|
| 1 | +/* |
| 2 | + * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one |
| 3 | + * or more contributor license agreements. Licensed under the Elastic License |
| 4 | + * 2.0; you may not use this file except in compliance with the Elastic License |
| 5 | + * 2.0. |
| 6 | + */ |
| 7 | + |
| 8 | +package org.elasticsearch.xpack.inference.queries; |
| 9 | + |
| 10 | +import org.apache.lucene.search.join.ScoreMode; |
| 11 | +import org.elasticsearch.features.NodeFeature; |
| 12 | +import org.elasticsearch.index.mapper.IndexFieldMapper; |
| 13 | +import org.elasticsearch.index.query.BoolQueryBuilder; |
| 14 | +import org.elasticsearch.index.query.QueryBuilder; |
| 15 | +import org.elasticsearch.index.query.QueryBuilders; |
| 16 | +import org.elasticsearch.index.query.TermsQueryBuilder; |
| 17 | +import org.elasticsearch.search.vectors.KnnVectorQueryBuilder; |
| 18 | +import org.elasticsearch.search.vectors.QueryVectorBuilder; |
| 19 | +import org.elasticsearch.xpack.core.ml.vectors.TextEmbeddingQueryVectorBuilder; |
| 20 | +import org.elasticsearch.xpack.inference.mapper.SemanticTextField; |
| 21 | + |
| 22 | +import java.util.Collection; |
| 23 | +import java.util.List; |
| 24 | +import java.util.Map; |
| 25 | + |
| 26 | +public class SemanticKnnVectorQueryRewriteInterceptor extends SemanticQueryRewriteInterceptor { |
| 27 | + |
| 28 | + public static final NodeFeature SEMANTIC_KNN_VECTOR_QUERY_REWRITE_INTERCEPTION_SUPPORTED = new NodeFeature( |
| 29 | + "search.semantic_knn_vector_query_rewrite_interception_supported" |
| 30 | + ); |
| 31 | + |
| 32 | + public SemanticKnnVectorQueryRewriteInterceptor() {} |
| 33 | + |
| 34 | + @Override |
| 35 | + protected String getFieldName(QueryBuilder queryBuilder) { |
| 36 | + assert (queryBuilder instanceof KnnVectorQueryBuilder); |
| 37 | + KnnVectorQueryBuilder knnVectorQueryBuilder = (KnnVectorQueryBuilder) queryBuilder; |
| 38 | + return knnVectorQueryBuilder.getFieldName(); |
| 39 | + } |
| 40 | + |
| 41 | + @Override |
| 42 | + protected String getQuery(QueryBuilder queryBuilder) { |
| 43 | + assert (queryBuilder instanceof KnnVectorQueryBuilder); |
| 44 | + KnnVectorQueryBuilder knnVectorQueryBuilder = (KnnVectorQueryBuilder) queryBuilder; |
| 45 | + TextEmbeddingQueryVectorBuilder queryVectorBuilder = getTextEmbeddingQueryBuilderFromQuery(knnVectorQueryBuilder); |
| 46 | + return queryVectorBuilder != null ? queryVectorBuilder.getModelText() : null; |
| 47 | + } |
| 48 | + |
| 49 | + @Override |
| 50 | + protected QueryBuilder buildInferenceQuery(QueryBuilder queryBuilder, InferenceIndexInformationForField indexInformation) { |
| 51 | + assert (queryBuilder instanceof KnnVectorQueryBuilder); |
| 52 | + KnnVectorQueryBuilder knnVectorQueryBuilder = (KnnVectorQueryBuilder) queryBuilder; |
| 53 | + Map<String, List<String>> inferenceIdsIndices = indexInformation.getInferenceIdsIndices(); |
| 54 | + if (inferenceIdsIndices.size() == 1) { |
| 55 | + // Simple case, everything uses the same inference ID |
| 56 | + Map.Entry<String, List<String>> inferenceIdIndex = inferenceIdsIndices.entrySet().iterator().next(); |
