Introduce support for Reciprocal Rank Fusion (combining queries) #2489
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,167 @@ | ||
| /* | ||
| * 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.solr.search.combining; | ||
|
|
||
| import java.io.IOException; | ||
| import java.util.ArrayList; | ||
| import java.util.List; | ||
| import java.util.Objects; | ||
| import org.apache.lucene.search.BooleanClause.Occur; | ||
| import org.apache.lucene.search.MatchNoDocsQuery; | ||
| import org.apache.lucene.search.Query; | ||
| import org.apache.lucene.search.QueryVisitor; | ||
| import org.apache.solr.common.params.CombinerParams; | ||
| import org.apache.solr.common.params.SolrParams; | ||
| import org.apache.solr.common.util.NamedList; | ||
| import org.apache.solr.common.util.SimpleOrderedMap; | ||
| import org.apache.solr.handler.component.QueryComponent; | ||
| import org.apache.solr.request.SolrQueryRequest; | ||
| import org.apache.solr.search.ExtendedQueryBase; | ||
| import org.apache.solr.search.QParser; | ||
| import org.apache.solr.search.QParserPlugin; | ||
| import org.apache.solr.search.QueryCommand; | ||
| import org.apache.solr.search.QueryParsing; | ||
| import org.apache.solr.search.QueryResult; | ||
| import org.apache.solr.search.SolrIndexSearcher; | ||
| import org.apache.solr.search.SyntaxError; | ||
|
|
||
| public class CombineQParserPlugin extends QParserPlugin { | ||
|
There was a problem hiding this comment. This source file captures the essential aspect of my proposal. Needs another iteration of refinement at least (e.g. javadocs, and param naming / defaulting). Maybe more. |
||
| public static final String NAME = "combine"; | ||
|
|
||
| @Override | ||
| public QParser createParser( | ||
| String qstr, SolrParams localParams, SolrParams params, SolrQueryRequest req) { | ||
| return new CombineQParser(qstr, localParams, params, req); | ||
| } | ||
|
|
||
| static class CombineQParser extends QParser { | ||
|
|
||
| private List<String> unparsedQueries; | ||
| private QueriesCombiner queriesCombiningStrategy; | ||
| private List<QParser> queryParsers; | ||
| private List<Query> queries; | ||
|
|
||
| public CombineQParser( | ||
| String qstr, SolrParams localParams, SolrParams params, SolrQueryRequest req) { | ||
| super(qstr, localParams, params, req); | ||
| } | ||
|
|
||
| @Override | ||
| public Query parse() throws SyntaxError { | ||
|
|
||
| var queriesToCombineKeys = localParams.getParams("keys"); | ||
|
There was a problem hiding this comment. I am not a huge fan of var as in my opinion it reduces readability, but if it's ok as stadnard in Solr, no strong opinion |
||
| if (queriesToCombineKeys == null) { | ||
| queriesToCombineKeys = params.getParams(CombinerParams.COMBINER_KEYS); | ||
| } | ||
|
|
||
| unparsedQueries = new ArrayList<>(queriesToCombineKeys.length); | ||
| queryParsers = new ArrayList<>(queriesToCombineKeys.length); | ||
| queries = new ArrayList<>(queriesToCombineKeys.length); | ||
|
|
||
| // nocommit blend localParams and params without needless suffix in local | ||
| var blendParams = SolrParams.wrapDefaults(localParams, params); | ||
| queriesCombiningStrategy = QueriesCombiner.getImplementation(blendParams); | ||
|
|
||
| String defType = blendParams.get(QueryParsing.DEFTYPE, DEFAULT_QTYPE); | ||
|
There was a problem hiding this comment. Maybe could support different |
||
|
|
||
| for (String queryKey : queriesToCombineKeys) { | ||
| final var unparsedQuery = blendParams.get(queryKey); | ||
| final var parser = QParser.getParser(unparsedQuery, defType, req); | ||
| var query = parser.getQuery(); | ||
| if (query == null) { // sad this can happen | ||
