Java IndexReader.leaves方法代码示例

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
@Override
public Query rewrite(IndexReader reader) throws IOException {
    Query rewritten = super.rewrite(reader);
    if (rewritten != this) {
        return rewritten;
    }
    boolean hasPayloads = false;
    for (LeafReaderContext context : reader.leaves()) {
        final Terms terms = context.reader().terms(term.field());
        if (terms != null) {
            if (terms.hasPayloads()) {
                hasPayloads = true;
                break;
            }
        }
    }
    // if the terms does not exist we could return a MatchNoDocsQuery but this would break the unified highlighter
    // which rewrites query with an empty reader.
    if (hasPayloads == false) {
        return new TermQuery(term);
    }
    return this;
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
private static FixedBitSet getSeqNosSet(final IndexReader reader, final long highestSeqNo) throws IOException {
    // _seq_no are stored as doc values for the time being, so this is how we get them
    // (as opposed to using an IndexSearcher or IndexReader)
    final FixedBitSet bitSet = new FixedBitSet((int) highestSeqNo + 1);
    final List<LeafReaderContext> leaves = reader.leaves();
    if (leaves.isEmpty()) {
        return bitSet;
    }

    for (int i = 0; i < leaves.size(); i++) {
        final LeafReader leaf = leaves.get(i).reader();
        final NumericDocValues values = leaf.getNumericDocValues(SeqNoFieldMapper.NAME);
        if (values == null) {
            continue;
        }
        final Bits bits = leaf.getLiveDocs();
        for (int docID = 0; docID < leaf.maxDoc(); docID++) {
            if (bits == null || bits.get(docID)) {
                final long seqNo = values.get(docID);
                assertFalse("should not have more than one document with the same seq_no[" + seqNo + "]", bitSet.get((int) seqNo));
                bitSet.set((int) seqNo);
            }
        }
    }
    return bitSet;
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
@Override
protected void doPostCollection() throws IOException {
    IndexReader indexReader = context().searcher().getIndexReader();
    for (LeafReaderContext ctx : indexReader.leaves()) {
        Scorer childDocsScorer = childFilter.scorer(ctx);
        if (childDocsScorer == null) {
            continue;
        }
        DocIdSetIterator childDocsIter = childDocsScorer.iterator();

        final LeafBucketCollector sub = collectableSubAggregators.getLeafCollector(ctx);
        final SortedDocValues globalOrdinals = valuesSource.globalOrdinalsValues(parentType, ctx);

        // Set the scorer, since we now replay only the child docIds
        sub.setScorer(new ConstantScoreScorer(null, 1f,childDocsIter));

        final Bits liveDocs = ctx.reader().getLiveDocs();
        for (int docId = childDocsIter.nextDoc(); docId != DocIdSetIterator.NO_MORE_DOCS; docId = childDocsIter.nextDoc()) {
            if (liveDocs != null && liveDocs.get(docId) == false) {
                continue;
            }
            long globalOrdinal = globalOrdinals.getOrd(docId);
            if (globalOrdinal != -1) {
                long bucketOrd = parentOrdToBuckets.get(globalOrdinal);
                if (bucketOrd != -1) {
                    collectBucket(sub, docId, bucketOrd);
                    if (multipleBucketsPerParentOrd) {
                        long[] otherBucketOrds = parentOrdToOtherBuckets.get(globalOrdinal);
                        if (otherBucketOrds != null) {
                            for (long otherBucketOrd : otherBucketOrds) {
                                collectBucket(sub, docId, otherBucketOrd);
                            }
                        }
                    }
                }
            }
        }
    }
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
/**
 * Returns total in-heap bytes used by all suggesters.  This method has CPU cost <code>O(numIndexedFields)</code>.
 *
 * @param fieldNamePatterns if non-null, any completion field name matching any of these patterns will break out its in-heap bytes
 * separately in the returned {@link CompletionStats}
 */
public static CompletionStats completionStats(IndexReader indexReader, String ... fieldNamePatterns) {
    long sizeInBytes = 0;
    ObjectLongHashMap<String> completionFields = null;
    if (fieldNamePatterns != null  && fieldNamePatterns.length > 0) {
        completionFields = new ObjectLongHashMap<>(fieldNamePatterns.length);
    }
    for (LeafReaderContext atomicReaderContext : indexReader.leaves()) {
        LeafReader atomicReader = atomicReaderContext.reader();
        try {
            Fields fields = atomicReader.fields();
            for (String fieldName : fields) {
                Terms terms = fields.terms(fieldName);
                if (terms instanceof CompletionTerms) {
                    // TODO: currently we load up the suggester for reporting its size
                    long fstSize = ((CompletionTerms) terms).suggester().ramBytesUsed();
                    if (fieldNamePatterns != null && fieldNamePatterns.length > 0 && Regex.simpleMatch(fieldNamePatterns, fieldName)) {
                        completionFields.addTo(fieldName, fstSize);
                    }
                    sizeInBytes += fstSize;
                }
            }
        } catch (IOException ioe) {
            throw new ElasticsearchException(ioe);
        }
    }
    return new CompletionStats(sizeInBytes, completionFields == null ? null : new FieldMemoryStats(completionFields));
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
GlobalFieldData(IndexReader reader, AtomicParentChildFieldData[] fielddata, long ramBytesUsed, Map<String, OrdinalMapAndAtomicFieldData> ordinalMapPerType) {
    this.coreCacheKey = reader.getCoreCacheKey();
    this.leaves = reader.leaves();
    this.ramBytesUsed = ramBytesUsed;
    this.fielddata = fielddata;
    this.ordinalMapPerType = ordinalMapPerType;
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
private void getPrefixTerms(ObjectHashSet<Term> terms, final Term prefix, final IndexReader reader) throws IOException {
    // SlowCompositeReaderWrapper could be used... but this would merge all terms from each segment into one terms
    // instance, which is very expensive. Therefore I think it is better to iterate over each leaf individually.
    List<LeafReaderContext> leaves = reader.leaves();
    for (LeafReaderContext leaf : leaves) {
        Terms _terms = leaf.reader().terms(field);
        if (_terms == null) {
            continue;
        }

