Revert "Default new semantic_text fields to use BBQ when models are compatible (elastic#126629)"

This reverts commit a72883e.

Author: kderusso

docs/changelog/126629.yaml

@@ -160,7 +160,6 @@ private static Version parseUnchecked(String version) {
     public static final IndexVersion SYNTHETIC_SOURCE_STORE_ARRAYS_NATIVELY_SCALED_FLOAT = def(9_020_0_00, Version.LUCENE_10_1_0);
     public static final IndexVersion USE_LUCENE101_POSTINGS_FORMAT = def(9_021_0_00, Version.LUCENE_10_1_0);
     public static final IndexVersion UPGRADE_TO_LUCENE_10_2_0 = def(9_022_00_0, Version.LUCENE_10_2_0);
-    public static final IndexVersion SEMANTIC_TEXT_DEFAULTS_TO_BBQ = def(9_023_0_00, Version.LUCENE_10_2_0);
     /*
      * STOP! READ THIS FIRST! No, really,
      *        ____ _____ ___  ____  _        ____  _____    _    ____    _____ _   _ ___ ____    _____ ___ ____  ____ _____ _

server/src/main/java/org/elasticsearch/index/IndexVersions.java

@@ -290,11 +290,6 @@ public Builder elementType(ElementType elementType) {
             return this;
         }
 
-        public Builder indexOptions(IndexOptions indexOptions) {
-            this.indexOptions.setValue(indexOptions);
-            return this;
-        }
-
         @Override
         public DenseVectorFieldMapper build(MapperBuilderContext context) {
             // Validate again here because the dimensions or element type could have been set programmatically,
@@ -1226,7 +1221,7 @@ public final String toString() {
         public abstract VectorSimilarityFunction vectorSimilarityFunction(IndexVersion indexVersion, ElementType elementType);
     }
 
-    public abstract static class IndexOptions implements ToXContent {
+    abstract static class IndexOptions implements ToXContent {
         final VectorIndexType type;
 
         IndexOptions(VectorIndexType type) {
@@ -1235,36 +1230,21 @@ public abstract static class IndexOptions implements ToXContent {
 
         abstract KnnVectorsFormat getVectorsFormat(ElementType elementType);
 
-        public boolean validate(ElementType elementType, int dim, boolean throwOnError) {
-            return validateElementType(elementType, throwOnError) && validateDimension(dim, throwOnError);
-        }
-
-        public boolean validateElementType(ElementType elementType) {
-            return validateElementType(elementType, true);
-        }
-
-        final boolean validateElementType(ElementType elementType, boolean throwOnError) {
-            boolean validElementType = type.supportsElementType(elementType);
-            if (throwOnError && validElementType == false) {
+        final void validateElementType(ElementType elementType) {
+            if (type.supportsElementType(elementType) == false) {
                 throw new IllegalArgumentException(
                     "[element_type] cannot be [" + elementType.toString() + "] when using index type [" + type + "]"
                 );
             }
-            return validElementType;
         }
 
         abstract boolean updatableTo(IndexOptions update);
 
-        public boolean validateDimension(int dim) {
-            return validateDimension(dim, true);
-        }
-
-        public boolean validateDimension(int dim, boolean throwOnError) {
-            boolean supportsDimension = type.supportsDimension(dim);
-            if (throwOnError && supportsDimension == false) {
-                throw new IllegalArgumentException(type.name + " only supports even dimensions; provided=" + dim);
+        public void validateDimension(int dim) {
+            if (type.supportsDimension(dim)) {
+                return;
             }
-            return supportsDimension;
+            throw new IllegalArgumentException(type.name + " only supports even dimensions; provided=" + dim);
         }
 
         abstract boolean doEquals(IndexOptions other);
@@ -1767,12 +1747,12 @@ boolean updatableTo(IndexOptions update) {
 
     }
 
-    public static class Int8HnswIndexOptions extends QuantizedIndexOptions {
+    static class Int8HnswIndexOptions extends QuantizedIndexOptions {
         private final int m;
         private final int efConstruction;
         private final Float confidenceInterval;
 
-        public Int8HnswIndexOptions(int m, int efConstruction, Float confidenceInterval, RescoreVector rescoreVector) {
+        Int8HnswIndexOptions(int m, int efConstruction, Float confidenceInterval, RescoreVector rescoreVector) {
             super(VectorIndexType.INT8_HNSW, rescoreVector);
             this.m = m;
             this.efConstruction = efConstruction;
@@ -1910,11 +1890,11 @@ public String toString() {
         }
     }
 
