Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
/*
* Licensed to Elasticsearch under one or more contributor
* license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright
* ownership. Elasticsearch licenses this file to you under
* the Apache License, Version 2.0 (the "License"); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package org.elasticsearch.search.aggregations.bucket.terms;
import com.google.common.base.Predicate;
import com.google.common.collect.Iterators;
import com.google.common.collect.Lists;
import com.google.common.collect.UnmodifiableIterator;
import org.apache.lucene.index.AtomicReaderContext;
import org.apache.lucene.util.BytesRef;
import org.elasticsearch.common.collect.Iterators2;
import org.elasticsearch.common.lease.Releasables;
import org.elasticsearch.common.lucene.ReaderContextAware;
import org.elasticsearch.common.util.BigArrays;
import org.elasticsearch.common.util.LongArray;
import org.elasticsearch.index.fielddata.BytesValues;
import org.elasticsearch.index.fielddata.ordinals.Ordinals;
import org.elasticsearch.search.aggregations.Aggregator;
import org.elasticsearch.search.aggregations.AggregatorFactories;
import org.elasticsearch.search.aggregations.bucket.BucketsAggregator;
import org.elasticsearch.search.aggregations.bucket.BytesRefHash;
import org.elasticsearch.search.aggregations.bucket.terms.support.BucketPriorityQueue;
import org.elasticsearch.search.aggregations.bucket.terms.support.IncludeExclude;
import org.elasticsearch.search.aggregations.support.AggregationContext;
import org.elasticsearch.search.aggregations.support.ValuesSource;
import org.elasticsearch.search.aggregations.support.bytes.BytesValuesSource;
import java.io.IOException;
import java.util.*;
/**
* An aggregator of string values.
*/
public class StringTermsAggregator extends BucketsAggregator {
private final ValuesSource valuesSource;
private final InternalOrder order;
private final int requiredSize;
private final int shardSize;
private final long minDocCount;
protected final BytesRefHash bucketOrds;
private final IncludeExclude includeExclude;
public StringTermsAggregator(String name, AggregatorFactories factories, ValuesSource valuesSource, long estimatedBucketCount,
InternalOrder order, int requiredSize, int shardSize, long minDocCount,
IncludeExclude includeExclude, AggregationContext aggregationContext, Aggregator parent) {
super(name, BucketAggregationMode.PER_BUCKET, factories, estimatedBucketCount, aggregationContext, parent);
this.valuesSource = valuesSource;
this.order = InternalOrder.validate(order, this);
this.requiredSize = requiredSize;
this.shardSize = shardSize;
this.minDocCount = minDocCount;
this.includeExclude = includeExclude;
bucketOrds = new BytesRefHash(estimatedBucketCount, aggregationContext.pageCacheRecycler());
}
@Override
public boolean shouldCollect() {
return true;
}
@Override
public void collect(int doc, long owningBucketOrdinal) throws IOException {
assert owningBucketOrdinal == 0;
final BytesValues values = valuesSource.bytesValues();
final int valuesCount = values.setDocument(doc);
for (int i = 0; i < valuesCount; ++i) {
final BytesRef bytes = values.nextValue();
if (includeExclude != null && !includeExclude.accept(bytes)) {
continue;
}
final int hash = values.currentValueHash();
assert hash == bytes.hashCode();
long bucketOrdinal = bucketOrds.add(bytes, hash);
if (bucketOrdinal < 0) { // already seen
bucketOrdinal = - 1 - bucketOrdinal;
}
collectBucket(doc, bucketOrdinal);
}
}
/** Returns an iterator over the field data terms. */
private static Iterator terms(final BytesValues.WithOrdinals bytesValues, boolean reverse) {
final Ordinals.Docs ordinals = bytesValues.ordinals();
if (reverse) {
return new UnmodifiableIterator() {
long i = ordinals.getMaxOrd() - 1;
@Override
public boolean hasNext() {
return i >= Ordinals.MIN_ORDINAL;
}
@Override
public BytesRef next() {
bytesValues.getValueByOrd(i--);
return bytesValues.copyShared();
}
};
} else {
return new UnmodifiableIterator() {
long i = Ordinals.MIN_ORDINAL;
@Override
public boolean hasNext() {
return i < ordinals.getMaxOrd();
}
@Override
public BytesRef next() {
bytesValues.getValueByOrd(i++);
return bytesValues.copyShared();
}
};
}
}
@Override
public StringTerms buildAggregation(long owningBucketOrdinal) {
assert owningBucketOrdinal == 0;
if (minDocCount == 0 && (order != InternalOrder.COUNT_DESC || bucketOrds.size() < requiredSize)) {
// we need to fill-in the blanks
List valuesWithOrdinals = Lists.newArrayList();
for (AtomicReaderContext ctx : context.searchContext().searcher().getTopReaderContext().leaves()) {
context.setNextReader(ctx);
final BytesValues values = valuesSource.bytesValues();
if (values instanceof BytesValues.WithOrdinals) {
valuesWithOrdinals.add((BytesValues.WithOrdinals) values);
} else {
// brute force
for (int docId = 0; docId < ctx.reader().maxDoc(); ++docId) {
final int valueCount = values.setDocument(docId);
for (int i = 0; i < valueCount; ++i) {
final BytesRef term = values.nextValue();
if (includeExclude == null || includeExclude.accept(term)) {
bucketOrds.add(term, values.currentValueHash());
}
}
}
}
}
// With ordinals we can be smarter and add just as many terms as necessary to the hash table
// For instance, if sorting by term asc, we only need to get the first `requiredSize` terms as other terms would
// either be excluded by the priority queue or at reduce time.
