elasticsearch 9.1.3__py3-none-any.whl → 9.2.1__py3-none-any.whl

elasticsearch/_sync/client/inference.py

@@ -51,7 +51,9 @@ class InferenceClient(NamespacedClient):
 
         :param inference_id: The inference Id
         :param input: Inference input. Either a string or an array of strings.
-        :param task_settings: Optional task settings
+        :param task_settings: Task settings for the individual inference request. These
+            settings are specific to the <task_type> you specified and override the task
+            settings specified when initializing the service.
         :param timeout: Specifies the amount of time to wait for the inference request
             to complete.
         """
@@ -117,14 +119,16 @@ class InferenceClient(NamespacedClient):
         .. raw:: html
 
           <p>Delete an inference endpoint.</p>
+          <p>This API requires the manage_inference cluster privilege (the built-in <code>inference_admin</code> role grants this privilege).</p>
 
 
         `<https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-inference-delete>`_
 
         :param inference_id: The inference identifier.
         :param task_type: The task type
-        :param dry_run: When true, the endpoint is not deleted and a list of ingest processors
-            which reference this endpoint is returned.
+        :param dry_run: When true, checks the semantic_text fields and inference processors
+            that reference the endpoint and returns them in a list, but does not delete
+            the endpoint.
         :param force: When true, the inference endpoint is forcefully deleted even if
             it is still being used by ingest processors or semantic text fields.
         """
@@ -191,6 +195,7 @@ class InferenceClient(NamespacedClient):
         .. raw:: html
 
           <p>Get an inference endpoint.</p>
+          <p>This API requires the <code>monitor_inference</code> cluster privilege (the built-in <code>inference_admin</code> and <code>inference_user</code> roles grant this privilege).</p>
 
 
         `<https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-inference-get>`_
@@ -381,11 +386,12 @@ class InferenceClient(NamespacedClient):
           However, if you do not plan to use the inference APIs to use these models or if you want to use non-NLP models, use the machine learning trained model APIs.</p>
           <p>The following integrations are available through the inference API. You can find the available task types next to the integration name:</p>
           <ul>
+          <li>AI21 (<code>chat_completion</code>, <code>completion</code>)</li>
           <li>AlibabaCloud AI Search (<code>completion</code>, <code>rerank</code>, <code>sparse_embedding</code>, <code>text_embedding</code>)</li>
           <li>Amazon Bedrock (<code>completion</code>, <code>text_embedding</code>)</li>
           <li>Amazon SageMaker (<code>chat_completion</code>, <code>completion</code>, <code>rerank</code>, <code>sparse_embedding</code>, <code>text_embedding</code>)</li>
           <li>Anthropic (<code>completion</code>)</li>
-          <li>Azure AI Studio (<code>completion</code>, <code>text_embedding</code>)</li>
+          <li>Azure AI Studio (<code>completion</code>, 'rerank', <code>text_embedding</code>)</li>
           <li>Azure OpenAI (<code>completion</code>, <code>text_embedding</code>)</li>
           <li>Cohere (<code>completion</code>, <code>rerank</code>, <code>text_embedding</code>)</li>
           <li>DeepSeek (<code>chat_completion</code>, <code>completion</code>)</li>
@@ -455,6 +461,82 @@ class InferenceClient(NamespacedClient):
             path_parts=__path_parts,
         )
 
