[Inference API] Deprecate elser service

docs/changelog/113216.yaml

@@ -0,0 +1,10 @@
+pr: 113216
+summary: "[Inference API] Deprecate elser service"
+area: Machine Learning
+type: deprecation
+issues: []
+deprecation:
+  title: "[Inference API] Deprecate elser service"
+  area: REST API
+  details: The `elser` service of the inference API will be removed in an upcoming release. Please use the elasticsearch service instead. 
+  impact: In the current version there is no impact. In a future version, users of the `elser` service will no longer be able to use it, and will be required to use the `elasticsearch` service to access elser through the inference API.

server/src/main/java/org/elasticsearch/inference/InferenceServiceRegistry.java

@@ -46,7 +46,13 @@ public Map<String, InferenceService> getServices() {
     }
 
     public Optional<InferenceService> getService(String serviceName) {
-        return Optional.ofNullable(services.get(serviceName));
+
+        if ("elser".equals(serviceName)) { // ElserService.NAME before removal
+            // here we are aliasing the elser service to use the elasticsearch service instead
+            return Optional.ofNullable(services.get("elasticsearch")); // ElasticsearchInternalService.NAME
+        } else {
+            return Optional.ofNullable(services.get(serviceName));
+        }
     }
 
     public List<NamedWriteableRegistry.Entry> getNamedWriteables() {

x-pack/plugin/inference/src/internalClusterTest/java/org/elasticsearch/xpack/inference/integration/ModelRegistryIT.java

@@ -27,11 +27,10 @@
 import org.elasticsearch.xcontent.XContentBuilder;
 import org.elasticsearch.xpack.inference.InferencePlugin;
 import org.elasticsearch.xpack.inference.registry.ModelRegistry;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalModel;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalService;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalServiceSettingsTests;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalServiceTests;
-import org.elasticsearch.xpack.inference.services.elser.ElserMlNodeTaskSettingsTests;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalService;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserInternalModel;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserInternalServiceSettingsTests;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserMlNodeTaskSettingsTests;
 import org.junit.Before;
 
 import java.io.IOException;
@@ -117,8 +116,10 @@ public void testGetModel() throws Exception {
 
         assertEquals(model.getConfigurations().getService(), modelHolder.get().service());
 
-        var elserService = new ElserInternalService(new InferenceServiceExtension.InferenceServiceFactoryContext(mock(Client.class)));
-        ElserInternalModel roundTripModel = elserService.parsePersistedConfigWithSecrets(
+        var elserService = new ElasticsearchInternalService(
+            new InferenceServiceExtension.InferenceServiceFactoryContext(mock(Client.class))
+        );
+        ElserInternalModel roundTripModel = (ElserInternalModel) elserService.parsePersistedConfigWithSecrets(
             modelHolder.get().inferenceEntityId(),
             modelHolder.get().taskType(),
             modelHolder.get().settings(),
@@ -274,7 +275,17 @@ public void testGetModelWithSecrets() throws InterruptedException {
     }
 
     private Model buildElserModelConfig(String inferenceEntityId, TaskType taskType) {
-        return ElserInternalServiceTests.randomModelConfig(inferenceEntityId, taskType);
+        return switch (taskType) {
+            case SPARSE_EMBEDDING -> new org.elasticsearch.xpack.inference.services.elasticsearch.ElserInternalModel(
+                inferenceEntityId,
+                taskType,
+                ElasticsearchInternalService.NAME,
+                ElserInternalServiceSettingsTests.createRandom(),
+                ElserMlNodeTaskSettingsTests.createRandom()
+            );
+            default -> throw new IllegalArgumentException("task type " + taskType + " is not supported");
+        };
+
     }
 
     protected <T> void blockingCall(Consumer<ActionListener<T>> function, AtomicReference<T> response, AtomicReference<Exception> error)
@@ -297,7 +308,7 @@ private static Model buildModelWithUnknownField(String inferenceEntityId) {
             new ModelWithUnknownField(
                 inferenceEntityId,
                 TaskType.SPARSE_EMBEDDING,
-                ElserInternalService.NAME,
+                ElasticsearchInternalService.NAME,
                 ElserInternalServiceSettingsTests.createRandom(),
                 ElserMlNodeTaskSettingsTests.createRandom()
             )

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

@@ -64,9 +64,9 @@
 import org.elasticsearch.xpack.inference.services.elasticsearch.CustomElandInternalTextEmbeddingServiceSettings;
 import org.elasticsearch.xpack.inference.services.elasticsearch.CustomElandRerankTaskSettings;
 import org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalServiceSettings;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserInternalServiceSettings;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserMlNodeTaskSettings;
 import org.elasticsearch.xpack.inference.services.elasticsearch.MultilingualE5SmallInternalServiceSettings;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalServiceSettings;
-import org.elasticsearch.xpack.inference.services.elser.ElserMlNodeTaskSettings;
 import org.elasticsearch.xpack.inference.services.googleaistudio.completion.GoogleAiStudioCompletionServiceSettings;
 import org.elasticsearch.xpack.inference.services.googleaistudio.embeddings.GoogleAiStudioEmbeddingsServiceSettings;
 import org.elasticsearch.xpack.inference.services.googlevertexai.GoogleVertexAiSecretSettings;

