Add Ibm Granite Completion and Chat Completion support

docs/changelog/129146.yaml

@@ -0,0 +1,5 @@
+pr: 129146
+summary: "[ML] Add IBM watsonx Completion and Chat Completion support to the Inference Plugin"
+area: Machine Learning
+type: enhancement
+issues: []

server/src/main/java/org/elasticsearch/TransportVersions.java

@@ -299,7 +299,10 @@ static TransportVersion def(int id) {
     public static final TransportVersion HEAP_USAGE_IN_CLUSTER_INFO = def(9_096_0_00);
     public static final TransportVersion NONE_CHUNKING_STRATEGY = def(9_097_0_00);
     public static final TransportVersion PROJECT_DELETION_GLOBAL_BLOCK = def(9_098_0_00);
-    public static final TransportVersion ML_INFERENCE_IBM_WATSONX_COMPLETION_ADDED = def(9_099_0_00);
+    public static final TransportVersion SECURITY_CLOUD_API_KEY_REALM_AND_TYPE = def(9_099_0_00);
+    public static final TransportVersion STATE_PARAM_GET_SNAPSHOT = def(9_100_0_00);
+    public static final TransportVersion PROJECT_ID_IN_SNAPSHOTS_DELETIONS_AND_REPO_CLEANUP = def(9_101_0_00);
+    public static final TransportVersion ML_INFERENCE_IBM_WATSONX_COMPLETION_ADDED = def(9_102_0_00);
 
     /*
      * STOP! READ THIS FIRST! No, really,

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

@@ -16,7 +16,7 @@ public class IbmWatsonxCompletionResponseHandler extends OpenAiChatCompletionRes
     /**
      * Constructs a IbmWatsonxCompletionResponseHandler with the specified request type and response parser.
      *
-     * @param requestType The type of request being handled (e.g., "IBM Watsonx completions").
+     * @param requestType The type of request being handled (e.g., "IBM watsonx completions").
      * @param parseFunction The function to parse the response.
      */
     public IbmWatsonxCompletionResponseHandler(String requestType, ResponseParser parseFunction) {

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

@@ -35,7 +35,7 @@ public class IbmWatsonxEmbeddingsRequestManager extends IbmWatsonxRequestManager
     private static final ResponseHandler HANDLER = createEmbeddingsHandler();
 
     private static ResponseHandler createEmbeddingsHandler() {
-        return new IbmWatsonxResponseHandler("IBM Watsonx embeddings", IbmWatsonxEmbeddingsResponseEntity::fromResponse);
+        return new IbmWatsonxResponseHandler("IBM watsonx embeddings", IbmWatsonxEmbeddingsResponseEntity::fromResponse);
     }
 
     private final IbmWatsonxEmbeddingsModel model;

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

@@ -16,16 +16,13 @@
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.action.IbmWatsonxActionVisitor;
 import org.elasticsearch.xpack.inference.services.settings.RateLimitSettings;
 
-import java.net.URI;
 import java.util.Map;
 import java.util.Objects;
 
 public abstract class IbmWatsonxModel extends RateLimitGroupingModel {
 
     private final IbmWatsonxRateLimitServiceSettings rateLimitServiceSettings;
 
-    protected URI uri;
-
     public IbmWatsonxModel(
         ModelConfigurations configurations,
         ModelSecrets secrets,
@@ -56,7 +53,7 @@ public IbmWatsonxRateLimitServiceSettings rateLimitServiceSettings() {
 
     @Override
     public int rateLimitGroupingHash() {
-        return Objects.hash(uri);
+        return Objects.hash(this.rateLimitServiceSettings);
     }
 
     @Override

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

@@ -31,7 +31,7 @@ public class IbmWatsonxRerankRequestManager extends IbmWatsonxRequestManager {
 
     private static ResponseHandler createIbmWatsonxResponseHandler() {
         return new IbmWatsonxResponseHandler(
-            "IBM Watsonx rerank",
+            "IBM watsonx rerank",
             (request, response) -> IbmWatsonxRankedResponseEntity.fromResponse(response)
         );
     }

