jpountz commented on a change in pull request #240:
URL: https://github.com/apache/lucene/pull/240#discussion_r709829843
##########
File path:
lucene/classification/src/java/org/apache/lucene/classification/CachingNaiveBayesClassifier.java
##########
@@ -179,10 +179,8 @@ public CachingNaiveBayesClassifier(
if (query != null) {
booleanQuery.add(query, BooleanClause.Occur.MUST);
}
- TotalHitCountCollector totalHitCountCollector = new
TotalHitCountCollector();
- indexSearcher.search(booleanQuery.build(), totalHitCountCollector);
- int ret = totalHitCountCollector.getTotalHits();
+ int ret = indexSearcher.search(booleanQuery.build(), new
TotalHitCountCollectorManager());
Review comment:
We could make it even better by using `IndexSearcher#count`.
##########
File path:
lucene/classification/src/java/org/apache/lucene/classification/SimpleNaiveBayesClassifier.java
##########
@@ -276,9 +274,7 @@ private int getWordFreqForClass(String word, Term term)
throws IOException {
if (query != null) {
booleanQuery.add(query, BooleanClause.Occur.MUST);
}
- TotalHitCountCollector totalHitCountCollector = new
TotalHitCountCollector();
- indexSearcher.search(booleanQuery.build(), totalHitCountCollector);
- return totalHitCountCollector.getTotalHits();
+ return indexSearcher.search(booleanQuery.build(), new
TotalHitCountCollectorManager());
Review comment:
likewise here
##########
File path:
lucene/benchmark/src/java/org/apache/lucene/benchmark/byTask/tasks/ReadTask.java
##########
@@ -180,6 +185,7 @@ protected int withTopDocs(IndexSearcher searcher, Query q,
TopDocs hits) throws
return res;
}
+ @Deprecated
Review comment:
I suspect it would still be useful to assess the performance of a
collector but you probably have a point that if performance is the only thing
that the benchmark module reports, then there's a chance that it's not actually
used much.
I did a quick GitHub search to see how much code I could find that extends
this code:
- [`"extends ReadTask"
createCollector`](https://github.com/search?q=%22extends+ReadTask%22+createCollector&type=code):
no hits
- [`"extends SearchTask" createCollector
-SearchWithCollectorTask`](https://github.com/search?q=%22extends+SearchTask%22+createCollector+-SearchWithCollectorTask&type=code):
no hits (I had to exclude `SearchWithCollectorTask` or all Lucene forks would
be returned).
Given that the benchmark module is quite expert, I'd be ok with breaking
hard here.
##########
File path:
lucene/sandbox/src/test/org/apache/lucene/sandbox/search/TestLargeNumHitsTopDocsCollector.java
##########
@@ -159,19 +159,18 @@ public void testNoPQHitsOrder() throws IOException {
private void runNumHits(int numHits) throws IOException {
IndexSearcher searcher = newSearcher(reader);
LargeNumHitsTopDocsCollector largeCollector = new
LargeNumHitsTopDocsCollector(numHits);
- TopScoreDocCollector regularCollector =
- TopScoreDocCollector.create(numHits, null, Integer.MAX_VALUE);
+ TopScoreDocCollectorManager regularCollector =
+ new TopScoreDocCollectorManager(numHits, Integer.MAX_VALUE);
searcher.search(testQuery, largeCollector);
- searcher.search(testQuery, regularCollector);
+ TopDocs topDocs = searcher.search(testQuery, regularCollector);
Review comment:
keep it called `secondTopDocs`, I think it helped readability
##########
File path: lucene/core/src/java/org/apache/lucene/search/TopFieldCollector.java
##########
@@ -407,97 +410,14 @@ public static TopFieldCollector create(Sort sort, int
numHits, int totalHitsThre
* field is indexed both with doc values and points. In this case, there
is an assumption that
* the same data is stored in these points and doc values.
* @return a {@link TopFieldCollector} instance which will sort the results
by the sort criteria.
+ * @deprecated This method is being deprecated in favor of using the
constructor of {@link
+ * TopFieldCollectorManager} due to its support for concurrency in
IndexSearcher
*/
+ @Deprecated
public static TopFieldCollector create(
Sort sort, int numHits, FieldDoc after, int totalHitsThreshold) {
- if (totalHitsThreshold < 0) {
- throw new IllegalArgumentException(
- "totalHitsThreshold must be >= 0, got " + totalHitsThreshold);
- }
-
- return create(
- sort,
- numHits,
- after,
- HitsThresholdChecker.create(Math.max(totalHitsThreshold, numHits)),
- null /* bottomValueChecker */);
- }
-
- /**
- * Same as above with additional parameters to allow passing in the
threshold checker and the max
- * score accumulator.
