Add Ibm Granite Completion and Chat Completion support (elastic#129146)

* Add Ibm Granite Completion and Chat Completion support

* Apply suggestions

* remove ibm watsonx transport version constant

* update transport version

Author: Evgenii-Kazannik

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

@@ -328,7 +328,7 @@ static TransportVersion def(int id) {
     public static final TransportVersion MAPPINGS_IN_DATA_STREAMS = def(9_112_0_00);
     public static final TransportVersion PROJECT_STATE_REGISTRY_RECORDS_DELETIONS = def(9_113_0_00);
     public static final TransportVersion ESQL_SERIALIZE_TIMESERIES_FIELD_TYPE = def(9_114_0_00);
-
+    public static final TransportVersion ML_INFERENCE_IBM_WATSONX_COMPLETION_ADDED = def(9_115_0_00);
     /*
      * STOP! READ THIS FIRST! No, really,
      *        ____ _____ ___  ____  _        ____  _____    _    ____    _____ _   _ ___ ____    _____ ___ ____  ____ _____ _

x-pack/plugin/inference/qa/inference-service-tests/src/javaRestTest/java/org/elasticsearch/xpack/inference/InferenceGetServicesIT.java

@@ -151,7 +151,8 @@ public void testGetServicesWithCompletionTaskType() throws IOException {
                     "completion_test_service",
                     "hugging_face",
                     "amazon_sagemaker",
-                    "mistral"
+                    "mistral",
+                    "watsonxai"
                 ).toArray()
             )
         );
@@ -169,7 +170,8 @@ public void testGetServicesWithChatCompletionTaskType() throws IOException {
                     "hugging_face",
                     "amazon_sagemaker",
                     "googlevertexai",
-                    "mistral"
+                    "mistral",
+                    "watsonxai"
                 ).toArray()
             )
         );

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

@@ -95,6 +95,7 @@
 import org.elasticsearch.xpack.inference.services.huggingface.elser.HuggingFaceElserServiceSettings;
 import org.elasticsearch.xpack.inference.services.huggingface.rerank.HuggingFaceRerankServiceSettings;
 import org.elasticsearch.xpack.inference.services.huggingface.rerank.HuggingFaceRerankTaskSettings;
+import org.elasticsearch.xpack.inference.services.ibmwatsonx.completion.IbmWatsonxChatCompletionServiceSettings;
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.embeddings.IbmWatsonxEmbeddingsServiceSettings;
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.rerank.IbmWatsonxRerankServiceSettings;
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.rerank.IbmWatsonxRerankTaskSettings;
@@ -469,6 +470,13 @@ private static void addIbmWatsonxNamedWritables(List<NamedWriteableRegistry.Entr
         namedWriteables.add(
             new NamedWriteableRegistry.Entry(TaskSettings.class, IbmWatsonxRerankTaskSettings.NAME, IbmWatsonxRerankTaskSettings::new)
         );
+        namedWriteables.add(
+            new NamedWriteableRegistry.Entry(
+                ServiceSettings.class,
+                IbmWatsonxChatCompletionServiceSettings.NAME,
+                IbmWatsonxChatCompletionServiceSettings::new
+            )
+        );
     }
 
     private static void addGoogleVertexAiNamedWriteables(List<NamedWriteableRegistry.Entry> namedWriteables) {

x-pack/plugin/inference/src/main/java/org/elasticsearch/xpack/inference/services/cohere/rerank/CohereRerankModel.java

@@ -80,8 +80,8 @@ public DefaultSecretSettings getSecretSettings() {
 
     /**
      * Accepts a visitor to create an executable action. The returned action will not return documents in the response.
-     * @param visitor _
-     * @param taskSettings _
+     * @param visitor          Interface for creating {@link ExecutableAction} instances for Cohere models.
+     * @param taskSettings     Settings in the request to override the model's defaults
      * @return the rerank action
      */
     @Override

