|
| 1 | +/* |
| 2 | + * SPDX-License-Identifier: Apache-2.0 |
| 3 | + * |
| 4 | + * The OpenSearch Contributors require contributions made to |
| 5 | + * this file be licensed under the Apache-2.0 license or a |
| 6 | + * compatible open source license. |
| 7 | + */ |
| 8 | + |
| 9 | +package org.opensearch.index.mapper; |
| 10 | + |
| 11 | +import org.apache.logging.log4j.LogManager; |
| 12 | +import org.apache.logging.log4j.Logger; |
| 13 | +import org.opensearch.common.regex.Regex; |
| 14 | +import org.opensearch.index.query.QueryShardContext; |
| 15 | +import org.opensearch.script.Script; |
| 16 | + |
| 17 | +import java.io.IOException; |
| 18 | +import java.util.HashMap; |
| 19 | +import java.util.HashSet; |
| 20 | +import java.util.List; |
| 21 | +import java.util.Map; |
| 22 | +import java.util.Optional; |
| 23 | +import java.util.Set; |
| 24 | +import java.util.concurrent.ConcurrentHashMap; |
| 25 | + |
| 26 | +/** |
| 27 | + * Accepts definition of DerivedField from search request in both forms: map parsed from SearchRequest and {@link DerivedField} defined using client. |
| 28 | + * The object is initialized per search request and is responsible to resolve {@link DerivedFieldType} given a field name. |
| 29 | + * It uses {@link FieldTypeInference} to infer field type for a nested field within DerivedField of {@link DerivedFieldSupportedTypes#OBJECT} type. |
| 30 | + */ |
| 31 | +public class DefaultDerivedFieldResolver implements DerivedFieldResolver { |
| 32 | + private final QueryShardContext queryShardContext; |
| 33 | + private final Map<String, DerivedFieldType> derivedFieldTypeMap = new ConcurrentHashMap<>(); |
| 34 | + private final FieldTypeInference typeInference; |
| 35 | + private static final Logger logger = LogManager.getLogger(DefaultDerivedFieldResolver.class); |
| 36 | + |
| 37 | + public DefaultDerivedFieldResolver( |
| 38 | + QueryShardContext queryShardContext, |
| 39 | + Map<String, Object> derivedFieldsObject, |
| 40 | + List<DerivedField> derivedFields |
| 41 | + ) { |
| 42 | + this( |
| 43 | + queryShardContext, |
| 44 | + derivedFieldsObject, |
| 45 | + derivedFields, |
| 46 | + new FieldTypeInference( |
| 47 | + queryShardContext.index().getName(), |
| 48 | + queryShardContext.getMapperService(), |
| 49 | + queryShardContext.getIndexReader() |
| 50 | + ) |
| 51 | + ); |
| 52 | + } |
| 53 | + |
| 54 | + public DefaultDerivedFieldResolver( |
| 55 | + QueryShardContext queryShardContext, |
| 56 | + Map<String, Object> derivedFieldsObject, |
| 57 | + List<DerivedField> derivedFields, |
| 58 | + FieldTypeInference typeInference |
| 59 | + ) { |
| 60 | + this.queryShardContext = queryShardContext; |
| 61 | + initDerivedFieldTypes(derivedFieldsObject, derivedFields); |
| 62 | + this.typeInference = typeInference; |
| 63 | + } |
| 64 | + |
| 65 | + @Override |
| 66 | + public Set<String> resolvePattern(String pattern) { |
| 67 | + Set<String> derivedFields = new HashSet<>(); |
| 68 | + for (MappedFieldType fieldType : queryShardContext.getMapperService().fieldTypes()) { |
| 69 | + if (Regex.simpleMatch(pattern, fieldType.name()) && fieldType instanceof DerivedFieldType) { |
| 70 | + derivedFields.add(fieldType.name()); |
| 71 | + } |
| 72 | + } |
| 73 | + for (String fieldName : derivedFieldTypeMap.keySet()) { |
| 74 | + if (Regex.simpleMatch(pattern, fieldName)) { |
| 75 | + derivedFields.add(fieldName); |
| 76 | + } |
| 77 | + } |
| 78 | + return derivedFields; |
| 79 | + } |
| 80 | + |
| 81 | + /** |
| 82 | + * Resolves the fieldName. The search request definitions are given precedence over derived fields definitions in the index mapping. |
| 83 | + * It caches the response for previously resolved field names |
| 84 | + * @param fieldName name of the field. It also accepts nested derived field |
| 85 | + * @return DerivedFieldType if resolved successfully, a null otherwise. |
| 86 | + */ |
| 87 | + @Override |
| 88 | + public DerivedFieldType resolve(String fieldName) { |
| 89 | + return Optional.ofNullable(resolveUsingSearchDefinitions(fieldName)) |
| 90 | + .orElseGet(() -> (DerivedFieldType) queryShardContext.getMapperService().fieldType(fieldName)); |
| 91 | + } |
| 92 | + |
| 93 | + private DerivedFieldType resolveUsingSearchDefinitions(String fieldName) { |
| 94 | + if (derivedFieldTypeMap.containsKey(fieldName)) { |