- */
- static TopFieldCollector create(
Review comment:
Indeed we don't support access to pkg-private code.
##########
File path:
lucene/classification/src/java/org/apache/lucene/classification/SimpleNaiveBayesClassifier.java
##########
@@ -179,8 +178,7 @@ protected int countDocsWithClass() throws IOException {
if (query != null) {
q.add(query, BooleanClause.Occur.MUST);
}
- indexSearcher.search(q.build(), classQueryCountCollector);
- docCount = classQueryCountCollector.getTotalHits();
+ docCount = indexSearcher.search(q.build(), new
TotalHitCountCollectorManager());
Review comment:
let's simplify by using `IndexSearcher#count`?
##########
File path:
lucene/core/src/java/org/apache/lucene/search/TopFieldCollectorManager.java
##########
@@ -0,0 +1,177 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF 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.apache.lucene.search;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Collection;
+import java.util.List;
+
+/**
+ * Create a TopFieldCollectorManager which uses a shared hit counter to
maintain number of hits and
+ * a shared {@link MaxScoreAccumulator} to propagate the minimum score across
segments if the
+ * primary sort is by relevancy.
+ *
+ * <p>Note that a new collectorManager should be created for each search due
to its internal states.
+ */
+public class TopFieldCollectorManager implements
CollectorManager<TopFieldCollector, TopFieldDocs> {
+ private final Sort sort;
+ private final int numHits;
+ private final FieldDoc after;
+ private final HitsThresholdChecker hitsThresholdChecker;
+ private final MaxScoreAccumulator minScoreAcc;
+ private final List<TopFieldCollector> collectors;
+ private final boolean supportsConcurrency;
+ private boolean collectorCreated;
+
+ /**
+ * Creates a new {@link TopFieldCollectorManager} from the given arguments.
+ *
+ * <p><b>NOTE</b>: The instances returned by this method pre-allocate a full
array of length
+ * <code>numHits</code>.
+ *
+ * @param sort the sort criteria (SortFields).
+ * @param numHits the number of results to collect.
+ * @param after the previous doc after which matching docs will be collected.
+ * @param totalHitsThreshold the number of docs to count accurately. If the
query matches more
+ * than {@code totalHitsThreshold} hits then its hit count will be a
lower bound. On the other
+ * hand if the query matches less than or exactly {@code
totalHitsThreshold} hits then the hit
+ * count of the result will be accurate. {@link Integer#MAX_VALUE} may
be used to make the hit
+ * count accurate, but this will also make query processing slower.
+ * @param supportsConcurrency to use thread-safe and slower internal states
for count tracking.
+ */
+ public TopFieldCollectorManager(
+ Sort sort, int numHits, FieldDoc after, int totalHitsThreshold, boolean
supportsConcurrency) {
+ if (totalHitsThreshold < 0) {
+ throw new IllegalArgumentException(
+ "totalHitsThreshold must be >= 0, got " + totalHitsThreshold);
+ }
+
+ this.sort = sort;
+ this.numHits = numHits;
+ this.after = after;
+ this.supportsConcurrency = supportsConcurrency;
+ this.hitsThresholdChecker =
+ supportsConcurrency
+ ? HitsThresholdChecker.createShared(Math.max(totalHitsThreshold,
numHits))
+ : HitsThresholdChecker.create(Math.max(totalHitsThreshold,
numHits));
+ this.minScoreAcc = supportsConcurrency ? new MaxScoreAccumulator() : null;
+ this.collectors = new ArrayList<>();
+ }
+
+ /**
+ * Creates a new {@link TopFieldCollectorManager} from the given arguments,
with thread-safe
+ * internal states.
+ *
+ * <p><b>NOTE</b>: The instances returned by this method pre-allocate a full
array of length
+ * <code>numHits</code>.
+ *
+ * @param sort the sort criteria (SortFields).
+ * @param numHits the number of results to collect.
+ * @param after the previous doc after which matching docs will be collected.
+ * @param totalHitsThreshold the number of docs to count accurately. If the
query matches more
+ * than {@code totalHitsThreshold} hits then its hit count will be a
lower bound. On the other
+ * hand if the query matches less than or exactly {@code
totalHitsThreshold} hits then the hit
+ * count of the result will be accurate. {@link Integer#MAX_VALUE} may
be used to make the hit
+ * count accurate, but this will also make query processing slower.
