cpoerschke commented on code in PR #3418: URL: https://github.com/apache/solr/pull/3418#discussion_r2314378342
########## solr/core/src/java/org/apache/solr/handler/component/CombinedQueryComponent.java: ########## @@ -0,0 +1,590 @@ +/* + * 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.handler.component; + +import java.io.IOException; +import java.io.PrintWriter; +import java.io.StringWriter; +import java.lang.invoke.MethodHandles; +import java.util.ArrayList; +import java.util.Collections; +import java.util.HashMap; +import java.util.List; +import java.util.Map; +import java.util.Objects; +import java.util.Set; +import java.util.stream.Collectors; +import org.apache.lucene.search.Explanation; +import org.apache.lucene.search.Query; +import org.apache.lucene.search.Sort; +import org.apache.lucene.search.SortField; +import org.apache.solr.client.solrj.SolrServerException; +import org.apache.solr.common.SolrDocument; +import org.apache.solr.common.SolrDocumentList; +import org.apache.solr.common.SolrException; +import org.apache.solr.common.params.CombinerParams; +import org.apache.solr.common.params.CursorMarkParams; +import org.apache.solr.common.params.GroupParams; +import org.apache.solr.common.params.ShardParams; +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.common.util.StrUtils; +import org.apache.solr.core.SolrCore; +import org.apache.solr.response.BasicResultContext; +import org.apache.solr.response.ResultContext; +import org.apache.solr.response.SolrQueryResponse; +import org.apache.solr.schema.IndexSchema; +import org.apache.solr.schema.SchemaField; +import org.apache.solr.search.DocListAndSet; +import org.apache.solr.search.QueryResult; +import org.apache.solr.search.SolrReturnFields; +import org.apache.solr.search.SortSpec; +import org.apache.solr.search.combine.QueryAndResponseCombiner; +import org.apache.solr.search.combine.ReciprocalRankFusion; +import org.apache.solr.util.SolrResponseUtil; +import org.apache.solr.util.plugin.SolrCoreAware; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +/** + * The CombinedQueryComponent class extends QueryComponent and provides support for executing + * multiple queries and combining their results. + */ +public class CombinedQueryComponent extends QueryComponent implements SolrCoreAware { + + private static final Logger log = LoggerFactory.getLogger(MethodHandles.lookup().lookupClass()); + public static final String COMPONENT_NAME = "combined_query"; + protected NamedList<?> initParams; + private final Map<String, QueryAndResponseCombiner> combiners = new HashMap<>(); + private int maxCombinerQueries; + + @Override + public void init(NamedList<?> args) { + super.init(args); + this.initParams = args; + this.maxCombinerQueries = CombinerParams.DEFAULT_MAX_COMBINER_QUERIES; + } + + @Override + public void inform(SolrCore core) { + if (initParams != null && initParams.size() > 0) { + for (Map.Entry<String, ?> initEntry : initParams) { + if ("combiners".equals(initEntry.getKey()) + && initEntry.getValue() instanceof NamedList<?> all) { + for (int i = 0; i < all.size(); i++) { + String name = all.getName(i); + NamedList<?> combinerConfig = (NamedList<?>) all.getVal(i); + String className = (String) combinerConfig.get("class"); + QueryAndResponseCombiner combiner = + core.getResourceLoader().newInstance(className, QueryAndResponseCombiner.class); + combiner.init(combinerConfig); + combiners.compute( + name, + (k, existingCombiner) -> { + if (existingCombiner == null) { + return combiner; + } + throw new SolrException( + SolrException.ErrorCode.BAD_REQUEST, + "Found more than one combiner with same name"); + }); + } + } + } + Object maxQueries = initParams.get("maxCombinerQueries"); + if (maxQueries != null) { + this.maxCombinerQueries = Integer.parseInt(maxQueries.toString()); + } + } + combiners.computeIfAbsent( + CombinerParams.RECIPROCAL_RANK_FUSION, + key -> { + ReciprocalRankFusion reciprocalRankFusion = new ReciprocalRankFusion(); + reciprocalRankFusion.init(initParams); + return reciprocalRankFusion; + }); + } + + /** + * Overrides the prepare method to handle combined queries. + * + * @param rb the ResponseBuilder to prepare + * @throws