        TermsEnum termsEnum = _terms.iterator();
        TermsEnum.SeekStatus seekStatus = termsEnum.seekCeil(prefix.bytes());
        if (TermsEnum.SeekStatus.END == seekStatus) {
            continue;
        }

        for (BytesRef term = termsEnum.term(); term != null; term = termsEnum.next()) {
            if (!StringHelper.startsWith(term, prefix.bytes())) {
                break;
            }

            terms.add(new Term(field, BytesRef.deepCopyOf(term)));
            if (terms.size() >= maxExpansions) {
                return;
            }
        }
    }
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
private static int matchCount(BitSetProducer producer, IndexReader reader) throws IOException {
    int count = 0;
    for (LeafReaderContext ctx : reader.leaves()) {
        final BitSet bitSet = producer.getBitSet(ctx);
        if (bitSet != null) {
            count += bitSet.cardinality();
        }
    }
    return count;
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
@Override
protected void doPostCollection() throws IOException {
    IndexReader indexReader = context().searchContext().searcher().getIndexReader();
    for (LeafReaderContext ctx : indexReader.leaves()) {
        Scorer childDocsScorer = childFilter.scorer(ctx);
        if (childDocsScorer == null) {
            continue;
        }
        DocIdSetIterator childDocsIter = childDocsScorer.iterator();

        final LeafBucketCollector sub = collectableSubAggregators.getLeafCollector(ctx);
        final SortedDocValues globalOrdinals = valuesSource.globalOrdinalsValues(parentType, ctx);

        // Set the scorer, since we now replay only the child docIds
        sub.setScorer(ConstantScorer.create(childDocsIter, null, 1f));

        final Bits liveDocs = ctx.reader().getLiveDocs();
        for (int docId = childDocsIter.nextDoc(); docId != DocIdSetIterator.NO_MORE_DOCS; docId = childDocsIter.nextDoc()) {
            if (liveDocs != null && liveDocs.get(docId) == false) {
                continue;
            }
            long globalOrdinal = globalOrdinals.getOrd(docId);
            if (globalOrdinal != -1) {
                long bucketOrd = parentOrdToBuckets.get(globalOrdinal);
                if (bucketOrd != -1) {
                    collectBucket(sub, docId, bucketOrd);
                    if (multipleBucketsPerParentOrd) {
                        long[] otherBucketOrds = parentOrdToOtherBuckets.get(globalOrdinal);
                        if (otherBucketOrds != null) {
                            for (long otherBucketOrd : otherBucketOrds) {
                                collectBucket(sub, docId, otherBucketOrd);
                            }
                        }
                    }
                }
            }
        }
    }
}
 

import org.apache.lucene.index.IndexReader; //导入方法依赖的package包/类
public void testNestedChildrenFilter() throws Exception {
    int numParentDocs = scaledRandomIntBetween(0, 32);
    int maxChildDocsPerParent = scaledRandomIntBetween(8, 16);

    Directory dir = newDirectory();
    RandomIndexWriter writer = new RandomIndexWriter(random(), dir);
    for (int i = 0; i < numParentDocs; i++) {
        int numChildDocs = scaledRandomIntBetween(0, maxChildDocsPerParent);
        List<Document> docs = new ArrayList<>(numChildDocs + 1);
        for (int j = 0; j < numChildDocs; j++) {
            Document childDoc = new Document();
            childDoc.add(new StringField("type", "child", Field.Store.NO));
            docs.add(childDoc);
        }

        Document parenDoc = new Document();
        parenDoc.add(new StringField("type", "parent", Field.Store.NO));
        parenDoc.add(new LegacyIntField("num_child_docs", numChildDocs, Field.Store.YES));
        docs.add(parenDoc);
        writer.addDocuments(docs);
    }

    IndexReader reader = writer.getReader();
    writer.close();

    IndexSearcher searcher = new IndexSearcher(reader);
    FetchSubPhase.HitContext hitContext = new FetchSubPhase.HitContext();
    BitSetProducer parentFilter = new QueryBitSetProducer(new TermQuery(new Term("type", "parent")));
    Query childFilter = new TermQuery(new Term("type", "child"));
    int checkedParents = 0;
    final Weight parentsWeight = searcher.createNormalizedWeight(new TermQuery(new Term("type", "parent")), false);
    for (LeafReaderContext leaf : reader.leaves()) {
        DocIdSetIterator parents = parentsWeight.scorer(leaf).iterator();
        for (int parentDoc = parents.nextDoc(); parentDoc != DocIdSetIterator.NO_MORE_DOCS ; parentDoc = parents.nextDoc()) {
            int expectedChildDocs = leaf.reader().document(parentDoc).getField("num_child_docs").numericValue().intValue();
            hitContext.reset(null, leaf, parentDoc, searcher);
            NestedChildrenQuery nestedChildrenFilter = new NestedChildrenQuery(parentFilter, childFilter, hitContext);
            TotalHitCountCollector totalHitCountCollector = new TotalHitCountCollector();
            searcher.search(new ConstantScoreQuery(nestedChildrenFilter), totalHitCountCollector);
            assertThat(totalHitCountCollector.getTotalHits(), equalTo(expectedChildDocs));
            checkedParents++;
        }
    }
    assertThat(checkedParents, equalTo(numParentDocs));
    reader.close();
    dir.close();
}