| 95 | + return derivedFieldTypeMap.get(fieldName); |
| 96 | + } |
| 97 | + DerivedFieldType resolvedNestedType = resolveNestedField(fieldName); |
| 98 | + if (resolvedNestedType != null) { |
| 99 | + derivedFieldTypeMap.put(fieldName, resolvedNestedType); |
| 100 | + } |
| 101 | + return resolvedNestedType; |
| 102 | + } |
| 103 | + |
| 104 | + private DerivedFieldType resolveNestedField(String fieldName) { |
| 105 | + DerivedFieldType parentDerivedField = (DerivedFieldType) getParentDerivedField(fieldName); |
| 106 | + if (parentDerivedField != null) { |
| 107 | + try { |
| 108 | + Script script = parentDerivedField.derivedField.getScript(); |
| 109 | + Mapper inferredFieldMapper = typeInference.infer(getValueFetcher(fieldName, script)); |
| 110 | + if (inferredFieldMapper != null) { |
| 111 | + return getDerivedFieldType( |
| 112 | + new DerivedField( |
| 113 | + fieldName, |
| 114 | + inferredFieldMapper.typeName(), |
| 115 | + script, |
| 116 | + parentDerivedField.derivedField.getSourceIndexedField() |
| 117 | + ) |
| 118 | + ); |
| 119 | + } else { |
| 120 | + logger.warn("Field type cannot be inferred. Ensure the field {} is not rare across entire index", fieldName); |
| 121 | + } |
| 122 | + } catch (IOException e) { |
| 123 | + logger.warn(e); |
| 124 | + } |
| 125 | + } |
| 126 | + return null; |
| 127 | + } |
| 128 | + |
| 129 | + private MappedFieldType getParentDerivedField(String fieldName) { |
| 130 | + if (fieldName.contains(".")) { |
| 131 | + return resolve(fieldName.split("\\.")[0]); |
| 132 | + } |
| 133 | + return null; |
| 134 | + } |
| 135 | + |
| 136 | + ValueFetcher getValueFetcher(String fieldName, Script script) { |
| 137 | + String subFieldName = fieldName.substring(fieldName.indexOf(".") + 1); |
| 138 | + return new ObjectDerivedFieldType.ObjectDerivedFieldValueFetcher( |
| 139 | + subFieldName, |
| 140 | + DerivedFieldType.getDerivedFieldLeafFactory(script, queryShardContext, queryShardContext.lookup()), |
| 141 | + o -> o // raw object returned will be used to infer the type without modifying it |
| 142 | + ); |
| 143 | + } |
| 144 | + |
| 145 | + private void initDerivedFieldTypes(Map<String, Object> derivedFieldsObject, List<DerivedField> derivedFields) { |
| 146 | + if (derivedFieldsObject != null && !derivedFieldsObject.isEmpty()) { |
| 147 | + Map<String, Object> derivedFieldObject = new HashMap<>(); |
| 148 | + derivedFieldObject.put(DerivedFieldMapper.CONTENT_TYPE, derivedFieldsObject); |
| 149 | + derivedFieldTypeMap.putAll(getAllDerivedFieldTypeFromObject(derivedFieldObject)); |
| 150 | + } |
| 151 | + if (derivedFields != null) { |
| 152 | + for (DerivedField derivedField : derivedFields) { |
| 153 | + derivedFieldTypeMap.put(derivedField.getName(), getDerivedFieldType(derivedField)); |
| 154 | + } |
| 155 | + } |
| 156 | + } |
| 157 | + |
| 158 | + private Map<String, DerivedFieldType> getAllDerivedFieldTypeFromObject(Map<String, Object> derivedFieldObject) { |
| 159 | + Map<String, DerivedFieldType> derivedFieldTypes = new HashMap<>(); |
| 160 | + DocumentMapper documentMapper = queryShardContext.getMapperService() |
| 161 | + .documentMapperParser() |
| 162 | + .parse(DerivedFieldMapper.CONTENT_TYPE, derivedFieldObject); |
| 163 | + if (documentMapper != null && documentMapper.mappers() != null) { |
| 164 | + for (Mapper mapper : documentMapper.mappers()) { |
| 165 | + if (mapper instanceof DerivedFieldMapper) { |
| 166 | + DerivedFieldType derivedFieldType = ((DerivedFieldMapper) mapper).fieldType(); |
| 167 | + derivedFieldTypes.put(derivedFieldType.name(), derivedFieldType); |
| 168 | + } |
| 169 | + } |
| 170 | + } |
| 171 | + return derivedFieldTypes; |
| 172 | + } |
| 173 | + |
| 174 | + private DerivedFieldType getDerivedFieldType(DerivedField derivedField) { |
| 175 | + Mapper.BuilderContext builderContext = new Mapper.BuilderContext( |
| 176 | + queryShardContext.getMapperService().getIndexSettings().getSettings(), |
| 177 | + new ContentPath(1) |
| 178 | + ); |
| 179 | + DerivedFieldMapper.Builder builder = new DerivedFieldMapper.Builder( |
| 180 | + derivedField, |
| 181 | + queryShardContext.getMapperService().getIndexAnalyzers() |
| 182 | + ); |
| 183 | + return builder.build(builderContext).fieldType(); |
| 184 | + } |
| 185 | +} |
0 commit comments