+ */
+ public TopFieldCollectorManager(Sort sort, int numHits, FieldDoc after, int
totalHitsThreshold) {
+ this(sort, numHits, after, totalHitsThreshold, true);
+ }
+
+ /**
+ * Creates a new {@link TopFieldCollectorManager} from the given arguments,
with thread-safe
+ * internal states.
+ *
+ * <p><b>NOTE</b>: The instances returned by this method pre-allocate a full
array of length
+ * <code>numHits</code>.
+ *
+ * @param sort the sort criteria (SortFields).
+ * @param numHits the number of results to collect.
+ * @param totalHitsThreshold the number of docs to count accurately. If the
query matches more
+ * than {@code totalHitsThreshold} hits then its hit count will be a
lower bound. On the other
+ * hand if the query matches less than or exactly {@code
totalHitsThreshold} hits then the hit
+ * count of the result will be accurate. {@link Integer#MAX_VALUE} may
be used to make the hit
+ * count accurate, but this will also make query processing slower.
+ */
+ public TopFieldCollectorManager(Sort sort, int numHits, int
totalHitsThreshold) {
+ this(sort, numHits, null, totalHitsThreshold, true);
+ }
+
+ @Override
+ public TopFieldCollector newCollector() {
+ if (collectorCreated && supportsConcurrency == false) {
+ throw new IllegalStateException(
+ "The instantiated TopFieldCollectorManager does not support
concurrency, but multiple collectors are being created");
+ } else {
+ collectorCreated = true;
+ }
+
+ if (sort.fields.length == 0) {
+ throw new IllegalArgumentException("Sort must contain at least one
field");
+ }
Review comment:
Can we move as much validation as possible to the constructor, so that
we would detect inconsistencies earlier: when users create the collector
manager rather than when Lucene starts collecting hits?
##########
File path:
lucene/core/src/java/org/apache/lucene/search/TopFieldCollectorManager.java
##########
@@ -0,0 +1,177 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF 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.apache.lucene.search;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Collection;
+import java.util.List;
+
+/**
+ * Create a TopFieldCollectorManager which uses a shared hit counter to
maintain number of hits and
+ * a shared {@link MaxScoreAccumulator} to propagate the minimum score across
segments if the
+ * primary sort is by relevancy.
+ *
+ * <p>Note that a new collectorManager should be created for each search due
to its internal states.
+ */
+public class TopFieldCollectorManager implements
CollectorManager<TopFieldCollector, TopFieldDocs> {
+ private final Sort sort;
+ private final int numHits;
+ private final FieldDoc after;
+ private final HitsThresholdChecker hitsThresholdChecker;
+ private final MaxScoreAccumulator minScoreAcc;
+ private final List<TopFieldCollector> collectors;
+ private final boolean supportsConcurrency;
+ private boolean collectorCreated;
+
+ /**
+ * Creates a new {@link TopFieldCollectorManager} from the given arguments.
+ *
+ * <p><b>NOTE</b>: The instances returned by this method pre-allocate a full
array of length
+ * <code>numHits</code>.
+ *
+ * @param sort the sort criteria (SortFields).
+ * @param numHits the number of results to collect.
+ * @param after the previous doc after which matching docs will be collected.
+ * @param totalHitsThreshold the number of docs to count accurately. If the
query matches more
+ * than {@code totalHitsThreshold} hits then its hit count will be a
lower bound. On the other
+ * hand if the query matches less than or exactly {@code
totalHitsThreshold} hits then the hit
+ * count of the result will be accurate. {@link Integer#MAX_VALUE} may
be used to make the hit
+ * count accurate, but this will also make query processing slower.
+ * @param supportsConcurrency to use thread-safe and slower internal states
for count tracking.
+ */
+ public TopFieldCollectorManager(
+ Sort sort, int numHits, FieldDoc after, int totalHitsThreshold, boolean
supportsConcurrency) {
+ if (totalHitsThreshold < 0) {
+ throw new IllegalArgumentException(
+ "totalHitsThreshold must be >= 0, got " + totalHitsThreshold);
+ }
+
+ this.sort = sort;
+ this.numHits = numHits;
+ this.after = after;
+ this.supportsConcurrency = supportsConcurrency;
+ this.hitsThresholdChecker =
+ supportsConcurrency
+ ? HitsThresholdChecker.createShared(Math.max(totalHitsThreshold,
numHits))
+ : HitsThresholdChecker.create(Math.max(totalHitsThreshold,
numHits));
+ this.minScoreAcc = supportsConcurrency ? new MaxScoreAccumulator() : null;
+ this.collectors = new ArrayList<>();
+ }
+
+ /**
+ * Creates a new {@link TopFieldCollectorManager} from the given arguments,
with thread-safe
+ * internal states.
