Support semantic_text in object fields (#114601)

Author: Mikep86

docs/changelog/114601.yaml

@@ -0,0 +1,6 @@
+pr: 114601
+summary: Support semantic_text in object fields
+area: Vector Search
+type: bug
+issues:
+ - 114401

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/InferenceFeatures.java

@@ -36,4 +36,8 @@ public Set<NodeFeature> getFeatures() {
         return Set.copyOf(features);
     }
 
+    @Override
+    public Set<NodeFeature> getTestFeatures() {
+        return Set.of(SemanticTextFieldMapper.SEMANTIC_TEXT_IN_OBJECT_FIELD_FIX);
+    }
 }

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/action/filter/ShardBulkInferenceActionFilter.java

@@ -397,7 +397,7 @@ private void applyInferenceResponses(BulkItemRequest item, FieldInferenceRespons
                     ),
                     indexRequest.getContentType()
                 );
-                newDocMap.put(fieldName, result);
+                SemanticTextFieldMapper.insertValue(fieldName, newDocMap, result);
             }
             indexRequest.source(newDocMap, indexRequest.getContentType());
         }

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/mapper/SemanticTextFieldMapper.java

@@ -32,6 +32,7 @@
 import org.elasticsearch.index.mapper.Mapper;
 import org.elasticsearch.index.mapper.MapperBuilderContext;
 import org.elasticsearch.index.mapper.MapperMergeContext;
+import org.elasticsearch.index.mapper.MappingLookup;
 import org.elasticsearch.index.mapper.NestedObjectMapper;
 import org.elasticsearch.index.mapper.ObjectMapper;
 import org.elasticsearch.index.mapper.SimpleMappedFieldType;
@@ -86,6 +87,7 @@
 public class SemanticTextFieldMapper extends FieldMapper implements InferenceFieldMapper {
     public static final NodeFeature SEMANTIC_TEXT_SEARCH_INFERENCE_ID = new NodeFeature("semantic_text.search_inference_id");
     public static final NodeFeature SEMANTIC_TEXT_DEFAULT_ELSER_2 = new NodeFeature("semantic_text.default_elser_2");
+    public static final NodeFeature SEMANTIC_TEXT_IN_OBJECT_FIELD_FIX = new NodeFeature("semantic_text.in_object_field_fix");
 
     public static final String CONTENT_TYPE = "semantic_text";
     public static final String DEFAULT_ELSER_2_INFERENCE_ID = DEFAULT_ELSER_ID;
@@ -398,6 +400,25 @@ public Object getOriginalValue(Map<String, Object> sourceAsMap) {
         return XContentMapValues.extractValue(TEXT_FIELD, fieldValueMap);
     }
 
+    @Override
+    protected void doValidate(MappingLookup mappers) {
+        int parentPathIndex = fullPath().lastIndexOf(leafName());
+        if (parentPathIndex > 0) {
+            // Check that the parent object field allows subobjects.
+            // Subtract one from the parent path index to omit the trailing dot delimiter.
+            ObjectMapper parentMapper = mappers.objectMappers().get(fullPath().substring(0, parentPathIndex - 1));
+            if (parentMapper == null) {
+                throw new IllegalStateException(CONTENT_TYPE + " field [" + fullPath() + "] does not have a parent object mapper");
+            }
+
+            if (parentMapper.subobjects() == ObjectMapper.Subobjects.DISABLED) {
+                throw new IllegalArgumentException(
+                    CONTENT_TYPE + " field [" + fullPath() + "] cannot be in an object field with subobjects disabled"
+                );
+            }
+        }
+    }
+
     public static class SemanticTextFieldType extends SimpleMappedFieldType {
         private final String inferenceId;
         private final String searchInferenceId;
@@ -592,6 +613,116 @@ private String generateInvalidQueryInferenceResultsMessage(StringBuilder baseMes
         }
     }
 