+    @_rewrite_parameters(
+        body_fields=("service", "service_settings"),
+    )
+    def put_ai21(
+        self,
+        *,
+        task_type: t.Union[str, t.Literal["chat_completion", "completion"]],
+        ai21_inference_id: str,
+        service: t.Optional[t.Union[str, t.Literal["ai21"]]] = None,
+        service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
+        error_trace: t.Optional[bool] = None,
+        filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
+        human: t.Optional[bool] = None,
+        pretty: t.Optional[bool] = None,
+        timeout: t.Optional[t.Union[str, t.Literal[-1], t.Literal[0]]] = None,
+        body: t.Optional[t.Dict[str, t.Any]] = None,
+    ) -> ObjectApiResponse[t.Any]:
+        """
+        .. raw:: html
+
+          <p>Create a AI21 inference endpoint.</p>
+          <p>Create an inference endpoint to perform an inference task with the <code>ai21</code> service.</p>
+
+
+        `<https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-inference-put-ai21>`_
+
+        :param task_type: The type of the inference task that the model will perform.
+        :param ai21_inference_id: The unique identifier of the inference endpoint.
+        :param service: The type of service supported for the specified task type. In
+            this case, `ai21`.
+        :param service_settings: Settings used to install the inference model. These
+            settings are specific to the `ai21` service.
+        :param timeout: Specifies the amount of time to wait for the inference endpoint
+            to be created.
+        """
+        if task_type in SKIP_IN_PATH:
+            raise ValueError("Empty value passed for parameter 'task_type'")
+        if ai21_inference_id in SKIP_IN_PATH:
+            raise ValueError("Empty value passed for parameter 'ai21_inference_id'")
+        if service is None and body is None:
+            raise ValueError("Empty value passed for parameter 'service'")
+        if service_settings is None and body is None:
+            raise ValueError("Empty value passed for parameter 'service_settings'")
+        __path_parts: t.Dict[str, str] = {
+            "task_type": _quote(task_type),
+            "ai21_inference_id": _quote(ai21_inference_id),
+        }
+        __path = f'/_inference/{__path_parts["task_type"]}/{__path_parts["ai21_inference_id"]}'
+        __query: t.Dict[str, t.Any] = {}
+        __body: t.Dict[str, t.Any] = body if body is not None else {}
+        if error_trace is not None:
+            __query["error_trace"] = error_trace
+        if filter_path is not None:
+            __query["filter_path"] = filter_path
+        if human is not None:
+            __query["human"] = human
+        if pretty is not None:
+            __query["pretty"] = pretty
+        if timeout is not None:
+            __query["timeout"] = timeout
+        if not __body:
+            if service is not None:
+                __body["service"] = service
+            if service_settings is not None:
+                __body["service_settings"] = service_settings
+        __headers = {"accept": "application/json", "content-type": "application/json"}
+        return self.perform_request(  # type: ignore[return-value]
+            "PUT",
+            __path,
+            params=__query,
+            headers=__headers,
+            body=__body,
+            endpoint_id="inference.put_ai21",
+            path_parts=__path_parts,
+        )
+
     @_rewrite_parameters(
         body_fields=(
             "service",
@@ -467,7 +549,7 @@ class InferenceClient(NamespacedClient):
         self,
         *,
         task_type: t.Union[
-            str, t.Literal["completion", "rerank", "space_embedding", "text_embedding"]
+            str, t.Literal["completion", "rerank", "sparse_embedding", "text_embedding"]
         ],
         alibabacloud_inference_id: str,
         service: t.Optional[t.Union[str, t.Literal["alibabacloud-ai-search"]]] = None,
@@ -496,7 +578,9 @@ class InferenceClient(NamespacedClient):
             this case, `alibabacloud-ai-search`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `alibabacloud-ai-search` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `sparse_embedding` or `text_embedding` task types. Not applicable to
+            the `rerank` or `completion` task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -592,7 +676,8 @@ class InferenceClient(NamespacedClient):
             this case, `amazonbedrock`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `amazonbedrock` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` task type.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -694,7 +779,9 @@ class InferenceClient(NamespacedClient):
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `amazon_sagemaker` service and `service_settings.api`
             you specified.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `sparse_embedding` or `text_embedding` task types. Not applicable to
+            the `rerank`, `completion`, or `chat_completion` task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type and `service_settings.api` you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -748,12 +835,7 @@ class InferenceClient(NamespacedClient):
         )
 