+ *
+ * <p><b>NOTE</b>: The instances returned by this method pre-allocate a full
array of length
+ * <code>numHits</code>.
+ *
+ * @param sort the sort criteria (SortFields).
+ * @param numHits the number of results to collect.
+ * @param after the previous doc after which matching docs will be collected.
+ * @param totalHitsThreshold the number of docs to count accurately. If the
query matches more
+ * than {@code totalHitsThreshold} hits then its hit count will be a
lower bound. On the other
+ * hand if the query matches less than or exactly {@code
totalHitsThreshold} hits then the hit
+ * count of the result will be accurate. {@link Integer#MAX_VALUE} may
be used to make the hit
+ * count accurate, but this will also make query processing slower.
+ */
+ public TopFieldCollectorManager(Sort sort, int numHits, FieldDoc after, int
totalHitsThreshold) {
+ this(sort, numHits, after, totalHitsThreshold, true);
+ }
+
+ /**
+ * Creates a new {@link TopFieldCollectorManager} from the given arguments,
with thread-safe
+ * internal states.
+ *
+ * <p><b>NOTE</b>: The instances returned by this method pre-allocate a full
array of length
+ * <code>numHits</code>.
+ *
+ * @param sort the sort criteria (SortFields).
+ * @param numHits the number of results to collect.
+ * @param totalHitsThreshold the number of docs to count accurately. If the
query matches more
+ * than {@code totalHitsThreshold} hits then its hit count will be a
lower bound. On the other
+ * hand if the query matches less than or exactly {@code
totalHitsThreshold} hits then the hit
+ * count of the result will be accurate. {@link Integer#MAX_VALUE} may
be used to make the hit
+ * count accurate, but this will also make query processing slower.
+ */
+ public TopFieldCollectorManager(Sort sort, int numHits, int
totalHitsThreshold) {
+ this(sort, numHits, null, totalHitsThreshold, true);
+ }
+
+ @Override
+ public TopFieldCollector newCollector() {
+ if (collectorCreated && supportsConcurrency == false) {
+ throw new IllegalStateException(
+ "The instantiated TopFieldCollectorManager does not support
concurrency, but multiple collectors are being created");
+ } else {
+ collectorCreated = true;
+ }
+
+ if (sort.fields.length == 0) {
+ throw new IllegalArgumentException("Sort must contain at least one
field");
+ }
+
+ if (numHits <= 0) {
+ throw new IllegalArgumentException(
+ "numHits must be > 0; please use TotalHitCountCollector if you just
need the total hit count");
+ }
+
+ FieldValueHitQueue<FieldValueHitQueue.Entry> queue =
+ FieldValueHitQueue.create(sort.fields, numHits);
+
+ TopFieldCollector collector;
+ if (after == null) {
+ collector =
+ new TopFieldCollector.SimpleFieldCollector(
+ sort, queue, numHits, hitsThresholdChecker, minScoreAcc);
+ } else {
+ if (after.fields == null) {
+ throw new IllegalArgumentException(
+ "after.fields wasn't set; you must pass fillFields=true for the
previous search");
+ }
+
+ if (after.fields.length != sort.getSort().length) {
+ throw new IllegalArgumentException(
+ "after.fields has "
+ + after.fields.length
+ + " values but sort has "
+ + sort.getSort().length);
+ }
+
+ collector =
+ new TopFieldCollector.PagingFieldCollector(
+ sort, queue, after, numHits, hitsThresholdChecker, minScoreAcc);
+ }
+
+ collectors.add(collector);
+ return collector;
+ }
+
+ @Override
+ public TopFieldDocs reduce(Collection<TopFieldCollector> collectors) throws
IOException {
+ final TopFieldDocs[] topDocs = new TopFieldDocs[collectors.size()];
+ int i = 0;
+ for (TopFieldCollector collector : collectors) {
+ topDocs[i++] = collector.topDocs();
+ }
+ return TopDocs.merge(sort, 0, numHits, topDocs);
+ }
+
+ public List<TopFieldCollector> getCollectors() {
Review comment:
Can we make it pkg-private?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