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

@@ -86,7 +86,6 @@
 import org.elasticsearch.xpack.inference.services.elastic.ElasticInferenceServiceFeature;
 import org.elasticsearch.xpack.inference.services.elastic.ElasticInferenceServiceSettings;
 import org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalService;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalService;
 import org.elasticsearch.xpack.inference.services.googleaistudio.GoogleAiStudioService;
 import org.elasticsearch.xpack.inference.services.googlevertexai.GoogleVertexAiService;
 import org.elasticsearch.xpack.inference.services.huggingface.HuggingFaceService;
@@ -229,7 +228,6 @@ public void loadExtensions(ExtensionLoader loader) {
 
     public List<InferenceServiceExtension.Factory> getInferenceServiceFactories() {
         return List.of(
-            ElserInternalService::new,
             context -> new HuggingFaceElserService(httpFactory.get(), serviceComponents.get()),
             context -> new HuggingFaceService(httpFactory.get(), serviceComponents.get()),
             context -> new OpenAiService(httpFactory.get(), serviceComponents.get()),

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

@@ -11,6 +11,8 @@
 import org.elasticsearch.action.ActionListener;
 import org.elasticsearch.action.support.ActionFilters;
 import org.elasticsearch.action.support.HandledTransportAction;
+import org.elasticsearch.common.logging.DeprecationCategory;
+import org.elasticsearch.common.logging.DeprecationLogger;
 import org.elasticsearch.common.util.concurrent.EsExecutors;
 import org.elasticsearch.common.xcontent.ChunkedToXContent;
 import org.elasticsearch.inference.InferenceService;
@@ -25,12 +27,14 @@
 import org.elasticsearch.xpack.core.inference.action.InferenceAction;
 import org.elasticsearch.xpack.inference.action.task.StreamingTaskManager;
 import org.elasticsearch.xpack.inference.registry.ModelRegistry;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalService;
 import org.elasticsearch.xpack.inference.telemetry.InferenceStats;
 
 import java.util.Set;
 import java.util.stream.Collectors;
 
 import static org.elasticsearch.core.Strings.format;
+import static org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalService.OLD_ELSER_SERVICE_NAME;
 
 public class TransportInferenceAction extends HandledTransportAction<InferenceAction.Request, InferenceAction.Response> {
     private static final String STREAMING_INFERENCE_TASK_TYPE = "streaming_inference";
@@ -42,6 +46,7 @@ public class TransportInferenceAction extends HandledTransportAction<InferenceAc
     private final InferenceServiceRegistry serviceRegistry;
     private final InferenceStats inferenceStats;
     private final StreamingTaskManager streamingTaskManager;
+    private static final DeprecationLogger DEPRECATION_LOGGER = DeprecationLogger.getLogger(TransportInferenceAction.class);
 
     @Inject
     public TransportInferenceAction(
@@ -75,6 +80,15 @@ protected void doExecute(Task task, InferenceAction.Request request, ActionListe
                 );
                 return;
             }
+            if (service.get().name().equals(OLD_ELSER_SERVICE_NAME)) {
+                DEPRECATION_LOGGER.warn(
+                    DeprecationCategory.API,
+                    "inference_api_elser_service",
+                    "The [{}] service is deprecated and will be removed in a future release. Use the [{}] service instead.",
+                    OLD_ELSER_SERVICE_NAME,
+                    ElasticsearchInternalService.NAME
+                );
+            }
 
             if (request.getTaskType().isAnyOrSame(unparsedModel.taskType()) == false) {
                 // not the wildcard task type and not the model task type

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

@@ -20,6 +20,8 @@
 import org.elasticsearch.cluster.block.ClusterBlockLevel;
 import org.elasticsearch.cluster.metadata.IndexNameExpressionResolver;
 import org.elasticsearch.cluster.service.ClusterService;
+import org.elasticsearch.common.logging.DeprecationCategory;
+import org.elasticsearch.common.logging.DeprecationLogger;
 import org.elasticsearch.common.settings.ClusterSettings;
 import org.elasticsearch.common.settings.Settings;
 import org.elasticsearch.common.util.concurrent.EsExecutors;
@@ -45,18 +47,24 @@
 import org.elasticsearch.xpack.core.ml.utils.MlPlatformArchitecturesUtil;
 import org.elasticsearch.xpack.inference.InferencePlugin;
 import org.elasticsearch.xpack.inference.registry.ModelRegistry;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalService;
 