     @_rewrite_parameters(
-        body_fields=(
-            "service",
-            "service_settings",
-            "chunking_settings",
-            "task_settings",
-        ),
+        body_fields=("service", "service_settings", "task_settings"),
     )
     def put_anthropic(
         self,
@@ -762,7 +844,6 @@ class InferenceClient(NamespacedClient):
         anthropic_inference_id: str,
         service: t.Optional[t.Union[str, t.Literal["anthropic"]]] = None,
         service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
-        chunking_settings: t.Optional[t.Mapping[str, t.Any]] = None,
         error_trace: t.Optional[bool] = None,
         filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
         human: t.Optional[bool] = None,
@@ -786,8 +867,7 @@ class InferenceClient(NamespacedClient):
         :param service: The type of service supported for the specified task type. In
             this case, `anthropic`.
         :param service_settings: Settings used to install the inference model. These
-            settings are specific to the `watsonxai` service.
-        :param chunking_settings: The chunking configuration object.
+            settings are specific to the `anthropic` service.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -825,8 +905,6 @@ class InferenceClient(NamespacedClient):
                 __body["service"] = service
             if service_settings is not None:
                 __body["service_settings"] = service_settings
-            if chunking_settings is not None:
-                __body["chunking_settings"] = chunking_settings
             if task_settings is not None:
                 __body["task_settings"] = task_settings
         __headers = {"accept": "application/json", "content-type": "application/json"}
@@ -851,7 +929,7 @@ class InferenceClient(NamespacedClient):
     def put_azureaistudio(
         self,
         *,
-        task_type: t.Union[str, t.Literal["completion", "text_embedding"]],
+        task_type: t.Union[str, t.Literal["completion", "rerank", "text_embedding"]],
         azureaistudio_inference_id: str,
         service: t.Optional[t.Union[str, t.Literal["azureaistudio"]]] = None,
         service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
@@ -878,8 +956,10 @@ class InferenceClient(NamespacedClient):
         :param service: The type of service supported for the specified task type. In
             this case, `azureaistudio`.
         :param service_settings: Settings used to install the inference model. These
-            settings are specific to the `openai` service.
-        :param chunking_settings: The chunking configuration object.
+            settings are specific to the `azureaistudio` service.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `rerank` or `completion`
+            task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -979,7 +1059,8 @@ class InferenceClient(NamespacedClient):
             this case, `azureopenai`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `azureopenai` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` task type.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -1071,7 +1152,9 @@ class InferenceClient(NamespacedClient):
             this case, `cohere`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `cohere` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `rerank` or `completion`
+            task type.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -1122,6 +1205,90 @@ class InferenceClient(NamespacedClient):
             path_parts=__path_parts,
         )
 
+    @_rewrite_parameters(
+        body_fields=("service", "service_settings", "task_settings"),
+    )
+    def put_contextualai(
+        self,
+        *,
+        task_type: t.Union[str, t.Literal["rerank"]],
+        contextualai_inference_id: str,
+        service: t.Optional[t.Union[str, t.Literal["contextualai"]]] = None,
+        service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
+        error_trace: t.Optional[bool] = None,
+        filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
+        human: t.Optional[bool] = None,
+        pretty: t.Optional[bool] = None,
+        task_settings: t.Optional[t.Mapping[str, t.Any]] = None,
+        timeout: t.Optional[t.Union[str, t.Literal[-1], t.Literal[0]]] = None,
+        body: t.Optional[t.Dict[str, t.Any]] = None,
+    ) -> ObjectApiResponse[t.Any]:
+        """
+        .. raw:: html
+
+          <p>Create an Contextual AI inference endpoint.</p>
+          <p>Create an inference endpoint to perform an inference task with the <code>contexualai</code> service.</p>
+          <p>To review the available <code>rerank</code> models, refer to <a href="https://docs.contextual.ai/api-reference/rerank/rerank#body-model">https://docs.contextual.ai/api-reference/rerank/rerank#body-model</a>.</p>
+
+
+        `<https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-inference-put-contextualai>`_
+
+        :param task_type: The type of the inference task that the model will perform.
+        :param contextualai_inference_id: The unique identifier of the inference endpoint.
+        :param service: The type of service supported for the specified task type. In
+            this case, `contextualai`.
+        :param service_settings: Settings used to install the inference model. These
+            settings are specific to the `contextualai` service.
+        :param task_settings: Settings to configure the inference task. These settings
+            are specific to the task type you specified.
+        :param timeout: Specifies the amount of time to wait for the inference endpoint
+            to be created.
+        """
+        if task_type in SKIP_IN_PATH:
+            raise ValueError("Empty value passed for parameter 'task_type'")
+        if contextualai_inference_id in SKIP_IN_PATH:
+            raise ValueError(
+                "Empty value passed for parameter 'contextualai_inference_id'"
+            )
+        if service is None and body is None:
+            raise ValueError("Empty value passed for parameter 'service'")
+        if service_settings is None and body is None:
+            raise ValueError("Empty value passed for parameter 'service_settings'")
+        __path_parts: t.Dict[str, str] = {
+            "task_type": _quote(task_type),
+            "contextualai_inference_id": _quote(contextualai_inference_id),
+        }
+        __path = f'/_inference/{__path_parts["task_type"]}/{__path_parts["contextualai_inference_id"]}'
+        __query: t.Dict[str, t.Any] = {}
+        __body: t.Dict[str, t.Any] = body if body is not None else {}
+        if error_trace is not None:
+            __query["error_trace"] = error_trace
+        if filter_path is not None:
+            __query["filter_path"] = filter_path
+        if human is not None:
+            __query["human"] = human
+        if pretty is not None:
+            __query["pretty"] = pretty
+        if timeout is not None:
+            __query["timeout"] = timeout
+        if not __body:
+            if service is not None:
+                __body["service"] = service
+            if service_settings is not None:
+                __body["service_settings"] = service_settings
+            if task_settings is not None:
+                __body["task_settings"] = task_settings
+        __headers = {"accept": "application/json", "content-type": "application/json"}
+        return self.perform_request(  # type: ignore[return-value]
+            "PUT",
+            __path,
+            params=__query,
+            headers=__headers,
+            body=__body,
+            endpoint_id="inference.put_contextualai",
+            path_parts=__path_parts,
+        )
+
     @_rewrite_parameters(
         body_fields=(
             "service",
@@ -1202,7 +1369,9 @@ class InferenceClient(NamespacedClient):
             this case, `custom`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `custom` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `sparse_embedding` or `text_embedding` task types. Not applicable to
+            the `rerank` or `completion` task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         """
@@ -1250,7 +1419,7 @@ class InferenceClient(NamespacedClient):
         )
 