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

@@ -72,14 +72,14 @@ public class IbmWatsonxService extends SenderService {
 
     public static final String NAME = "watsonxai";
 
-    private static final String SERVICE_NAME = "IBM Watsonx";
+    private static final String SERVICE_NAME = "IBM watsonx";
     private static final EnumSet<TaskType> supportedTaskTypes = EnumSet.of(
         TaskType.TEXT_EMBEDDING,
         TaskType.COMPLETION,
         TaskType.CHAT_COMPLETION
     );
     private static final ResponseHandler UNIFIED_CHAT_COMPLETION_HANDLER = new IbmWatsonUnifiedChatCompletionResponseHandler(
-        "IBM Watsonx chat completions",
+        "IBM watsonx chat completions",
         OpenAiChatCompletionResponseEntity::fromResponse
     );
 
@@ -375,7 +375,7 @@ public static InferenceServiceConfiguration get() {
 
                 configurationMap.put(
                     API_VERSION,
-                    new SettingsConfiguration.Builder(supportedTaskTypes).setDescription("The IBM Watsonx API version ID to use.")
+                    new SettingsConfiguration.Builder(supportedTaskTypes).setDescription("The IBM watsonx API version ID to use.")
                         .setLabel("API Version")
                         .setRequired(true)
                         .setSensitive(false)

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

@@ -11,7 +11,7 @@ public class IbmWatsonxServiceFields {
 
     /**
      * Taken from <a href="https://cloud.ibm.com/apidocs/watsonx-ai#text-embeddings">
-     *     Introduction to IBM Watsonx.ai as a Service</a>
+     *     Introduction to IBM watsonx.ai as a Service</a>
      */
     static final int EMBEDDING_MAX_BATCH_SIZE = 1000;
     public static final String API_VERSION = "api_version";

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/action/IbmWatsonxActionCreator.java

@@ -42,7 +42,7 @@ public class IbmWatsonxActionCreator implements IbmWatsonxActionVisitor {
     private final Sender sender;
     private final ServiceComponents serviceComponents;
 
-    static final String COMPLETION_REQUEST_TYPE = "IBM Watsonx completions";
+    static final String COMPLETION_REQUEST_TYPE = "IBM watsonx completions";
     static final String USER_ROLE = "user";
     static final ResponseHandler COMPLETION_HANDLER = new IbmWatsonxCompletionResponseHandler(
         COMPLETION_REQUEST_TYPE,
@@ -56,7 +56,7 @@ public IbmWatsonxActionCreator(Sender sender, ServiceComponents serviceComponent
 
     @Override
     public ExecutableAction create(IbmWatsonxEmbeddingsModel model, Map<String, Object> taskSettings) {
-        var failedToSendRequestErrorMessage = constructFailedToSendRequestMessage("IBM Watsonx embeddings");
+        var failedToSendRequestErrorMessage = constructFailedToSendRequestMessage("IBM watsonx embeddings");
         return new SenderExecutableAction(
             sender,
             getEmbeddingsRequestManager(model, serviceComponents.truncator(), serviceComponents.threadPool()),
@@ -95,13 +95,13 @@ protected IbmWatsonxEmbeddingsRequestManager getEmbeddingsRequestManager(
     }
 
     /**
-     * Builds an error message for IBM Watsonx actions.
+     * Builds an error message for IBM watsonx actions.
      *
      * @param requestType The type of request (e.g. COMPLETION, EMBEDDING, RERANK).
      * @param inferenceId The ID of the inference entity.
      * @return A formatted error message.
      */
     public static String buildErrorMessage(TaskType requestType, String inferenceId) {
-        return format("Failed to send IBM Watsonx %s request from inference entity id [%s]", requestType.toString(), inferenceId);
+        return format("Failed to send IBM watsonx %s request from inference entity id [%s]", requestType.toString(), inferenceId);
     }
 }

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/action/IbmWatsonxActionVisitor.java

@@ -15,9 +15,9 @@
 import java.util.Map;
 