+    /**
+     * <p>
+     * Insert or replace the path's value in the map with the provided new value. The map will be modified in-place.
+     * If the complete path does not exist in the map, it will be added to the deepest (sub-)map possible.
+     * </p>
+     * <p>
+     * For example, given the map:
+     * </p>
+     * <pre>
+     * {
+     *   "path1": {
+     *     "path2": {
+     *       "key1": "value1"
+     *     }
+     *   }
+     * }
+     * </pre>
+     * <p>
+     * And the caller wanted to insert {@code "path1.path2.path3.key2": "value2"}, the method would emit the modified map:
+     * </p>
+     * <pre>
+     * {
+     *   "path1": {
+     *     "path2": {
+     *       "key1": "value1",
+     *       "path3.key2": "value2"
+     *     }
+     *   }
+     * }
+     * </pre>
+     *
+     * @param path the value's path in the map.
+     * @param map the map to search and modify in-place.
+     * @param newValue the new value to assign to the path.
+     *
+     * @throws IllegalArgumentException If either the path cannot be fully traversed or there is ambiguity about where to insert the new
+     *                                  value.
+     */
+    public static void insertValue(String path, Map<?, ?> map, Object newValue) {
+        String[] pathElements = path.split("\\.");
+        if (pathElements.length == 0) {
+            return;
+        }
+
+        List<SuffixMap> suffixMaps = extractSuffixMaps(pathElements, 0, map);
+        if (suffixMaps.isEmpty()) {
+            // This should never happen. Throw in case it does for some reason.
+            throw new IllegalStateException("extractSuffixMaps returned an empty suffix map list");
+        } else if (suffixMaps.size() == 1) {
+            SuffixMap suffixMap = suffixMaps.getFirst();
+            suffixMap.map().put(suffixMap.suffix(), newValue);
+        } else {
+            throw new IllegalArgumentException(
+                "Path [" + path + "] could be inserted in " + suffixMaps.size() + " distinct ways, it is ambiguous which one to use"
+            );
+        }
+    }
+
+    private record SuffixMap(String suffix, Map<String, Object> map) {}
+
+    private static List<SuffixMap> extractSuffixMaps(String[] pathElements, int index, Object currentValue) {
+        if (currentValue instanceof List<?> valueList) {
+            List<SuffixMap> suffixMaps = new ArrayList<>(valueList.size());
+            for (Object o : valueList) {
+                suffixMaps.addAll(extractSuffixMaps(pathElements, index, o));
+            }
+
+            return suffixMaps;
+        } else if (currentValue instanceof Map<?, ?>) {
+            @SuppressWarnings("unchecked")
+            Map<String, Object> map = (Map<String, Object>) currentValue;
+            List<SuffixMap> suffixMaps = new ArrayList<>(map.size());
+
+            String key = pathElements[index];
+            while (index < pathElements.length) {
+                if (map.containsKey(key)) {
+                    if (index + 1 == pathElements.length) {
+                        // We found the complete path
+                        suffixMaps.add(new SuffixMap(key, map));
+                    } else {
+                        // We've matched that path partially, keep traversing to try to match it fully
+                        suffixMaps.addAll(extractSuffixMaps(pathElements, index + 1, map.get(key)));
+                    }
+                }
+
+                if (++index < pathElements.length) {
+                    key += "." + pathElements[index];
+                }
+            }
+
+            if (suffixMaps.isEmpty()) {
+                // We checked for all remaining elements in the path, and they do not exist. This means we found a leaf map that we should
+                // add the value to.
+                suffixMaps.add(new SuffixMap(key, map));
+            }
+
+            return suffixMaps;
+        } else {
+            throw new IllegalArgumentException(
+                "Path ["
+                    + String.join(".", Arrays.copyOfRange(pathElements, 0, index))
+                    + "] has value ["
+                    + currentValue
+                    + "] of type ["
+                    + currentValue.getClass().getSimpleName()
+                    + "], which cannot be traversed into further"
+            );
+        }
+    }
+
     private static ObjectMapper createInferenceField(
         MapperBuilderContext context,
         IndexVersion indexVersionCreated,