if (valuesWithOrdinals.size() > 0) {
final boolean reverse = order == InternalOrder.TERM_DESC;
Comparator comparator = BytesRef.getUTF8SortedAsUnicodeComparator();
if (reverse) {
comparator = Collections.reverseOrder(comparator);
}
Iterator[] iterators = new Iterator[valuesWithOrdinals.size()];
for (int i = 0; i < valuesWithOrdinals.size(); ++i) {
iterators[i] = terms(valuesWithOrdinals.get(i), reverse);
}
Iterator terms = Iterators2.mergeSorted(Arrays.asList(iterators), comparator, true);
if (includeExclude != null) {
terms = Iterators.filter(terms, new Predicate() {
@Override
public boolean apply(BytesRef input) {
return includeExclude.accept(input);
}
});
}
if (order == InternalOrder.COUNT_ASC) {
// let's try to find `shardSize` terms that matched no hit
// this one needs shardSize and not requiredSize because even though terms have a count of 0 here,
// they might have higher counts on other shards
for (int added = 0; added < shardSize && terms.hasNext(); ) {
if (bucketOrds.add(terms.next()) >= 0) {
++added;
}
}
} else if (order == InternalOrder.COUNT_DESC) {
// add terms until there are enough buckets
while (bucketOrds.size() < requiredSize && terms.hasNext()) {
bucketOrds.add(terms.next());
}
} else if (order == InternalOrder.TERM_ASC || order == InternalOrder.TERM_DESC) {
// add the `requiredSize` least terms
for (int i = 0; i < requiredSize && terms.hasNext(); ++i) {
bucketOrds.add(terms.next());
}
} else {
// other orders (aggregations) are not optimizable
while (terms.hasNext()) {
bucketOrds.add(terms.next());
}
}
}
}
final int size = (int) Math.min(bucketOrds.size(), shardSize);
BucketPriorityQueue ordered = new BucketPriorityQueue(size, order.comparator(this));
StringTerms.Bucket spare = null;
for (int i = 0; i < bucketOrds.size(); i++) {
if (spare == null) {
spare = new StringTerms.Bucket(new BytesRef(), 0, null);
}
bucketOrds.get(i, spare.termBytes);
spare.docCount = bucketDocCount(i);
spare.bucketOrd = i;
spare = (StringTerms.Bucket) ordered.insertWithOverflow(spare);
}
final InternalTerms.Bucket[] list = new InternalTerms.Bucket[ordered.size()];
for (int i = ordered.size() - 1; i >= 0; --i) {
final StringTerms.Bucket bucket = (StringTerms.Bucket) ordered.pop();
bucket.aggregations = bucketAggregations(bucket.bucketOrd);
list[i] = bucket;
}
return new StringTerms(name, order, requiredSize, minDocCount, Arrays.asList(list));
}
@Override
public StringTerms buildEmptyAggregation() {
return new StringTerms(name, order, requiredSize, minDocCount, Collections.emptyList());
}
@Override
public void doRelease() {
Releasables.release(bucketOrds);
}
/**
* Extension of StringTermsAggregator that caches bucket ords using terms ordinals.
*/
public static class WithOrdinals extends StringTermsAggregator implements ReaderContextAware {
private final BytesValuesSource.WithOrdinals valuesSource;
private BytesValues.WithOrdinals bytesValues;
private Ordinals.Docs ordinals;
private LongArray ordinalToBucket;
public WithOrdinals(String name, AggregatorFactories factories, BytesValuesSource.WithOrdinals valuesSource, long esitmatedBucketCount,
InternalOrder order, int requiredSize, int shardSize, long minDocCount, AggregationContext aggregationContext, Aggregator parent) {
super(name, factories, valuesSource, esitmatedBucketCount, order, requiredSize, shardSize, minDocCount, null, aggregationContext, parent);
this.valuesSource = valuesSource;
}
@Override
public void setNextReader(AtomicReaderContext reader) {
bytesValues = valuesSource.bytesValues();
ordinals = bytesValues.ordinals();
final long maxOrd = ordinals.getMaxOrd();
if (ordinalToBucket == null || ordinalToBucket.size() < maxOrd) {
if (ordinalToBucket != null) {
ordinalToBucket.release();
}
ordinalToBucket = BigArrays.newLongArray(BigArrays.overSize(maxOrd), context().pageCacheRecycler(), false);
}
ordinalToBucket.fill(0, maxOrd, -1L);
}
@Override
public void collect(int doc, long owningBucketOrdinal) throws IOException {
assert owningBucketOrdinal == 0 : "this is a per_bucket aggregator";
final int valuesCount = ordinals.setDocument(doc);
for (int i = 0; i < valuesCount; ++i) {
final long ord = ordinals.nextOrd();
long bucketOrd = ordinalToBucket.get(ord);
if (bucketOrd < 0) { // unlikely condition on a low-cardinality field
final BytesRef bytes = bytesValues.getValueByOrd(ord);
final int hash = bytesValues.currentValueHash();
assert hash == bytes.hashCode();
bucketOrd = bucketOrds.add(bytes, hash);
if (bucketOrd < 0) { // already seen in another segment
bucketOrd = - 1 - bucketOrd;
}
ordinalToBucket.set(ord, bucketOrd);
}
collectBucket(doc, bucketOrd);
}
}
@Override
public void doRelease() {
Releasables.release(bucketOrds, ordinalToBucket);
}
}
}