IOException if an I/O error occurs during preparation + */ + @Override + public void prepare(ResponseBuilder rb) throws IOException { + if (rb instanceof CombinedQueryResponseBuilder crb) { + SolrParams params = crb.req.getParams(); + if (params.get(CursorMarkParams.CURSOR_MARK_PARAM) != null + || params.getBool(GroupParams.GROUP, false)) { + throw new SolrException( + SolrException.ErrorCode.BAD_REQUEST, "Unsupported functionality for Combined Queries."); + } + String[] queriesToCombineKeys = params.getParams(CombinerParams.COMBINER_QUERY); + if (queriesToCombineKeys.length > maxCombinerQueries) { + throw new SolrException( + SolrException.ErrorCode.BAD_REQUEST, + "Too many queries to combine: limit is " + maxCombinerQueries); + } + for (String queryKey : queriesToCombineKeys) { + final var unparsedQuery = params.get(queryKey); + ResponseBuilder rbNew = new ResponseBuilder(rb.req, new SolrQueryResponse(), rb.components); + rbNew.setQueryString(unparsedQuery); + super.prepare(rbNew); + crb.responseBuilders.add(rbNew); + } + } + super.prepare(rb); + } + + /** + * Overrides the process method to handle CombinedQueryResponseBuilder instances. This method + * processes the responses from multiple shards, combines them using the specified + * QueryAndResponseCombiner strategy, and sets the appropriate results and metadata in the + * CombinedQueryResponseBuilder. + * + * @param rb the ResponseBuilder object to process + * @throws IOException if an I/O error occurs during processing + */ + @Override + public void process(ResponseBuilder rb) throws IOException { + if (rb instanceof CombinedQueryResponseBuilder crb) { + boolean partialResults = false; + boolean segmentTerminatedEarly = false; + boolean setMaxHitsTerminatedEarly = false; + List<QueryResult> queryResults = new ArrayList<>(); + for (ResponseBuilder rbNow : crb.responseBuilders) { + // Just a placeholder for future implementation for Cursors + rbNow.setCursorMark(crb.getCursorMark()); + super.process(rbNow); + DocListAndSet docListAndSet = rbNow.getResults(); + QueryResult queryResult = new QueryResult(); + queryResult.setDocListAndSet(docListAndSet); + queryResults.add(queryResult); + partialResults |= queryResult.isPartialResults(); + if (queryResult.getSegmentTerminatedEarly() != null) { + segmentTerminatedEarly |= queryResult.getSegmentTerminatedEarly(); + } + if (queryResult.getMaxHitsTerminatedEarly() != null) { + setMaxHitsTerminatedEarly |= queryResult.getMaxHitsTerminatedEarly(); + } + } + String algorithm = + rb.req + .getParams() + .get(CombinerParams.COMBINER_ALGORITHM, CombinerParams.DEFAULT_COMBINER); + QueryAndResponseCombiner combinerStrategy = + QueryAndResponseCombiner.getImplementation(algorithm, combiners); + QueryResult combinedQueryResult = combinerStrategy.combine(queryResults, rb.req.getParams()); + combinedQueryResult.setPartialResults(partialResults); + combinedQueryResult.setSegmentTerminatedEarly(segmentTerminatedEarly); + combinedQueryResult.setMaxHitsTerminatedEarly(setMaxHitsTerminatedEarly); + crb.setResult(combinedQueryResult); + if (rb.isDebug()) { + String[] queryKeys = rb.req.getParams().getParams(CombinerParams.COMBINER_QUERY); + List<Query> queries = crb.responseBuilders.stream().map(ResponseBuilder::getQuery).toList(); + NamedList<Explanation> explanations = + combinerStrategy.getExplanations( + queryKeys, + queries, + queryResults, + rb.req.getSearcher(), + rb.req.getSchema(), + rb.req.getParams()); + rb.addDebugInfo("combinerExplanations", explanations); + } + ResultContext ctx = new BasicResultContext(crb); + crb.rsp.addResponse(ctx); + crb.rsp.addToLog( + "hits", + crb.getResults() == null || crb.getResults().docList == null + ? 