 import java.io.IOException;
 import java.util.Map;
 import java.util.Set;
 
 import static org.elasticsearch.core.Strings.format;
+import static org.elasticsearch.xpack.inference.services.elasticsearch.BaseElasticsearchInternalService.selectDefaultModelVariantBasedOnClusterArchitecture;
+import static org.elasticsearch.xpack.inference.services.elasticsearch.ElasticsearchInternalService.OLD_ELSER_SERVICE_NAME;
+import static org.elasticsearch.xpack.inference.services.elasticsearch.ElserModels.ELSER_V2_MODEL;
+import static org.elasticsearch.xpack.inference.services.elasticsearch.ElserModels.ELSER_V2_MODEL_LINUX_X86;
 
 public class TransportPutInferenceModelAction extends TransportMasterNodeAction<
     PutInferenceModelAction.Request,
     PutInferenceModelAction.Response> {
 
     private static final Logger logger = LogManager.getLogger(TransportPutInferenceModelAction.class);
+    private static final DeprecationLogger DEPRECATION_LOGGER = DeprecationLogger.getLogger(PutInferenceModelAction.class);
 
     private final ModelRegistry modelRegistry;
     private final InferenceServiceRegistry serviceRegistry;
@@ -157,9 +165,31 @@ protected void masterOperation(
         }
 
         if (service.get().isInClusterService()) {
+
+            // required for BWC of elser service in elasticsearch service TODO remove when elser service deprecated
+            requestAsMap.put(ModelConfigurations.SERVICE, serviceName);
+
             // Find the cluster platform as the service may need that
             // information when creating the model
             MlPlatformArchitecturesUtil.getMlNodesArchitecturesSet(listener.delegateFailureAndWrap((delegate, architectures) -> {
+                if (serviceName.equals(OLD_ELSER_SERVICE_NAME)) {
+                    String modelId = selectDefaultModelVariantBasedOnClusterArchitecture(
+                        architectures,
+                        ELSER_V2_MODEL_LINUX_X86,
+                        ELSER_V2_MODEL
+                    );
+
+                    DEPRECATION_LOGGER.warn(
+                        DeprecationCategory.API,
+                        "inference_api_elser_service",
+                        "The [{}] service is deprecated and will be removed in a future release. Use the [{}] service instead, with"
+                            + " [model_id] set to [{}] in the [service_settings]",
+                        OLD_ELSER_SERVICE_NAME,
+                        ElasticsearchInternalService.NAME,
+                        modelId
+                    );
+                }
+
                 if (architectures.isEmpty() && clusterIsInElasticCloud(clusterService.getClusterSettings())) {
                     parseAndStoreModel(
                         service.get(),
@@ -224,7 +254,6 @@ private void parseAndStoreModel(
         });
 
         service.parseRequestConfig(inferenceEntityId, taskType, config, platformArchitectures, parsedModelListener);
-
     }
 
     private void putAndStartModel(InferenceService service, Model model, ActionListener<PutInferenceModelAction.Response> finalListener) {

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/elastic/ElasticInferenceServiceSparseEmbeddingsModel.java

@@ -18,7 +18,7 @@
 import org.elasticsearch.xpack.inference.external.action.ExecutableAction;
 import org.elasticsearch.xpack.inference.external.action.elastic.ElasticInferenceServiceActionVisitor;
 import org.elasticsearch.xpack.inference.services.ConfigurationParseContext;
-import org.elasticsearch.xpack.inference.services.elser.ElserModels;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserModels;
 
 import java.net.URI;
 import java.net.URISyntaxException;

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/elastic/ElasticInferenceServiceSparseEmbeddingsServiceSettings.java

@@ -17,7 +17,7 @@
 import org.elasticsearch.inference.ServiceSettings;
 import org.elasticsearch.xcontent.XContentBuilder;
 import org.elasticsearch.xpack.inference.services.ConfigurationParseContext;
-import org.elasticsearch.xpack.inference.services.elser.ElserModels;
+import org.elasticsearch.xpack.inference.services.elasticsearch.ElserModels;
 import org.elasticsearch.xpack.inference.services.settings.FilteredXContentObject;
 import org.elasticsearch.xpack.inference.services.settings.RateLimitSettings;
 

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

@@ -28,7 +28,6 @@
 import org.elasticsearch.xpack.core.ml.inference.TrainedModelInput;
 import org.elasticsearch.xpack.core.ml.inference.TrainedModelPrefixStrings;
 import org.elasticsearch.xpack.core.ml.inference.trainedmodel.InferenceConfigUpdate;
-import org.elasticsearch.xpack.inference.services.elser.ElserInternalModel;
 
 import java.io.IOException;
 import java.util.EnumSet;