Add custom serializer

Author: kderusso

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

@@ -69,6 +69,7 @@
 import org.elasticsearch.inference.InferenceResults;
 import org.elasticsearch.inference.MinimalServiceSettings;
 import org.elasticsearch.inference.SimilarityMeasure;
+import org.elasticsearch.inference.TaskType;
 import org.elasticsearch.search.fetch.StoredFieldsSpec;
 import org.elasticsearch.search.lookup.Source;
 import org.elasticsearch.search.vectors.KnnVectorQueryBuilder;
@@ -166,19 +167,9 @@ public static BiConsumer<String, MappingParserContext> validateParserContext(Str
     public static class Builder extends FieldMapper.Builder {
         private final ModelRegistry modelRegistry;
         private final boolean useLegacyFormat;
+        private final IndexVersion indexVersionCreated;
 
-        private final Parameter<String> inferenceId = Parameter.stringParam(
-            INFERENCE_ID_FIELD,
-            false,
-            mapper -> ((SemanticTextFieldType) mapper.fieldType()).inferenceId,
-            DEFAULT_ELSER_2_INFERENCE_ID
-        ).addValidator(v -> {
-            if (Strings.isEmpty(v)) {
-                throw new IllegalArgumentException(
-                    "[" + INFERENCE_ID_FIELD + "] on mapper [" + leafName() + "] of type [" + CONTENT_TYPE + "] must not be empty"
-                );
-            }
-        }).alwaysSerialize();
+        private final Parameter<String> inferenceId;
 
         private final Parameter<String> searchInferenceId = Parameter.stringParam(
             SEARCH_INFERENCE_ID_FIELD,
@@ -193,25 +184,9 @@ public static class Builder extends FieldMapper.Builder {
             }
         });
 
-        private final Parameter<MinimalServiceSettings> modelSettings = new Parameter<>(
-            MODEL_SETTINGS_FIELD,
-            true,
-            () -> null,
-            (n, c, o) -> SemanticTextField.parseModelSettingsFromMap(o),
-            mapper -> ((SemanticTextFieldType) mapper.fieldType()).modelSettings,
-            XContentBuilder::field,
-            Objects::toString
-        ).acceptsNull().setMergeValidator(SemanticTextFieldMapper::canMergeModelSettings);
+        private final Parameter<MinimalServiceSettings> modelSettings;
 
-        private final Parameter<SemanticTextIndexOptions> indexOptions = new Parameter<>(
-            INDEX_OPTIONS_FIELD,
-            true,
-            () -> null,
-            (n, c, o) -> parseIndexOptionsFromMap(n, o, c.indexVersionCreated()),
-            mapper -> ((SemanticTextFieldType) mapper.fieldType()).indexOptions,
-            XContentBuilder::field,
-            Objects::toString
-        ).acceptsNull();
+        private final Parameter<SemanticTextIndexOptions> indexOptions;
 
         @SuppressWarnings("unchecked")
         private final Parameter<ChunkingSettings> chunkingSettings = new Parameter<>(
@@ -248,6 +223,50 @@ public Builder(
             super(name);
             this.modelRegistry = modelRegistry;
             this.useLegacyFormat = InferenceMetadataFieldsMapper.isEnabled(indexSettings.getSettings()) == false;
+            this.indexVersionCreated = indexSettings.getIndexVersionCreated();
+
+            this.inferenceId = Parameter.stringParam(
+                INFERENCE_ID_FIELD,
+                false,
+                mapper -> ((SemanticTextFieldType) mapper.fieldType()).inferenceId,
+                DEFAULT_ELSER_2_INFERENCE_ID
+            ).addValidator(v -> {
+                if (Strings.isEmpty(v)) {
+                    throw new IllegalArgumentException(
+                        "[" + INFERENCE_ID_FIELD + "] on mapper [" + leafName() + "] of type [" + CONTENT_TYPE + "] must not be empty"
+                    );
+                }
+            }).alwaysSerialize();
+
+            this.modelSettings = new Parameter<>(
+                MODEL_SETTINGS_FIELD,
+                true,
+                () -> null,
+                (n, c, o) -> SemanticTextField.parseModelSettingsFromMap(o),
+                mapper -> ((SemanticTextFieldType) mapper.fieldType()).modelSettings,
+                XContentBuilder::field,
+                Objects::toString
+            ).acceptsNull().setMergeValidator(SemanticTextFieldMapper::canMergeModelSettings);
+
+            this.indexOptions = new Parameter<>(
+                INDEX_OPTIONS_FIELD,
+                true,
+                () -> null,
+                (n, c, o) -> parseIndexOptionsFromMap(n, o, c.indexVersionCreated()),
+                mapper -> ((SemanticTextFieldType) mapper.fieldType()).indexOptions,
+                (b, n, v) -> {
+                    if (v == null) {
+                        MinimalServiceSettings resolvedModelSettings = modelSettings.get() != null
+                            ? modelSettings.get()
+                            : modelRegistry.getMinimalServiceSettings(inferenceId.get());
+                        b.field(INDEX_OPTIONS_FIELD, defaultIndexOptions(indexVersionCreated, resolvedModelSettings));
+                    } else {
+                        b.field(INDEX_OPTIONS_FIELD, v);
+                    }
+                },
+                Objects::toString
+            ).acceptsNull().setSerializerCheck(this::indexOptionsSerializerCheck);
+
             this.inferenceFieldBuilder = c -> {
                 // Resolve the model setting from the registry if it has not been set yet.
                 var resolvedModelSettings = modelSettings.get() != null ? modelSettings.get() : getResolvedModelSettings(c, false);
@@ -263,6 +282,17 @@ public Builder(
             };
         }
 