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

@@ -0,0 +1,51 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+package org.elasticsearch.xpack.inference.services.ibmwatsonx;
+
+import org.elasticsearch.xpack.core.inference.results.UnifiedChatCompletionException;
+import org.elasticsearch.xpack.inference.external.http.HttpResult;
+import org.elasticsearch.xpack.inference.external.http.retry.ErrorResponse;
+import org.elasticsearch.xpack.inference.external.http.retry.ResponseParser;
+import org.elasticsearch.xpack.inference.external.request.Request;
+import org.elasticsearch.xpack.inference.services.ibmwatsonx.response.IbmWatsonxErrorResponseEntity;
+import org.elasticsearch.xpack.inference.services.openai.OpenAiUnifiedChatCompletionResponseHandler;
+
+import java.util.Locale;
+
+/**
+ * Handles streaming chat completion responses and error parsing for Watsonx inference endpoints.
+ * Adapts the OpenAI handler to support Watsonx's error schema.
+ */
+public class IbmWatsonUnifiedChatCompletionResponseHandler extends OpenAiUnifiedChatCompletionResponseHandler {
+
+    private static final String WATSONX_ERROR = "watsonx_error";
+
+    public IbmWatsonUnifiedChatCompletionResponseHandler(String requestType, ResponseParser parseFunction) {
+        super(requestType, parseFunction, IbmWatsonxErrorResponseEntity::fromResponse);
+    }
+
+    @Override
+    protected Exception buildError(String message, Request request, HttpResult result, ErrorResponse errorResponse) {
+        assert request.isStreaming() : "Only streaming requests support this format";
+        var responseStatusCode = result.response().getStatusLine().getStatusCode();
+        if (request.isStreaming()) {
+            var errorMessage = errorMessage(message, request, result, errorResponse, responseStatusCode);
+            var restStatus = toRestStatus(responseStatusCode);
+            return errorResponse instanceof IbmWatsonxErrorResponseEntity
+                ? new UnifiedChatCompletionException(restStatus, errorMessage, WATSONX_ERROR, restStatus.name().toLowerCase(Locale.ROOT))
+                : new UnifiedChatCompletionException(
+                    restStatus,
+                    errorMessage,
+                    createErrorType(errorResponse),
+                    restStatus.name().toLowerCase(Locale.ROOT)
+                );
+        } else {
+            return super.buildError(message, request, result, errorResponse);
+        }
+    }
+}

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

@@ -0,0 +1,25 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+package org.elasticsearch.xpack.inference.services.ibmwatsonx;
+
+import org.elasticsearch.xpack.inference.external.http.retry.ResponseParser;
+import org.elasticsearch.xpack.inference.services.ibmwatsonx.response.IbmWatsonxErrorResponseEntity;
+import org.elasticsearch.xpack.inference.services.openai.OpenAiChatCompletionResponseHandler;
+
+public class IbmWatsonxCompletionResponseHandler extends OpenAiChatCompletionResponseHandler {
+
+    /**
+     * 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 parseFunction The function to parse the response.
+     */
+    public IbmWatsonxCompletionResponseHandler(String requestType, ResponseParser parseFunction) {
+        super(requestType, parseFunction, IbmWatsonxErrorResponseEntity::fromResponse);
+    }
+}

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

@@ -7,18 +7,19 @@
 
 package org.elasticsearch.xpack.inference.services.ibmwatsonx;
 
-import org.elasticsearch.inference.Model;
 import org.elasticsearch.inference.ModelConfigurations;
 import org.elasticsearch.inference.ModelSecrets;
 import org.elasticsearch.inference.ServiceSettings;
 import org.elasticsearch.inference.TaskSettings;
 import org.elasticsearch.xpack.inference.external.action.ExecutableAction;
+import org.elasticsearch.xpack.inference.services.RateLimitGroupingModel;
 import org.elasticsearch.xpack.inference.services.ibmwatsonx.action.IbmWatsonxActionVisitor;
+import org.elasticsearch.xpack.inference.services.settings.RateLimitSettings;
 
 import java.util.Map;
 import java.util.Objects;
 
-public abstract class IbmWatsonxModel extends Model {
+public abstract class IbmWatsonxModel extends RateLimitGroupingModel {
 
     private final IbmWatsonxRateLimitServiceSettings rateLimitServiceSettings;
 
@@ -49,4 +50,14 @@ public IbmWatsonxModel(IbmWatsonxModel model, TaskSettings taskSettings) {
     public IbmWatsonxRateLimitServiceSettings rateLimitServiceSettings() {
         return rateLimitServiceSettings;
     }
+
+    @Override
+    public int rateLimitGroupingHash() {
+        return Objects.hash(this.rateLimitServiceSettings);
+    }
+
+    @Override
+    public RateLimitSettings rateLimitSettings() {
+        return this.rateLimitServiceSettings().rateLimitSettings();
+    }
 }

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)
         );
     }