 /**
- * Interface for creating {@link ExecutableAction} instances for IBM Watsonx models.
+ * Interface for creating {@link ExecutableAction} instances for IBM watsonx models.
  * <p>
- * This interface is used to create {@link ExecutableAction} instances for different types of IBM Watsonx models, such as
+ * This interface is used to create {@link ExecutableAction} instances for different types of IBM watsonx models, such as
  * {@link IbmWatsonxEmbeddingsModel} and {@link IbmWatsonxRerankModel} and {@link IbmWatsonxChatCompletionModel}.
  */
 public interface IbmWatsonxActionVisitor {

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/completion/IbmWatsonxChatCompletionModel.java

@@ -16,7 +16,6 @@
 import org.elasticsearch.xpack.inference.external.action.ExecutableAction;
 import org.elasticsearch.xpack.inference.services.ConfigurationParseContext;
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.IbmWatsonxModel;
-import org.elasticsearch.xpack.inference.services.ibmwatsonx.IbmWatsonxRateLimitServiceSettings;
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.action.IbmWatsonxActionVisitor;
 import org.elasticsearch.xpack.inference.services.settings.DefaultSecretSettings;
 
@@ -75,7 +74,7 @@ public static IbmWatsonxChatCompletionModel of(IbmWatsonxChatCompletionModel mod
         var overriddenServiceSettings = new IbmWatsonxChatCompletionServiceSettings(
             originalModelServiceSettings.uri(),
             originalModelServiceSettings.apiVersion(),
-            originalModelServiceSettings.modelId(),
+            request.model(),
             originalModelServiceSettings.projectId(),
             originalModelServiceSettings.rateLimitSettings()
         );
@@ -104,11 +103,6 @@ public static IbmWatsonxChatCompletionModel of(IbmWatsonxChatCompletionModel mod
         );
     }
 
-    @Override
-    public IbmWatsonxRateLimitServiceSettings rateLimitServiceSettings() {
-        return super.rateLimitServiceSettings();
-    }
-
     @Override
     public IbmWatsonxChatCompletionServiceSettings getServiceSettings() {
         return (IbmWatsonxChatCompletionServiceSettings) super.getServiceSettings();
@@ -132,7 +126,7 @@ public URI uri() {
 
     /**
      * Accepts a visitor to create an executable action. The returned action will not return documents in the response.
-     * @param visitor          Interface for creating {@link ExecutableAction} instances for IBM Watsonx models.
+     * @param visitor          Interface for creating {@link ExecutableAction} instances for IBM watsonx models.
      * @return the completion action
      */
     public ExecutableAction accept(IbmWatsonxActionVisitor visitor, Map<String, Object> taskSettings) {

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/request/IbmWatsonxChatCompletionRequest.java

@@ -21,8 +21,6 @@
 import java.nio.charset.StandardCharsets;
 import java.util.Objects;
 
-import static org.elasticsearch.xpack.inference.external.request.RequestUtils.createAuthBearerHeader;
-
 public class IbmWatsonxChatCompletionRequest implements IbmWatsonxRequest {
     private final IbmWatsonxChatCompletionModel model;
     private final UnifiedChatInput chatInput;
@@ -42,7 +40,6 @@ public HttpRequest createHttpRequest() {
         httpPost.setEntity(byteEntity);
 
         httpPost.setHeader(HttpHeaders.CONTENT_TYPE, XContentType.JSON.mediaType());
-        httpPost.setHeader(createAuthBearerHeader(model.getSecretSettings().apiKey()));
 
         decorateWithAuth(httpPost);
 
@@ -60,13 +57,13 @@ public void decorateWithAuth(HttpPost httpPost) {
 
     @Override
     public Request truncate() {
-        // No truncation for IBM Watsonx chat completions
+        // No truncation for IBM watsonx chat completions
         return this;
     }
 
     @Override
     public boolean[] getTruncationInfo() {
-        // No truncation for IBM Watsonx chat completions
+        // No truncation for IBM watsonx chat completions
         return null;
     }
 