+        private boolean indexOptionsSerializerCheck(boolean includeDefaults, boolean isConfigured, SemanticTextIndexOptions value) {
+            if (includeDefaults && value == null) {
+                MinimalServiceSettings resolvedModelSettings = modelSettings.get() != null
+                    ? modelSettings.get()
+                    : modelRegistry.getMinimalServiceSettings(inferenceId.get());
+                return defaultIndexOptions(indexVersionCreated, resolvedModelSettings) != null;
+            } else {
+                return isConfigured;
+            }
+        }
+
         public Builder setInferenceId(String id) {
             this.inferenceId.setValue(id);
             return this;
@@ -365,8 +395,9 @@ public SemanticTextFieldMapper build(MapperBuilderContext context) {
                 validateServiceSettings(modelSettings.get(), resolvedModelSettings);
             }
 
-            if (context.getMergeReason() != MapperService.MergeReason.MAPPING_RECOVERY && indexOptions.get() != null) {
-                validateIndexOptions(indexOptions.get(), inferenceId.getValue(), resolvedModelSettings);
+            SemanticTextIndexOptions builderIndexOptions = indexOptions.get();
+            if (context.getMergeReason() != MapperService.MergeReason.MAPPING_RECOVERY && builderIndexOptions != null) {
+                validateIndexOptions(builderIndexOptions, inferenceId.getValue(), resolvedModelSettings);
             }
 
             final String fullName = context.buildFullName(leafName());
@@ -1208,7 +1239,6 @@ static DenseVectorFieldMapper.DenseVectorIndexOptions defaultDenseVectorIndexOpt
         // As embedding models for text perform better with BBQ, we aggressively default semantic_text fields to use optimized index
         // options
         if (indexVersionDefaultsToBbqHnsw(indexVersionCreated)) {
-
             DenseVectorFieldMapper.DenseVectorIndexOptions defaultBbqHnswIndexOptions = defaultBbqHnswDenseVectorIndexOptions();
             return defaultBbqHnswIndexOptions.validate(modelSettings.elementType(), modelSettings.dimensions(), false)
                 ? defaultBbqHnswIndexOptions
@@ -1230,11 +1260,24 @@ static DenseVectorFieldMapper.DenseVectorIndexOptions defaultBbqHnswDenseVectorI
         return new DenseVectorFieldMapper.BBQHnswIndexOptions(m, efConstruction, rescoreVector);
     }
 
-    static SemanticTextIndexOptions defaultBbqHnswSemanticTextIndexOptions() {
-        return new SemanticTextIndexOptions(
-            SemanticTextIndexOptions.SupportedIndexOptions.DENSE_VECTOR,
-            defaultBbqHnswDenseVectorIndexOptions()
-        );
+    static SemanticTextIndexOptions defaultIndexOptions(IndexVersion indexVersionCreated, MinimalServiceSettings modelSettings) {
+
+        if (modelSettings == null) {
+            return null;
+        }
+
+        SemanticTextIndexOptions defaultIndexOptions = null;
+        if (modelSettings.taskType() == TaskType.TEXT_EMBEDDING) {
+            DenseVectorFieldMapper.DenseVectorIndexOptions denseVectorIndexOptions = defaultDenseVectorIndexOptions(
+                indexVersionCreated,
+                modelSettings
+            );
+            defaultIndexOptions = denseVectorIndexOptions == null
+                ? null
+                : new SemanticTextIndexOptions(SemanticTextIndexOptions.SupportedIndexOptions.DENSE_VECTOR, denseVectorIndexOptions);
+        }
+
+        return defaultIndexOptions;
     }
 
     private static boolean canMergeModelSettings(MinimalServiceSettings previous, MinimalServiceSettings current, Conflicts conflicts) {

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

@@ -20,6 +20,7 @@
 import java.util.Arrays;
 import java.util.Locale;
 import java.util.Map;
+import java.util.Objects;
 
 /**
  * Represents index options for a semantic_text field.
@@ -50,6 +51,18 @@ public IndexOptions indexOptions() {
         return indexOptions;
     }
 
+    @Override
+    public boolean equals(Object o) {
+        if (o instanceof SemanticTextIndexOptions == false) return false;
+        SemanticTextIndexOptions that = (SemanticTextIndexOptions) o;
+        return type == that.type && Objects.equals(indexOptions, that.indexOptions);
+    }
+
+    @Override
+    public int hashCode() {
+        return Objects.hash(type, indexOptions);
+    }
+
     public enum SupportedIndexOptions {
         DENSE_VECTOR("dense_vector") {
             @Override