0 + : crb.getResults().docList.matches()); + if (!crb.req.getParams().getBool(ShardParams.IS_SHARD, false)) { + // for non-distributed request and future cursor improvement + if (null != crb.getNextCursorMark()) { + crb.rsp.add( + CursorMarkParams.CURSOR_MARK_NEXT, + crb.responseBuilders.getFirst().getNextCursorMark().getSerializedTotem()); + } + } + + if (crb.mergeFieldHandler != null) { + crb.mergeFieldHandler.handleMergeFields(crb, crb.req.getSearcher()); + } else { + doFieldSortValues(rb, crb.req.getSearcher()); + } + doPrefetch(crb); + } else { + super.process(rb); + } + } + + @Override + protected void mergeIds(ResponseBuilder rb, ShardRequest sreq) { + List<MergeStrategy> mergeStrategies = rb.getMergeStrategies(); + if (mergeStrategies != null) { + mergeStrategies.sort(MergeStrategy.MERGE_COMP); + boolean idsMerged = false; + for (MergeStrategy mergeStrategy : mergeStrategies) { + mergeStrategy.merge(rb, sreq); + if (mergeStrategy.mergesIds()) { + idsMerged = true; + } + } + + if (idsMerged) { + return; // ids were merged above so return. + } + } + + SortSpec ss = rb.getSortSpec(); + + // If the shard request was also used to get fields (along with the scores), there is no reason + // to copy over the score dependent fields, since those will already exist in the document with + // the return fields + Set<String> scoreDependentFields; + if ((sreq.purpose & ShardRequest.PURPOSE_GET_FIELDS) == 0) { + scoreDependentFields = + rb.rsp.getReturnFields().getScoreDependentReturnFields().keySet().stream() + .filter(field -> !field.equals(SolrReturnFields.SCORE)) + .collect(Collectors.toSet()); + } else { + scoreDependentFields = Collections.emptySet(); + } + + IndexSchema schema = rb.req.getSchema(); + SchemaField uniqueKeyField = schema.getUniqueKeyField(); + + // id to shard mapping, to eliminate any accidental dups + HashMap<Object, String> uniqueDoc = new HashMap<>(); + + NamedList<Object> shardInfo = null; + if (rb.req.getParams().getBool(ShardParams.SHARDS_INFO, false)) { + shardInfo = new SimpleOrderedMap<>(); + rb.rsp.getValues().add(ShardParams.SHARDS_INFO, shardInfo); + } + + long numFound = 0; + boolean hitCountIsExact = true; + Float maxScore = null; + boolean thereArePartialResults = false; + Boolean segmentTerminatedEarly = null; + boolean maxHitsTerminatedEarly = false; + long approximateTotalHits = 0; + int failedShardCount = 0; + Map<String, List<ShardDoc>> shardDocMap = new HashMap<>(); + for (ShardResponse srsp : sreq.responses) { + SolrDocumentList docs = null; + NamedList<?> responseHeader = null; + + if (shardInfo != null) { + SimpleOrderedMap<Object> nl = new SimpleOrderedMap<>(); + + if (srsp.getException() != null) { + Throwable t = srsp.getException(); + if (t instanceof SolrServerException && t.getCause() != null) { + t = t.getCause(); + } + nl.add("error", t.toString()); + if (!rb.req.getCore().getCoreContainer().hideStackTrace()) { + StringWriter trace = new StringWriter(); + t.printStackTrace(new PrintWriter(trace)); + nl.add("trace", trace.toString()); + } + if (!StrUtils.isNullOrEmpty(srsp.getShardAddress())) { + nl.add("shardAddress", srsp.getShardAddress()); + } + } else { + responseHeader = + (NamedList<?>) + SolrResponseUtil.getSubsectionFromShardResponse( + rb, srsp, "responseHeader", false); + if (responseHeader == null) { + continue; + } + final Object rhste = + responseHeader.get(SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY); + if (rhste != null) { + nl.add(SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY, rhste); + } + final Object rhmhte = + responseHeader.get(SolrQueryResponse.RESPONSE_HEADER_MAX_HITS_TERMINATED_EARLY_KEY); + if (rhmhte != null) { + nl.add(SolrQueryResponse.RESPONSE_HEADER_MAX_HITS_TERMINATED_EARLY_KEY, rhmhte); + } + final Object rhath = + responseHeader.get(SolrQueryResponse.RESPONSE_HEADER_APPROXIMATE_TOTAL_HITS_KEY); + if (rhath != null) { + nl.add(SolrQueryResponse.RESPONSE_HEADER_APPROXIMATE_TOTAL_HITS_KEY, rhath); + } + docs = + (SolrDocumentList) + SolrResponseUtil.getSubsectionFromShardResponse(rb, srsp, "response", false); + if (docs == null) { + continue; + } + nl.add("numFound", docs.getNumFound()); + nl.add("numFoundExact", docs.getNumFoundExact()); + nl.add("maxScore", docs.getMaxScore()); + nl.add("shardAddress", srsp.getShardAddress()); + } + if (srsp.getSolrResponse() != null) { + nl.add("time", srsp.getSolrResponse().getElapsedTime()); + } + // This ought to be better, but at least this ensures no duplicate keys in JSON result + String shard = srsp.getShard(); + if (StrUtils.isNullOrEmpty(shard)) { + failedShardCount += 1; + shard = "unknown_shard_" + failedShardCount; + } + shardInfo.add(shard, nl); + } + // now that we've added the shard info, let's only proceed if we have no error. + if (srsp.getException() != null) { + thereArePartialResults = true; + continue; + } + + if (docs == null) { // could have been initialized in the shards info block above + docs = + Objects.requireNonNull( + (SolrDocumentList) + SolrResponseUtil.getSubsectionFromShardResponse(rb, srsp, "response", false)); + } + + if (responseHeader == null) { // could have been initialized in the shards info block above + responseHeader = + Objects.requireNonNull( + (NamedList<?