| query = new MatchNoDocsQuery(); | ||
| } | ||
| unparsedQueries.add(unparsedQuery); | ||
| queryParsers.add(parser); | ||
| queries.add(query); | ||
| } | ||
|
|
||
| return new CombineQuery(queriesCombiningStrategy, queries); | ||
| } | ||
|
|
||
| @Override | ||
| public void addDebugInfo(NamedList<Object> dbg) { | ||
|
There was a problem hiding this comment. maybe 'addQueryDebugInfo' ? |
||
| super.addDebugInfo(dbg); | ||
|
|
||
| NamedList<NamedList<Object>> queriesDebug = new SimpleOrderedMap<>(); | ||
| String[] queryKeys = req.getParams().getParams(CombinerParams.COMBINER_KEYS); | ||
| for (int i = 0; i < queries.size(); i++) { | ||
| NamedList<Object> singleQueryDebug = new SimpleOrderedMap<>(); | ||
| singleQueryDebug.add("querystring", unparsedQueries.get(i)); | ||
| singleQueryDebug.add("queryparser", queryParsers.get(i).getClass().getSimpleName()); | ||
| singleQueryDebug.add("parsedquery", QueryParsing.toString(queries.get(i), req.getSchema())); | ||
| singleQueryDebug.add("parsedquery_toString", queries.get(i).toString()); | ||
| queriesDebug.add(queryKeys[i], singleQueryDebug); | ||
| } | ||
| dbg.add("queriesToCombine", queriesDebug); | ||
| } | ||
| } | ||
|
|
||
| static final class CombineQuery extends ExtendedQueryBase | ||
|
There was a problem hiding this comment. A "SelfExecutingQuery" isn't normal; Lucene isn't going to work with this thing since it doesn't implement getWeight nor can it. But users don't need to know/care. |
||
| implements QueryComponent.SelfExecutingQuery { | ||
|
|
||
| private final QueriesCombiner combiner; | ||
| private final List<Query> queries; | ||
|
|
||
| public CombineQuery(QueriesCombiner combiner, List<Query> queries) { | ||
| this.combiner = combiner; | ||
| this.queries = queries; | ||
| } | ||
|
|
||
| @Override | ||
| public String toString(String field) { | ||
| return super.toString(field) + "{!" + NAME + "}"; // TODO others | ||
| } | ||
|
|
||
| @Override | ||
| public boolean equals(Object o) { | ||
| if (this == o) return true; | ||
| if (!(o instanceof CombineQuery)) return false; | ||
| CombineQuery that = (CombineQuery) o; | ||
| return Objects.equals(combiner, that.combiner) && Objects.equals(queries, that.queries); | ||
| } | ||
|
|
||
| @Override | ||
| public int hashCode() { | ||
| return Objects.hash(combiner, queries); | ||
| } | ||
|
|
||
| @Override | ||
| public void visit(QueryVisitor visitor) { | ||
| for (Query query : queries) { | ||
| query.visit(visitor.getSubVisitor(Occur.MUST, this)); | ||
| } | ||
| } | ||
|
|
||
| @Override | ||
| public QueryResult search(SolrIndexSearcher searcher, QueryCommand cmd) throws IOException { | ||
| QueryResult[] results = new QueryResult[queries.size()]; | ||
| // TODO do in multiple threads? | ||
| for (int i = 0; i < queries.size(); i++) { | ||
| cmd.setQuery(queries.get(i)); | ||
| QueryResult qr = new QueryResult(); | ||
| searcher.search(qr, cmd); | ||
| results[i] = qr; | ||
| } | ||
|
|
||
| // nocommit but how is the docSet (e.g. for faceting) or maybe other things supported? | ||
|
There was a problem hiding this comment. This seems like a fundamental limitation in the whole feature honestly. What Hossman said about re-rank query is interesting... not sure if it addresses this. There was a problem hiding this comment. mmm probably I have lost in the comments why the way I created the docSet was incorrect. Can you elaborate? (a nice occasion to learn how that docSet part works, as I didn't have time yet) |
||
|
|
||
| return combiner.combine(results); | ||
| } | ||
| } | ||
| } | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,93 @@ | ||
| /* | ||
| * 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.solr.search.combining; | ||