     @_rewrite_parameters(
-        body_fields=("service", "service_settings", "chunking_settings"),
+        body_fields=("service", "service_settings"),
     )
     def put_deepseek(
         self,
@@ -1259,7 +1428,6 @@ class InferenceClient(NamespacedClient):
         deepseek_inference_id: str,
         service: t.Optional[t.Union[str, t.Literal["deepseek"]]] = None,
         service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
-        chunking_settings: t.Optional[t.Mapping[str, t.Any]] = None,
         error_trace: t.Optional[bool] = None,
         filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
         human: t.Optional[bool] = None,
@@ -1282,7 +1450,6 @@ class InferenceClient(NamespacedClient):
             this case, `deepseek`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `deepseek` service.
-        :param chunking_settings: The chunking configuration object.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
             to be created.
         """
@@ -1316,8 +1483,6 @@ class InferenceClient(NamespacedClient):
                 __body["service"] = service
             if service_settings is not None:
                 __body["service_settings"] = service_settings
-            if chunking_settings is not None:
-                __body["chunking_settings"] = chunking_settings
         __headers = {"accept": "application/json", "content-type": "application/json"}
         return self.perform_request(  # type: ignore[return-value]
             "PUT",
@@ -1384,7 +1549,9 @@ class InferenceClient(NamespacedClient):
             this case, `elasticsearch`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `elasticsearch` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `sparse_embedding` and `text_embedding` task types. Not applicable to
+            the `rerank` task type.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -1484,7 +1651,8 @@ class InferenceClient(NamespacedClient):
             this case, `elser`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `elser` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Note that for ELSER
+            endpoints, the max_chunk_size may not exceed `300`.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
             to be created.
         """
@@ -1564,7 +1732,8 @@ class InferenceClient(NamespacedClient):
             this case, `googleaistudio`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `googleaistudio` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` task type.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
             to be created.
         """
@@ -1654,7 +1823,9 @@ class InferenceClient(NamespacedClient):
             this case, `googlevertexai`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `googlevertexai` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `rerank`, `completion`,
+            or `chat_completion` task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -1782,7 +1953,9 @@ class InferenceClient(NamespacedClient):
             this case, `hugging_face`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `hugging_face` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `rerank`, `completion`,
+            or `chat_completion` task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -1876,7 +2049,8 @@ class InferenceClient(NamespacedClient):
             this case, `jinaai`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `jinaai` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `rerank` task type.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -1927,6 +2101,90 @@ class InferenceClient(NamespacedClient):
             path_parts=__path_parts,
         )
 