>) + SolrResponseUtil.getSubsectionFromShardResponse( + rb, srsp, "responseHeader", false)); + } + + final boolean thisResponseIsPartial; + thisResponseIsPartial = + Boolean.TRUE.equals( + responseHeader.getBooleanArg(SolrQueryResponse.RESPONSE_HEADER_PARTIAL_RESULTS_KEY)); + thereArePartialResults |= thisResponseIsPartial; + + if (!Boolean.TRUE.equals(segmentTerminatedEarly)) { + final Object ste = + responseHeader.get(SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY); + if (Boolean.TRUE.equals(ste)) { + segmentTerminatedEarly = Boolean.TRUE; + } else if (Boolean.FALSE.equals(ste)) { + segmentTerminatedEarly = Boolean.FALSE; + } + } + + if (!maxHitsTerminatedEarly) { + if (Boolean.TRUE.equals( + responseHeader.get(SolrQueryResponse.RESPONSE_HEADER_MAX_HITS_TERMINATED_EARLY_KEY))) { + maxHitsTerminatedEarly = true; + } + } + Object ath = responseHeader.get(SolrQueryResponse.RESPONSE_HEADER_APPROXIMATE_TOTAL_HITS_KEY); + if (ath == null) { + approximateTotalHits += numFound; + } else { + approximateTotalHits += ((Number) ath).longValue(); + } + + // calculate global maxScore and numDocsFound + if (docs.getMaxScore() != null) { + maxScore = maxScore == null ? docs.getMaxScore() : Math.max(maxScore, docs.getMaxScore()); + } + numFound += docs.getNumFound(); + + if (hitCountIsExact && Boolean.FALSE.equals(docs.getNumFoundExact())) { + hitCountIsExact = false; + } + + @SuppressWarnings("unchecked") + NamedList<List<Object>> sortFieldValues = + (NamedList<List<Object>>) + SolrResponseUtil.getSubsectionFromShardResponse(rb, srsp, "sort_values", true); + if (null == sortFieldValues) { + sortFieldValues = new NamedList<>(); + } + + // if the SortSpec contains a field besides score or the Lucene docid, then the values will + // need to be unmarshalled from sortFieldValues. + boolean needsUnmarshalling = ss.includesNonScoreOrDocField(); + + // if we need to unmarshal the sortFieldValues for sorting but we have none, which can happen + // if partial results are being returned from the shard, then skip merging the results for the + // shard. This avoids an exception below. if the shard returned partial results but we don't + // need to unmarshal (a normal scoring query), then merge what we got. + if (thisResponseIsPartial && sortFieldValues.size() == 0 && needsUnmarshalling) { + continue; + } + + // Checking needsUnmarshalling saves on iterating the SortFields in the SortSpec again. + NamedList<List<Object>> unmarshalledSortFieldValues = + needsUnmarshalling ? unmarshalSortValues(ss, sortFieldValues, schema) : new NamedList<>(); + + // go through every doc in this response, construct a ShardDoc, and + // put it in the priority queue so it can be ordered. + for (int i = 0; i < docs.size(); i++) { + SolrDocument doc = docs.get(i); + Object id = doc.getFieldValue(uniqueKeyField.getName()); + ShardDoc shardDoc = new ShardDoc(); + shardDoc.id = id; + shardDoc.shard = srsp.getShard(); + shardDoc.orderInShard = i; + Object scoreObj = doc.getFieldValue(SolrReturnFields.SCORE); + if (scoreObj != null) { + if (scoreObj instanceof String) { + shardDoc.score = Float.parseFloat((String) scoreObj); + } else { + shardDoc.score = ((Number) scoreObj).floatValue(); + } + } + if (!scoreDependentFields.isEmpty()) { + shardDoc.scoreDependentFields = doc.getSubsetOfFields(scoreDependentFields); + } + + shardDoc.sortFieldValues = unmarshalledSortFieldValues; + shardDocMap.computeIfAbsent(srsp.getShard(), list -> new ArrayList<>()).add(shardDoc); + String prevShard = uniqueDoc.put(id, srsp.getShard()); + if (prevShard != null) { + // duplicate detected + numFound--; + } + } // end