-    public static class BBQHnswIndexOptions extends QuantizedIndexOptions {
+    static class BBQHnswIndexOptions extends QuantizedIndexOptions {
         private final int m;
         private final int efConstruction;
 
-        public BBQHnswIndexOptions(int m, int efConstruction, RescoreVector rescoreVector) {
+        BBQHnswIndexOptions(int m, int efConstruction, RescoreVector rescoreVector) {
             super(VectorIndexType.BBQ_HNSW, rescoreVector);
             this.m = m;
             this.efConstruction = efConstruction;
@@ -1956,14 +1936,11 @@ public XContentBuilder toXContent(XContentBuilder builder, Params params) throws
         }
 
         @Override
-        public boolean validateDimension(int dim, boolean throwOnError) {
-            boolean supportsDimension = type.supportsDimension(dim);
-            if (throwOnError && supportsDimension == false) {
-                throw new IllegalArgumentException(
-                    type.name + " does not support dimensions fewer than " + BBQ_MIN_DIMS + "; provided=" + dim
-                );
+        public void validateDimension(int dim) {
+            if (type.supportsDimension(dim)) {
+                return;
             }
-            return supportsDimension;
+            throw new IllegalArgumentException(type.name + " does not support dimensions fewer than " + BBQ_MIN_DIMS + "; provided=" + dim);
         }
     }
 
@@ -2007,19 +1984,15 @@ public XContentBuilder toXContent(XContentBuilder builder, Params params) throws
         }
 
         @Override
-        public boolean validateDimension(int dim, boolean throwOnError) {
-            boolean supportsDimension = type.supportsDimension(dim);
-            if (throwOnError && supportsDimension == false) {
-                throw new IllegalArgumentException(
-                    type.name + " does not support dimensions fewer than " + BBQ_MIN_DIMS + "; provided=" + dim
-                );
+        public void validateDimension(int dim) {
+            if (type.supportsDimension(dim)) {
+                return;
             }
-            return supportsDimension;
+            throw new IllegalArgumentException(type.name + " does not support dimensions fewer than " + BBQ_MIN_DIMS + "; provided=" + dim);
         }
-
     }
 
-    public record RescoreVector(float oversample) implements ToXContentObject {
+    record RescoreVector(float oversample) implements ToXContentObject {
         static final String NAME = "rescore_vector";
         static final String OVERSAMPLE = "oversample";
 
@@ -2338,10 +2311,6 @@ int getVectorDimensions() {
         ElementType getElementType() {
             return elementType;
         }
-
-        public IndexOptions getIndexOptions() {
-            return indexOptions;
-        }
     }
 
     private final IndexOptions indexOptions;

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

@@ -207,13 +207,6 @@ protected final MapperService createMapperService(Settings settings, String mapp
         return mapperService;
     }
 
-    protected final MapperService createMapperService(IndexVersion indexVersion, Settings settings, XContentBuilder mappings)
-        throws IOException {
-        MapperService mapperService = createMapperService(indexVersion, settings, () -> true, mappings);
-        merge(mapperService, mappings);
-        return mapperService;
-    }
-
     protected final MapperService createMapperService(IndexVersion version, XContentBuilder mapping) throws IOException {
         return createMapperService(version, getIndexSettings(), () -> true, mapping);
     }

test/framework/src/main/java/org/elasticsearch/index/mapper/MapperServiceTestCase.java

@@ -9,7 +9,6 @@
 
 import org.apache.logging.log4j.LogManager;
 import org.apache.logging.log4j.Logger;
-import org.apache.lucene.codecs.lucene99.Lucene99HnswVectorsFormat;
 import org.apache.lucene.index.FieldInfos;
 import org.apache.lucene.index.LeafReaderContext;
 import org.apache.lucene.search.DocIdSetIterator;
@@ -96,7 +95,6 @@
 import java.util.function.Function;
 import java.util.function.Supplier;
 
-import static org.elasticsearch.index.IndexVersions.SEMANTIC_TEXT_DEFAULTS_TO_BBQ;
 import static org.elasticsearch.inference.TaskType.SPARSE_EMBEDDING;
 import static org.elasticsearch.inference.TaskType.TEXT_EMBEDDING;
 import static org.elasticsearch.search.SearchService.DEFAULT_SIZE;
@@ -135,8 +133,6 @@ public class SemanticTextFieldMapper extends FieldMapper implements InferenceFie
     public static final String CONTENT_TYPE = "semantic_text";
     public static final String DEFAULT_ELSER_2_INFERENCE_ID = DEFAULT_ELSER_ID;
 
