir);
}
 

import org.apache.lucene.index.AtomicReader; //导入方法依赖的package包/类
public BaseTermsEnumTraverser(AtomicReaderContext context, Bits acceptDocs) throws IOException {
  this.context = context;
  AtomicReader reader = context.reader();
  this.acceptDocs = acceptDocs;
  this.maxDoc = reader.maxDoc();
  Terms terms = reader.terms(fieldName);
  if (terms != null)
    this.termsEnum = terms.iterator(null);
}
 

import org.apache.lucene.index.AtomicReader; //导入方法依赖的package包/类
/**
 * Takes the categories from the given taxonomy directory, and adds the
 * missing ones to this taxonomy. Additionally, it fills the given
 * {@link OrdinalMap} with a mapping from the original ordinal to the new
 * ordinal.
 */
public void addTaxonomy(Directory taxoDir, OrdinalMap map) throws IOException {
  ensureOpen();
  DirectoryReader r = DirectoryReader.open(taxoDir);
  try {
    final int size = r.numDocs();
    final OrdinalMap ordinalMap = map;
    ordinalMap.setSize(size);
    int base = 0;
    TermsEnum te = null;
    DocsEnum docs = null;
    for (final AtomicReaderContext ctx : r.leaves()) {
      final AtomicReader ar = ctx.reader();
      final Terms terms = ar.terms(Consts.FULL);
      te = terms.iterator(te);
      while (te.next() != null) {
        FacetLabel cp = new FacetLabel(FacetsConfig.stringToPath(te.term().utf8ToString()));
        final int ordinal = addCategory(cp);
        docs = te.docs(null, docs, DocsEnum.FLAG_NONE);
        ordinalMap.addMapping(docs.nextDoc() + base, ordinal);
      }
      base += ar.maxDoc(); // no deletions, so we're ok
    }
    ordinalMap.addDone();
  } finally {
    r.close();
  }
}
 

import org.apache.lucene.index.AtomicReader; //导入方法依赖的package包/类
public DocsEnum randomDocsEnum(String field, BytesRef term, List<AtomicReaderContext> readers, Bits bits) throws IOException {
  if (random().nextInt(10) == 0) {
    return null;
  }
  AtomicReader indexReader = readers.get(random().nextInt(readers.size())).reader();
  Terms terms = indexReader.terms(field);
  if (terms == null) {
    return null;
  }
  TermsEnum iterator = terms.iterator(null);
  if (iterator.seekExact(term)) {
    return iterator.docs(bits, null, random().nextBoolean() ? DocsEnum.FLAG_FREQS : DocsEnum.FLAG_NONE);
  }
  return null;
}
 

import org.apache.lucene.index.AtomicReader; //导入方法依赖的package包/类
@Override
protected Accountable createValue(AtomicReader reader, CacheKey key, boolean setDocsWithField /* ignored */)
    throws IOException {

  final int maxDoc = reader.maxDoc();

  Terms terms = reader.terms(key.field);

  final float acceptableOverheadRatio = ((Float) key.custom).floatValue();

  final PagedBytes bytes = new PagedBytes(15);

  int startTermsBPV;

  final int termCountHardLimit;
  if (maxDoc == Integer.MAX_VALUE) {
    termCountHardLimit = Integer.MAX_VALUE;
  } else {
    termCountHardLimit = maxDoc+1;
  }

  // TODO: use Uninvert?
  if (terms != null) {
    // Try for coarse estimate for number of bits; this
    // should be an underestimate most of the time, which
    // is fine -- GrowableWriter will reallocate as needed
    long numUniqueTerms = terms.size();
    if (numUniqueTerms != -1L) {
      if (numUniqueTerms > termCountHardLimit) {
        // app is misusing the API (there is more than
        // one term per doc); in this case we make best
        // effort to load what we can (see LUCENE-2142)
        numUniqueTerms = termCountHardLimit;
      }

      startTermsBPV = PackedInts.bitsRequired(numUniqueTerms);
    } else {
      startTermsBPV = 1;
    }
  } else {
    startTermsBPV = 1;
  }

  PackedLongValues.Builder termOrdToBytesOffset = PackedLongValues.monotonicBuilder(PackedInts.COMPACT);
  final GrowableWriter docToTermOrd = new GrowableWriter(startTermsBPV, maxDoc, acceptableOverheadRatio);

  int termOrd = 0;

  // TODO: use Uninvert?

  if (terms != null) {
    final TermsEnum termsEnum = terms.iterator(null);
    DocsEnum docs = null;

    while(true) {
      final BytesRef term = termsEnum.next();
      if (term == null) {
        break;
      }
      if (termOrd >= termCountHardLimit) {
        break;
      }

      termOrdToBytesOffset.add(bytes.copyUsingLengthPrefix(term));
      docs = termsEnum.docs(null, docs, DocsEnum.FLAG_NONE);
      while (true) {
        final int docID = docs.nextDoc();
        if (docID == DocIdSetIterator.NO_MORE_DOCS) {
          break;
        }
        // Store 1+ ord into packed bits
        docToTermOrd.set(docID, 1+termOrd);
      }
      termOrd++;
    }
  }

  // maybe an int-only impl?
  return new SortedDocValuesImpl(bytes.freeze(true), termOrdToBytesOffset.build(), docToTermOrd.getMutable(), termOrd);
}