Examples with FixedBitSet org.apache.lucene.util.FixedBitSet used on opensource projects

Example 11 with FixedBitSet

use of org.apache.lucene.util.FixedBitSet in project lucene-solr by apache.

the class DocSetBase method getBits.

/**
   * Return a {@link FixedBitSet} with a bit set for every document in this
   * {@link DocSet}. The default implementation iterates on all docs and sets
   * the relevant bits. You should override if you can provide a more efficient
   * implementation.
   */
protected FixedBitSet getBits() {
    FixedBitSet bits = new FixedBitSet(size());
    for (DocIterator iter = iterator(); iter.hasNext(); ) {
        int nextDoc = iter.nextDoc();
        bits = FixedBitSet.ensureCapacity(bits, nextDoc);
        bits.set(nextDoc);
    }
    return bits;
}

Example 12 with FixedBitSet

use of org.apache.lucene.util.FixedBitSet in project lucene-solr by apache.

the class DocSetUtil method createBigSet.

private static DocSet createBigSet(List<LeafReaderContext> leaves, PostingsEnum[] postList, int maxDoc, int firstReader) throws IOException {
    long[] bits = new long[FixedBitSet.bits2words(maxDoc)];
    int sz = 0;
    for (int i = firstReader; i < postList.length; i++) {
        PostingsEnum postings = postList[i];
        if (postings == null)
            continue;
        LeafReaderContext ctx = leaves.get(i);
        Bits liveDocs = ctx.reader().getLiveDocs();
        int base = ctx.docBase;
        for (; ; ) {
            int subId = postings.nextDoc();
            if (subId == DocIdSetIterator.NO_MORE_DOCS)
                break;
            if (liveDocs != null && !liveDocs.get(subId))
                continue;
            int globalId = subId + base;
            bits[globalId >> 6] |= (1L << globalId);
            sz++;
        }
    }
    BitDocSet docSet = new BitDocSet(new FixedBitSet(bits, maxDoc), sz);
    int smallSetSize = smallSetSize(maxDoc);
    if (sz < smallSetSize) {
        // make this optional?
        DocSet smallSet = toSmallSet(docSet);
        // assert equals(docSet, smallSet);
        return smallSet;
    }
    return docSet;
}

Example 13 with FixedBitSet

use of org.apache.lucene.util.FixedBitSet in project lucene-solr by apache.

the class DocSetBase method andNot.

@Override
public DocSet andNot(DocSet other) {
    FixedBitSet newbits = getBits().clone();
    newbits.andNot(toBitSet(other));
    return new BitDocSet(newbits);
}

Example 14 with FixedBitSet

use of org.apache.lucene.util.FixedBitSet in project lucene-solr by apache.

the class DocSetBase method toBitSet.

public static FixedBitSet toBitSet(DocSet set) {
    if (set instanceof DocSetBase) {
        return ((DocSetBase) set).getBits();
    } else {
        FixedBitSet bits = new FixedBitSet(64);
        for (DocIterator iter = set.iterator(); iter.hasNext(); ) {
            int nextDoc = iter.nextDoc();
            bits = FixedBitSet.ensureCapacity(bits, nextDoc);
            bits.set(nextDoc);
        }
        return bits;
    }
}

Example 15 with FixedBitSet

use of org.apache.lucene.util.FixedBitSet in project lucene-solr by apache.

the class RandomSamplingFacetsCollector method createSample.

/** Create a sampled of the given hits. */
private MatchingDocs createSample(MatchingDocs docs) {
    int maxdoc = docs.context.reader().maxDoc();
    // TODO: we could try the WAH8DocIdSet here as well, as the results will be sparse
    FixedBitSet sampleDocs = new FixedBitSet(maxdoc);
    int binSize = (int) (1.0 / samplingRate);
    try {
        int counter = 0;
        int limit, randomIndex;
        if (leftoverBin != NOT_CALCULATED) {
            limit = leftoverBin;
            // either NOT_CALCULATED, which means we already sampled from that bin,
            // or the next document to sample
            randomIndex = leftoverIndex;
        } else {
            limit = binSize;
            randomIndex = random.nextInt(binSize);
        }
        final DocIdSetIterator it = docs.bits.iterator();
        for (int doc = it.nextDoc(); doc != DocIdSetIterator.NO_MORE_DOCS; doc = it.nextDoc()) {
            if (counter == randomIndex) {
                sampleDocs.set(doc);
            }
            counter++;
            if (counter >= limit) {
                counter = 0;
                limit = binSize;
                randomIndex = random.nextInt(binSize);
            }
        }
        if (counter == 0) {
            // we either exhausted the bin and the iterator at the same time, or
            // this segment had no results. in the latter case we might want to
            // carry leftover to the next segment as is, but that complicates the
            // code and doesn't seem so important.
            leftoverBin = leftoverIndex = NOT_CALCULATED;
        } else {
            leftoverBin = limit - counter;
            if (randomIndex > counter) {
                // the document to sample is in the next bin
                leftoverIndex = randomIndex - counter;
            } else if (randomIndex < counter) {
                // we sampled a document from the bin, so just skip over remaining
                // documents in the bin in the next segment.
                leftoverIndex = NOT_CALCULATED;
            }
        }
        return new MatchingDocs(docs.context, new BitDocIdSet(sampleDocs), docs.totalHits, null);
    } catch (IOException e) {
        throw new RuntimeException(e);
    }
}