-    public static final float DEFAULT_RESCORE_OVERSAMPLE = 3.0f;
-
     public static final TypeParser parser(Supplier<ModelRegistry> modelRegistry) {
         return new TypeParser(
             (n, c) -> new Builder(n, c::bitSetProducer, c.getIndexSettings(), modelRegistry.get()),
@@ -1058,30 +1054,12 @@ private static Mapper.Builder createEmbeddingsField(
                 denseVectorMapperBuilder.dimensions(modelSettings.dimensions());
                 denseVectorMapperBuilder.elementType(modelSettings.elementType());
 
-                DenseVectorFieldMapper.IndexOptions defaultIndexOptions = null;
-                if (indexVersionCreated.onOrAfter(SEMANTIC_TEXT_DEFAULTS_TO_BBQ)) {
-                    defaultIndexOptions = defaultSemanticDenseIndexOptions();
-                }
-                if (defaultIndexOptions != null
-                    && defaultIndexOptions.validate(modelSettings.elementType(), modelSettings.dimensions(), false)) {
-                    denseVectorMapperBuilder.indexOptions(defaultIndexOptions);
-                }
-
                 yield denseVectorMapperBuilder;
             }
             default -> throw new IllegalArgumentException("Invalid task_type in model_settings [" + modelSettings.taskType().name() + "]");
         };
     }
 
-    static DenseVectorFieldMapper.IndexOptions defaultSemanticDenseIndexOptions() {
-        // As embedding models for text perform better with BBQ, we aggressively default semantic_text fields to use optimized index
-        // options outside of dense_vector defaults
-        int m = Lucene99HnswVectorsFormat.DEFAULT_MAX_CONN;
-        int efConstruction = Lucene99HnswVectorsFormat.DEFAULT_BEAM_WIDTH;
-        DenseVectorFieldMapper.RescoreVector rescoreVector = new DenseVectorFieldMapper.RescoreVector(DEFAULT_RESCORE_OVERSAMPLE);
-        return new DenseVectorFieldMapper.BBQHnswIndexOptions(m, efConstruction, rescoreVector);
-    }
-
     private static boolean canMergeModelSettings(MinimalServiceSettings previous, MinimalServiceSettings current, Conflicts conflicts) {
         if (previous != null && current != null && previous.canMergeWith(current)) {
             return true;

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

@@ -37,10 +37,7 @@ public void testIsEnabled() {
         assertFalse(InferenceMetadataFieldsMapper.isEnabled(settings));
 
         settings = Settings.builder()
-            .put(
-                IndexMetadata.SETTING_INDEX_VERSION_CREATED.getKey(),
-                getRandomCompatibleIndexVersion(true, IndexVersionUtils.getPreviousVersion(IndexVersions.INFERENCE_METADATA_FIELDS))
-            )
+            .put(IndexMetadata.SETTING_INDEX_VERSION_CREATED.getKey(), getRandomCompatibleIndexVersion(true))
             .put(InferenceMetadataFieldsMapper.USE_LEGACY_SEMANTIC_TEXT_FORMAT.getKey(), false)
             .build();
         assertFalse(InferenceMetadataFieldsMapper.isEnabled(settings));
@@ -117,18 +114,18 @@ public MappedFieldType getMappedFieldType() {
     }
 
     static IndexVersion getRandomCompatibleIndexVersion(boolean useLegacyFormat) {
-        return getRandomCompatibleIndexVersion(useLegacyFormat, IndexVersion.current());
-    }
-
-    static IndexVersion getRandomCompatibleIndexVersion(boolean useLegacyFormat, IndexVersion maxVersion) {
         if (useLegacyFormat) {
             if (randomBoolean()) {
-                return IndexVersionUtils.randomVersionBetween(random(), IndexVersions.UPGRADE_TO_LUCENE_10_0_0, maxVersion);
+                return IndexVersionUtils.randomVersionBetween(
+                    random(),
+                    IndexVersions.UPGRADE_TO_LUCENE_10_0_0,
+                    IndexVersionUtils.getPreviousVersion(IndexVersions.INFERENCE_METADATA_FIELDS)
+                );
             }
             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, IndexVersion.current());
             }
             return IndexVersionUtils.randomVersionBetween(
                 random(),
@@ -137,5 +134,4 @@ static IndexVersion getRandomCompatibleIndexVersion(boolean useLegacyFormat, Ind
             );
         }
     }
-
 }