| 57 | + String searchInferenceId = inferenceIdIndex.getKey(); |
| 58 | + List<String> indices = inferenceIdIndex.getValue(); |
| 59 | + return buildNestedQueryFromKnnVectorQuery(knnVectorQueryBuilder, indices, searchInferenceId); |
| 60 | + } else { |
| 61 | + // Multiple inference IDs, construct a boolean query |
| 62 | + return buildInferenceQueryWithMultipleInferenceIds(knnVectorQueryBuilder, inferenceIdsIndices); |
| 63 | + } |
| 64 | + } |
| 65 | + |
| 66 | + private QueryBuilder buildInferenceQueryWithMultipleInferenceIds( |
| 67 | + KnnVectorQueryBuilder queryBuilder, |
| 68 | + Map<String, List<String>> inferenceIdsIndices |
| 69 | + ) { |
| 70 | + BoolQueryBuilder boolQueryBuilder = new BoolQueryBuilder(); |
| 71 | + for (String inferenceId : inferenceIdsIndices.keySet()) { |
| 72 | + boolQueryBuilder.should( |
| 73 | + createSubQueryForIndices( |
| 74 | + inferenceIdsIndices.get(inferenceId), |
| 75 | + buildNestedQueryFromKnnVectorQuery(queryBuilder, inferenceIdsIndices.get(inferenceId), inferenceId) |
| 76 | + ) |
| 77 | + ); |
| 78 | + } |
| 79 | + return boolQueryBuilder; |
| 80 | + } |
| 81 | + |
| 82 | + @Override |
| 83 | + protected QueryBuilder buildCombinedInferenceAndNonInferenceQuery( |
| 84 | + QueryBuilder queryBuilder, |
| 85 | + InferenceIndexInformationForField indexInformation |
| 86 | + ) { |
| 87 | + assert (queryBuilder instanceof KnnVectorQueryBuilder); |
| 88 | + KnnVectorQueryBuilder knnVectorQueryBuilder = (KnnVectorQueryBuilder) queryBuilder; |
| 89 | + Map<String, List<String>> inferenceIdsIndices = indexInformation.getInferenceIdsIndices(); |
| 90 | + |
| 91 | + BoolQueryBuilder boolQueryBuilder = new BoolQueryBuilder(); |
| 92 | + boolQueryBuilder.should(addIndexFilterToKnnVectorQuery(indexInformation.nonInferenceIndices(), knnVectorQueryBuilder)); |
| 93 | + |
| 94 | + // We always perform nested subqueries on semantic_text fields, to support knn queries using query vectors. |
| 95 | + // Both pre and post filtering are required here to ensure we get the results we need without errors based on field types. |
| 96 | + for (String inferenceId : inferenceIdsIndices.keySet()) { |
| 97 | + boolQueryBuilder.should( |
| 98 | + createSubQueryForIndices( |
| 99 | + inferenceIdsIndices.get(inferenceId), |
| 100 | + buildNestedQueryFromKnnVectorQuery(knnVectorQueryBuilder, inferenceIdsIndices.get(inferenceId), inferenceId) |
| 101 | + ) |
| 102 | + ); |
| 103 | + } |
| 104 | + return boolQueryBuilder; |
| 105 | + } |
| 106 | + |
| 107 | + private QueryBuilder buildNestedQueryFromKnnVectorQuery( |
| 108 | + KnnVectorQueryBuilder knnVectorQueryBuilder, |
| 109 | + List<String> indices, |
| 110 | + String searchInferenceId |
| 111 | + ) { |
| 112 | + KnnVectorQueryBuilder filteredKnnVectorQueryBuilder = addIndexFilterToKnnVectorQuery(indices, knnVectorQueryBuilder); |
| 113 | + TextEmbeddingQueryVectorBuilder queryVectorBuilder = getTextEmbeddingQueryBuilderFromQuery(filteredKnnVectorQueryBuilder); |
| 114 | + if (queryVectorBuilder != null && queryVectorBuilder.getModelId() == null && searchInferenceId != null) { |