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/rerank/IbmWatsonxRerankModel.java

@@ -100,7 +100,7 @@ public URI uri() {
 
     /**
      * Accepts a visitor to create an executable action. The returned action will not return documents in the response.
-     * @param visitor          Interface for creating {@link ExecutableAction} instances for IBM Watsonx models.
+     * @param visitor          Interface for creating {@link ExecutableAction} instances for IBM watsonx models.
      * @param taskSettings     Settings in the request to override the model's defaults
      * @return the rerank action
      */

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/response/IbmWatsonxEmbeddingsResponseEntity.java

@@ -28,7 +28,7 @@
 
 public class IbmWatsonxEmbeddingsResponseEntity {
 
-    private static final String FAILED_TO_FIND_FIELD_TEMPLATE = "Failed to find required field [%s] in IBM Watsonx embeddings response";
+    private static final String FAILED_TO_FIND_FIELD_TEMPLATE = "Failed to find required field [%s] in IBM watsonx embeddings response";
 
     public static TextEmbeddingFloatResults fromResponse(Request request, HttpResult response) throws IOException {
         var parserConfig = XContentParserConfiguration.EMPTY.withDeprecationHandler(LoggingDeprecationHandler.INSTANCE);

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/ibmwatsonx/response/IbmWatsonxRankedResponseEntity.java

@@ -32,7 +32,7 @@ public class IbmWatsonxRankedResponseEntity {
     private static final Logger logger = LogManager.getLogger(IbmWatsonxRankedResponseEntity.class);
 
     /**
-     * Parses the IBM Watsonx ranked response.
+     * Parses the IBM watsonx ranked response.
      *
      * For a request like:
      *     "model": "rerank-english-v2.0",
@@ -71,7 +71,7 @@ public class IbmWatsonxRankedResponseEntity {
      *   ],
      *   }
      *
-     * @param response the http response from IBM Watsonx
+     * @param response the http response from IBM watsonx
      * @return the parsed response
      * @throws IOException if there is an error parsing the response
      */

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/mistral/action/MistralActionCreator.java

@@ -37,8 +37,8 @@
 public class MistralActionCreator implements MistralActionVisitor {
 
     public static final String COMPLETION_ERROR_PREFIX = "Mistral completions";
-    static final String USER_ROLE = "user";
-    static final ResponseHandler COMPLETION_HANDLER = new MistralCompletionResponseHandler(
+    public static final String USER_ROLE = "user";
+    public static final ResponseHandler COMPLETION_HANDLER = new MistralCompletionResponseHandler(
         "mistral completions",
         OpenAiChatCompletionResponseEntity::fromResponse
     );

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/voyageai/rerank/VoyageAIRerankModel.java

@@ -109,7 +109,7 @@ public DefaultSecretSettings getSecretSettings() {
 
     /**
      * Accepts a visitor to create an executable action. The returned action will not return documents in the response.
-     * @param visitor          Interface for creating {@link ExecutableAction} instances for IBM Voyage AI models.
+     * @param visitor          Interface for creating {@link ExecutableAction} instances for Voyage AI models.
      * @param taskSettings     Settings in the request to override the model's defaults
      * @return the rerank action
      */

x-pack/plugin/inference/src/test/java/org/elasticsearch/xpack/inference/external/unified/UnifiedChatCompletionRequestEntityTests.java

@@ -293,7 +293,6 @@ public void testSerializationWithNestedObjects() throws IOException {
         String randomStop = "stop" + random.nextInt(1000);
         float randomTemperature = (float) ((float) Math.round(0.5d + (double) random.nextFloat() * 0.5d * 100000d) / 100000d);
         float randomTopP = (float) ((float) Math.round(0.5d + (double) random.nextFloat() * 0.5d * 100000d) / 100000d);
-        int randomTimeLimit = random.nextInt(1000);
 
         UnifiedCompletionRequest.Message message = new UnifiedCompletionRequest.Message(
             new UnifiedCompletionRequest.ContentString(randomContent),