Cleanup

Author: kderusso

docs/reference/mapping/types/semantic-text.asciidoc

@@ -26,6 +26,15 @@
 import java.util.Locale;
 import java.util.Map;
 
+/**
+ * Represents index options for a semantic_text field.
+ * We represent semantic_text index_options as nested within their respective type. For example:
+ * "index_options": {
+ *   "dense_vector": {
+ *    "type": "bbq_hnsw
+ *    }
+ *  }
+ */
 public class SemanticTextIndexOptions implements ToXContent, Writeable {
 
     private static final String TYPE_FIELD = "type";
@@ -103,11 +112,11 @@ private static DenseVectorFieldMapper.DenseVectorIndexOptions parseDenseVectorIn
         try {
             Object type = map.remove(TYPE_FIELD);
             if (type == null) {
-                throw new IllegalArgumentException("Required [" + TYPE_FIELD + "]");
+                throw new IllegalArgumentException("Required " + TYPE_FIELD);
             }
             DenseVectorFieldMapper.VectorIndexType vectorIndexType = DenseVectorFieldMapper.VectorIndexType.fromString(
-                XContentMapValues.nodeStringValue(type.toString(), null)
-            ).orElseThrow(() -> new IllegalArgumentException("Unsupported index options " + TYPE_FIELD + " [" + type + "]"));
+                XContentMapValues.nodeStringValue(type, null)
+            ).orElseThrow(() -> new IllegalArgumentException("Unsupported index options " + TYPE_FIELD + type));
 
             return vectorIndexType.parseIndexOptions(fieldName, map, indexVersion);
         } catch (Exception exc) {

server/src/main/java/org/elasticsearch/cluster/metadata/SemanticTextIndexOptions.java

@@ -16,6 +16,9 @@
 
 import java.io.IOException;
 
+/**
+ * Represents general index options that can be attached to a semantic or vector field. 
+ */
 public abstract class IndexOptions implements ToXContent, Writeable {
 
     public IndexOptions() {}

server/src/main/java/org/elasticsearch/index/mapper/vectors/IndexOptions.java

@@ -38,7 +38,6 @@
 import org.elasticsearch.index.shard.ShardId;
 import org.elasticsearch.index.shard.ShardLongFieldRange;
 import org.elasticsearch.indices.IndicesModule;
-import org.elasticsearch.inference.TaskType;
 import org.elasticsearch.test.ESTestCase;
 import org.elasticsearch.test.index.IndexVersionUtils;
 import org.elasticsearch.xcontent.NamedXContentRegistry;
@@ -730,7 +729,7 @@ private static InferenceFieldMetadata randomInferenceFieldMetadata(String name)
             randomIdentifier(),
             randomSet(1, 5, ESTestCase::randomIdentifier).toArray(String[]::new),
             InferenceFieldMetadataTests.generateRandomChunkingSettings(),
-            InferenceFieldMetadataTests.randomSemanticTextIndexOptions(TaskType.TEXT_EMBEDDING)
+            InferenceFieldMetadataTests.randomSemanticTextIndexOptions()
         );
     }
 

server/src/test/java/org/elasticsearch/cluster/metadata/IndexMetadataTests.java

@@ -70,7 +70,7 @@ private static InferenceFieldMetadata createTestItem() {
         String searchInferenceId = randomIdentifier();
         String[] inputFields = generateRandomStringArray(5, 10, false, false);
         Map<String, Object> chunkingSettings = generateRandomChunkingSettings();
-        SemanticTextIndexOptions indexOptions = randomSemanticTextIndexOptions(TaskType.TEXT_EMBEDDING);
+        SemanticTextIndexOptions indexOptions = randomSemanticTextIndexOptions();
         return new InferenceFieldMetadata(name, inferenceId, searchInferenceId, inputFields, chunkingSettings, indexOptions);
     }
 
@@ -81,6 +81,11 @@ public static Map<String, Object> generateRandomChunkingSettings() {
         return randomBoolean() ? generateRandomWordBoundaryChunkingSettings() : generateRandomSentenceBoundaryChunkingSettings();
     }
 
+    public static SemanticTextIndexOptions randomSemanticTextIndexOptions() {
+        TaskType taskType = randomFrom(TaskType.SPARSE_EMBEDDING, TaskType.TEXT_EMBEDDING);
+        return randomSemanticTextIndexOptions(taskType);
+    }
+
     public static SemanticTextIndexOptions randomSemanticTextIndexOptions(TaskType taskType) {
 
         if (taskType == TaskType.TEXT_EMBEDDING) {
@@ -89,7 +94,6 @@ public static SemanticTextIndexOptions randomSemanticTextIndexOptions(TaskType t
                 : new SemanticTextIndexOptions(SemanticTextIndexOptions.SupportedIndexOptions.DENSE_VECTOR, randomIndexOptionsAll());
         }
 
-        // No other supported task types
         return null;
     }
 

server/src/test/java/org/elasticsearch/cluster/metadata/InferenceFieldMetadataTests.java

@@ -1135,6 +1135,20 @@ private static Mapper.Builder createEmbeddingsField(
                     }
                 }
 
+                boolean hasUserSpecifiedIndexOptions = indexOptions != null;
+                DenseVectorFieldMapper.DenseVectorIndexOptions denseVectorIndexOptions = hasUserSpecifiedIndexOptions
+                    ? (DenseVectorFieldMapper.DenseVectorIndexOptions) indexOptions.indexOptions()
+                    : (indexVersionCreated.onOrAfter(SEMANTIC_TEXT_DEFAULTS_TO_BBQ) ? defaultSemanticDenseIndexOptions() : null);
+
+                if (denseVectorIndexOptions != null
+                    && denseVectorIndexOptions.validate(
+                        modelSettings.elementType(),
+                        modelSettings.dimensions(),
+                        hasUserSpecifiedIndexOptions
+                    )) {
+                    denseVectorMapperBuilder.indexOptions(denseVectorIndexOptions);
+                }
+
                 yield denseVectorMapperBuilder;
             }
             default -> throw new IllegalArgumentException("Invalid task_type in model_settings [" + modelSettings.taskType().name() + "]");

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

@@ -137,23 +137,4 @@ static IndexVersion getRandomCompatibleIndexVersion(boolean useLegacyFormat, Ind
             );
         }
     }
-
-    static IndexVersion getRandomCompatibleIndexVersion(boolean useLegacyFormat, IndexVersion minVersion, IndexVersion maxVersion) {
-        if (useLegacyFormat) {
-            if (randomBoolean()) {
-                return IndexVersionUtils.randomVersionBetween(random(), IndexVersions.UPGRADE_TO_LUCENE_10_0_0, maxVersion);
-            }
-            return IndexVersionUtils.randomPreviousCompatibleVersion(random(), IndexVersions.INFERENCE_METADATA_FIELDS_BACKPORT);
-        } else {
-            if (randomBoolean()) {
-                return IndexVersionUtils.randomVersionBetween(random(), IndexVersions.INFERENCE_METADATA_FIELDS, maxVersion);
-            }
-            return IndexVersionUtils.randomVersionBetween(
-                random(),
-                IndexVersions.INFERENCE_METADATA_FIELDS_BACKPORT,
-                IndexVersionUtils.getPreviousVersion(IndexVersions.UPGRADE_TO_LUCENE_10_0_0)
-            );
-        }
-    }
-
 }