| 115 | + // If the model ID was not specified, we infer the inference ID associated with the semantic_text field. |
| 116 | + queryVectorBuilder = new TextEmbeddingQueryVectorBuilder(searchInferenceId, queryVectorBuilder.getModelText()); |
| 117 | + } |
| 118 | + return QueryBuilders.nestedQuery( |
| 119 | + SemanticTextField.getChunksFieldName(filteredKnnVectorQueryBuilder.getFieldName()), |
| 120 | + buildNewKnnVectorQuery( |
| 121 | + SemanticTextField.getEmbeddingsFieldName(filteredKnnVectorQueryBuilder.getFieldName()), |
| 122 | + filteredKnnVectorQueryBuilder, |
| 123 | + queryVectorBuilder |
| 124 | + ), |
| 125 | + ScoreMode.Max |
| 126 | + ); |
| 127 | + } |
| 128 | + |
| 129 | + private KnnVectorQueryBuilder addIndexFilterToKnnVectorQuery(Collection<String> indices, KnnVectorQueryBuilder original) { |
| 130 | + KnnVectorQueryBuilder copy; |
| 131 | + if (original.queryVectorBuilder() != null) { |
| 132 | + copy = new KnnVectorQueryBuilder( |
| 133 | + original.getFieldName(), |
| 134 | + original.queryVectorBuilder(), |
| 135 | + original.k(), |
| 136 | + original.numCands(), |
| 137 | + original.getVectorSimilarity() |
| 138 | + ); |
| 139 | + } else { |
| 140 | + copy = new KnnVectorQueryBuilder( |
| 141 | + original.getFieldName(), |
| 142 | + original.queryVector(), |
| 143 | + original.k(), |
| 144 | + original.numCands(), |
| 145 | + original.rescoreVectorBuilder(), |
| 146 | + original.getVectorSimilarity() |
| 147 | + ); |
| 148 | + } |
| 149 | + |
| 150 | + copy.addFilterQuery(new TermsQueryBuilder(IndexFieldMapper.NAME, indices)); |
| 151 | + return copy; |
| 152 | + } |
| 153 | + |
| 154 | + private TextEmbeddingQueryVectorBuilder getTextEmbeddingQueryBuilderFromQuery(KnnVectorQueryBuilder knnVectorQueryBuilder) { |
| 155 | + QueryVectorBuilder queryVectorBuilder = knnVectorQueryBuilder.queryVectorBuilder(); |
| 156 | + if (queryVectorBuilder == null) { |
| 157 | + return null; |
| 158 | + } |
| 159 | + assert (queryVectorBuilder instanceof TextEmbeddingQueryVectorBuilder); |
| 160 | + return (TextEmbeddingQueryVectorBuilder) queryVectorBuilder; |
| 161 | + } |
| 162 | + |
| 163 | + private KnnVectorQueryBuilder buildNewKnnVectorQuery( |
| 164 | + String fieldName, |
| 165 | + KnnVectorQueryBuilder original, |
| 166 | + QueryVectorBuilder queryVectorBuilder |
| 167 | + ) { |
| 168 | + if (original.queryVectorBuilder() != null) { |
| 169 | + return new KnnVectorQueryBuilder( |
| 170 | + fieldName, |
| 171 | + queryVectorBuilder, |
| 172 | + original.k(), |
| 173 | + original.numCands(), |
| 174 | + original.getVectorSimilarity() |
| 175 | + ); |
| 176 | + } else { |
| 177 | + return new KnnVectorQueryBuilder( |
| 178 | + fieldName, |
| 179 | + original.queryVector(), |
| 180 | + original.k(), |
| 181 | + original.numCands(), |
| 182 | + original.rescoreVectorBuilder(), |
| 183 | + original.getVectorSimilarity() |
| 184 | + ); |
| 185 | + } |
| 186 | + } |
| 187 | + |
| 188 | + @Override |
| 189 | + public String getQueryName() { |
| 190 | + return KnnVectorQueryBuilder.NAME; |
| 191 | + } |
| 192 | +} |
0 commit comments