for-each-doc-in-response + } // end for-each-response + + SolrDocumentList responseDocs = new SolrDocumentList(); + if (maxScore != null) responseDocs.setMaxScore(maxScore); + rb.rsp.addToLog("hits", numFound); + + responseDocs.setNumFound(numFound); + responseDocs.setNumFoundExact(hitCountIsExact); + responseDocs.setStart(ss.getOffset()); + + // save these results in a private area so we can access them + // again when retrieving stored fields. + // TODO: use ResponseBuilder (w/ comments) or the request context? + rb.resultIds = createShardResult(rb, shardDocMap, responseDocs); + rb.setResponseDocs(responseDocs); + + populateNextCursorMarkFromMergedShards(rb); + + if (thereArePartialResults) { + rb.rsp + .getResponseHeader() + .asShallowMap() + .put(SolrQueryResponse.RESPONSE_HEADER_PARTIAL_RESULTS_KEY, Boolean.TRUE); + } + if (segmentTerminatedEarly != null) { + final Object existingSegmentTerminatedEarly = + rb.rsp + .getResponseHeader() + .get(SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY); + if (existingSegmentTerminatedEarly == null) { + rb.rsp + .getResponseHeader() + .add( + SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY, + segmentTerminatedEarly); + } else if (!Boolean.TRUE.equals(existingSegmentTerminatedEarly) + && Boolean.TRUE.equals(segmentTerminatedEarly)) { + rb.rsp + .getResponseHeader() + .remove(SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY); + rb.rsp + .getResponseHeader() + .add( + SolrQueryResponse.RESPONSE_HEADER_SEGMENT_TERMINATED_EARLY_KEY, + segmentTerminatedEarly); + } + } + if (maxHitsTerminatedEarly) { + rb.rsp + .getResponseHeader() + .add(SolrQueryResponse.RESPONSE_HEADER_MAX_HITS_TERMINATED_EARLY_KEY, Boolean.TRUE); + if (approximateTotalHits > 0) { + rb.rsp + .getResponseHeader() + .add( + SolrQueryResponse.RESPONSE_HEADER_APPROXIMATE_TOTAL_HITS_KEY, approximateTotalHits); + } + } + } + + /** + * Combines and sorts documents from multiple shards to create the final result set. This method + * uses a combiner strategy to merge shard responses, then sorts the resulting documents using a + * priority queue based on the request's sort specification. It handles pagination (offset and + * count) and calculates the maximum score for the response. + * + * @param rb The ResponseBuilder containing the request and context, such as sort specifications. + * @param shardDocMap A map from shard addresses to the list of documents returned by each shard. + * @param responseDocs The final response document list, which will be populated with null + * placeholders and have its max score set. + * @return A map from document IDs to the corresponding ShardDoc objects for the documents in the + * final sorted page of results. + */ + protected Map<Object, ShardDoc> createShardResult( + ResponseBuilder rb, Map<String, List<ShardDoc>> shardDocMap, SolrDocumentList responseDocs) { + String algorithm = + rb.req.getParams().get(CombinerParams.COMBINER_ALGORITHM, CombinerParams.DEFAULT_COMBINER); + QueryAndResponseCombiner combinerStrategy = + QueryAndResponseCombiner.getImplementation(algorithm, combiners); + List<ShardDoc> combinedShardDocs = combinerStrategy.combine(shardDocMap, rb.req.getParams()); + Map<String, ShardDoc> shardDocIdMap = new HashMap<>(); + shardDocMap.forEach( + (shardKey, shardDocs) -> + shardDocs.forEach(shardDoc -> shardDocIdMap.put(shardDoc.id.toString(), shardDoc))); + Map<Object, ShardDoc> resultIds = new HashMap<>(); + float maxScore = 0.0f; + Sort sort = rb.getSortSpec().getSort(); + SortField[] sortFields; + if (sort != null) { + sortFields = sort.getSort(); + } else { + sortFields = new SortField[] {SortField.FIELD_SCORE}; + } + final ShardFieldSortedHitQueue queue = + new ShardFieldSortedHitQueue( + sortFields, + rb.getSortSpec().getOffset() + rb.getSortSpec().getCount(), + rb.req.getSearcher()); + combinedShardDocs.forEach(queue::insertWithOverflow); Review Comment: Could you help me understand why the combined shard docs need to go through the sorting queue here? -- 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...@solr.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org For additional commands, e-mail: issues-h...@solr.apache.org
