refactor(semantic_text)raise mapper_parsing_exception for incompatible sparse_v index versions

Author: mromaios

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

@@ -101,6 +101,7 @@
 import java.util.function.Function;
 import java.util.function.Supplier;
 
+import static org.elasticsearch.index.IndexVersions.NEW_SPARSE_VECTOR;
 import static org.elasticsearch.index.IndexVersions.SEMANTIC_TEXT_DEFAULTS_TO_BBQ;
 import static org.elasticsearch.index.IndexVersions.SEMANTIC_TEXT_DEFAULTS_TO_BBQ_BACKPORT_8_X;
 import static org.elasticsearch.inference.TaskType.SPARSE_EMBEDDING;
@@ -155,6 +156,8 @@ public class SemanticTextFieldMapper extends FieldMapper implements InferenceFie
     public static final float DEFAULT_RESCORE_OVERSAMPLE = 3.0f;
 
     static final String INDEX_OPTIONS_FIELD = "index_options";
+    public static final String ERROR_MESSAGE_UNSUPPORTED_SPARSE_VECTOR = "Creating a [semantic_text] field with [sparse_vector] models"
+        + " is not supported on this index. Try using a [dense_vector] model or create a new index with version 8.11+.";
 
     public static final TypeParser parser(Supplier<ModelRegistry> modelRegistry) {
         return new TypeParser(
@@ -1246,6 +1249,9 @@ private static Mapper.Builder createEmbeddingsField(
     ) {
         return switch (modelSettings.taskType()) {
             case SPARSE_EMBEDDING -> {
+                if (indexVersionCreated.before(NEW_SPARSE_VECTOR)) {
+                    throw new IllegalArgumentException(ERROR_MESSAGE_UNSUPPORTED_SPARSE_VECTOR);
+                }
                 SparseVectorFieldMapper.Builder sparseVectorMapperBuilder = new SparseVectorFieldMapper.Builder(
                     CHUNKED_EMBEDDINGS_FIELD,
                     indexVersionCreated,

x-pack/plugin/inference/src/test/java/org/elasticsearch/xpack/inference/mapper/SemanticTextFieldMapperTests.java

@@ -110,6 +110,7 @@
 import static org.elasticsearch.xpack.inference.mapper.SemanticTextField.getEmbeddingsFieldName;
 import static org.elasticsearch.xpack.inference.mapper.SemanticTextFieldMapper.DEFAULT_ELSER_2_INFERENCE_ID;
 import static org.elasticsearch.xpack.inference.mapper.SemanticTextFieldMapper.DEFAULT_RESCORE_OVERSAMPLE;
+import static org.elasticsearch.xpack.inference.mapper.SemanticTextFieldMapper.ERROR_MESSAGE_UNSUPPORTED_SPARSE_VECTOR;
 import static org.elasticsearch.xpack.inference.mapper.SemanticTextFieldMapper.INDEX_OPTIONS_FIELD;
 import static org.elasticsearch.xpack.inference.mapper.SemanticTextFieldTests.generateRandomChunkingSettings;
 import static org.elasticsearch.xpack.inference.mapper.SemanticTextFieldTests.generateRandomChunkingSettingsOtherThan;
@@ -1748,6 +1749,28 @@ public static void addSemanticTextInferenceResults(
         }
     }
 
+    public void testOldIndexSemanticTextSparseVersionRaisesError() throws IOException {
+        final XContentBuilder fieldMapping = fieldMapping(b -> {
+            b.field("type", "semantic_text");
+            b.field("inference_id", "another_inference_id");
+            b.startObject("model_settings");
+            b.field("task_type", "sparse_embedding");
+            b.endObject();
+        });
+
+        MapperParsingException exception = assertThrows(
+            MapperParsingException.class,
+            () -> createMapperService(
+                fieldMapping,
+                true,
+                IndexVersions.V_8_0_0,
+                IndexVersionUtils.getPreviousVersion(IndexVersions.NEW_SPARSE_VECTOR)
+            )
+        );
+        assertTrue(exception.getMessage().contains(ERROR_MESSAGE_UNSUPPORTED_SPARSE_VECTOR));
+        assertTrue(exception.getRootCause() instanceof IllegalArgumentException);
+    }
+
     static String randomFieldName(int numLevel) {
         StringBuilder builder = new StringBuilder();
         for (int i = 0; i < numLevel; i++) {