|
|
||
| import static org.apache.solr.common.params.CombinerParams.RECIPROCAl_RANK_FUSION; | ||
|
|
||
| import java.io.IOException; | ||
| import java.util.List; | ||
| import org.apache.lucene.search.Explanation; | ||
| import org.apache.lucene.search.Query; | ||
| import org.apache.solr.common.SolrException; | ||
| import org.apache.solr.common.params.CombinerParams; | ||
| import org.apache.solr.common.params.SolrParams; | ||
| import org.apache.solr.common.util.NamedList; | ||
| import org.apache.solr.schema.IndexSchema; | ||
| import org.apache.solr.search.DocList; | ||
| import org.apache.solr.search.QueryResult; | ||
| import org.apache.solr.search.SolrIndexSearcher; | ||
|
|
||
| /** | ||
| * Combining considers two or more query rankedLists: resultA, resultB ...<br> | ||
| * For a given query, each query result is a ranked list of documents La = (a1,a2,...), Lb = (b1, | ||
| * b2, ...)...<br> | ||
| * A combining algorithm creates a unique ranked list I = (i1, i2, ...).<br> | ||
| * This list is created by combining elements from the lists la and lb as described by the | ||
| * implementation algorithm.<br> | ||
| * since @Version 9.7 Used by {@link org.apache.solr.handler.component.QueryComponent} | ||
| */ | ||
| public abstract class QueriesCombiner { | ||
dsmiley marked this conversation as resolved.
Show resolved
Hide resolved
|
||
|
|
||
| protected int upTo; | ||
|
|
||
| public QueriesCombiner(SolrParams requestParams) { | ||
| this.upTo = | ||
| requestParams.getInt(CombinerParams.COMBINER_UP_TO, CombinerParams.COMBINER_UP_TO_DEFAULT); | ||
| } | ||
|
|
||
| public abstract QueryResult combine(QueryResult[] rankedLists); | ||
|
|
||
| protected QueryResult initCombinedResult(QueryResult[] rankedLists) { | ||
| QueryResult combinedRankedList = new QueryResult(); | ||
| boolean partialResults = false; | ||
| for (QueryResult result : rankedLists) { | ||
| partialResults |= result.isPartialResults(); | ||
| } | ||
| combinedRankedList.setPartialResults(partialResults); | ||
|
|
||
| boolean segmentTerminatedEarly = false; | ||
| for (QueryResult result : rankedLists) { | ||
| if (result.getSegmentTerminatedEarly() != null) { | ||
| segmentTerminatedEarly |= result.getSegmentTerminatedEarly(); | ||
| } | ||
| } | ||
| combinedRankedList.setSegmentTerminatedEarly(segmentTerminatedEarly); | ||
|
|
||
| combinedRankedList.setNextCursorMark(rankedLists[0].getNextCursorMark()); | ||
|
There was a problem hiding this comment. Not sure about picking the first list here and/or what does use of a cursor mark mean when combining queries. Alternative might be to just disallow cursor marks when combining queries. |
||
| return combinedRankedList; | ||
| } | ||
|
|
||
| public abstract NamedList<Explanation> getExplanations( | ||
| String[] queryKeys, | ||
| List<Query> queries, | ||
| List<DocList> resultsPerQuery, | ||
| SolrIndexSearcher searcher, | ||
| IndexSchema schema) | ||
| throws IOException; | ||
|
|
||
| public static QueriesCombiner getImplementation(SolrParams requestParams) { | ||
| String algorithm = requestParams.get(CombinerParams.COMBINER_ALGORITHM, RECIPROCAl_RANK_FUSION); | ||
| switch (algorithm) { | ||
| case RECIPROCAl_RANK_FUSION: | ||
| return new ReciprocalRankFusion(requestParams); | ||
| default: | ||
| throw new SolrException( | ||
| SolrException.ErrorCode.BAD_REQUEST, "Unknown Combining algorithm: " + algorithm); | ||
| } | ||
| } | ||
| } | ||
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I tend to not be a huge fan of comments in the code, what does this mean? it's just for your convenience in the draft?
If that's the case, ignore my comment