+    @_rewrite_parameters(
+        body_fields=("service", "service_settings", "chunking_settings"),
+    )
+    def put_llama(
+        self,
+        *,
+        task_type: t.Union[
+            str, t.Literal["chat_completion", "completion", "text_embedding"]
+        ],
+        llama_inference_id: str,
+        service: t.Optional[t.Union[str, t.Literal["llama"]]] = None,
+        service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
+        chunking_settings: t.Optional[t.Mapping[str, t.Any]] = None,
+        error_trace: t.Optional[bool] = None,
+        filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
+        human: t.Optional[bool] = None,
+        pretty: t.Optional[bool] = None,
+        timeout: t.Optional[t.Union[str, t.Literal[-1], t.Literal[0]]] = None,
+        body: t.Optional[t.Dict[str, t.Any]] = None,
+    ) -> ObjectApiResponse[t.Any]:
+        """
+        .. raw:: html
+
+          <p>Create a Llama inference endpoint.</p>
+          <p>Create an inference endpoint to perform an inference task with the <code>llama</code> service.</p>
+
+
+        `<https://www.elastic.co/docs/api/doc/elasticsearch/operation/operation-inference-put-llama>`_
+
+        :param task_type: The type of the inference task that the model will perform.
+        :param llama_inference_id: The unique identifier of the inference endpoint.
+        :param service: The type of service supported for the specified task type. In
+            this case, `llama`.
+        :param service_settings: Settings used to install the inference model. These
+            settings are specific to the `llama` service.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` or `chat_completion`
+            task types.
+        :param timeout: Specifies the amount of time to wait for the inference endpoint
+            to be created.
+        """
+        if task_type in SKIP_IN_PATH:
+            raise ValueError("Empty value passed for parameter 'task_type'")
+        if llama_inference_id in SKIP_IN_PATH:
+            raise ValueError("Empty value passed for parameter 'llama_inference_id'")
+        if service is None and body is None:
+            raise ValueError("Empty value passed for parameter 'service'")
+        if service_settings is None and body is None:
+            raise ValueError("Empty value passed for parameter 'service_settings'")
+        __path_parts: t.Dict[str, str] = {
+            "task_type": _quote(task_type),
+            "llama_inference_id": _quote(llama_inference_id),
+        }
+        __path = f'/_inference/{__path_parts["task_type"]}/{__path_parts["llama_inference_id"]}'
+        __query: t.Dict[str, t.Any] = {}
+        __body: t.Dict[str, t.Any] = body if body is not None else {}
+        if error_trace is not None:
+            __query["error_trace"] = error_trace
+        if filter_path is not None:
+            __query["filter_path"] = filter_path
+        if human is not None:
+            __query["human"] = human
+        if pretty is not None:
+            __query["pretty"] = pretty
+        if timeout is not None:
+            __query["timeout"] = timeout
+        if not __body:
+            if service is not None:
+                __body["service"] = service
+            if service_settings is not None:
+                __body["service_settings"] = service_settings
+            if chunking_settings is not None:
+                __body["chunking_settings"] = chunking_settings
+        __headers = {"accept": "application/json", "content-type": "application/json"}
+        return self.perform_request(  # type: ignore[return-value]
+            "PUT",
+            __path,
+            params=__query,
+            headers=__headers,
+            body=__body,
+            endpoint_id="inference.put_llama",
+            path_parts=__path_parts,
+        )
+
     @_rewrite_parameters(
         body_fields=("service", "service_settings", "chunking_settings"),
     )
@@ -1962,7 +2220,9 @@ class InferenceClient(NamespacedClient):
             this case, `mistral`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `mistral` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` or `chat_completion`
+            task types.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
             to be created.
         """
@@ -2052,7 +2312,9 @@ class InferenceClient(NamespacedClient):
             this case, `openai`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `openai` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` or `chat_completion`
+            task types.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -2143,7 +2405,8 @@ class InferenceClient(NamespacedClient):
             this case, `voyageai`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `voyageai` service.
-        :param chunking_settings: The chunking configuration object.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `rerank` task type.
         :param task_settings: Settings to configure the inference task. These settings
             are specific to the task type you specified.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
@@ -2195,7 +2458,7 @@ class InferenceClient(NamespacedClient):
         )
 
     @_rewrite_parameters(
-        body_fields=("service", "service_settings"),
+        body_fields=("service", "service_settings", "chunking_settings"),
     )
     def put_watsonx(
         self,
@@ -2206,6 +2469,7 @@ class InferenceClient(NamespacedClient):
         watsonx_inference_id: str,
         service: t.Optional[t.Union[str, t.Literal["watsonxai"]]] = None,
         service_settings: t.Optional[t.Mapping[str, t.Any]] = None,
+        chunking_settings: t.Optional[t.Mapping[str, t.Any]] = None,
         error_trace: t.Optional[bool] = None,
         filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
         human: t.Optional[bool] = None,
@@ -2230,6 +2494,9 @@ class InferenceClient(NamespacedClient):
             this case, `watsonxai`.
         :param service_settings: Settings used to install the inference model. These
             settings are specific to the `watsonxai` service.
+        :param chunking_settings: The chunking configuration object. Applies only to
+            the `text_embedding` task type. Not applicable to the `completion` or `chat_completion`
+            task types.
         :param timeout: Specifies the amount of time to wait for the inference endpoint
             to be created.
         """
@@ -2263,6 +2530,8 @@ class InferenceClient(NamespacedClient):
                 __body["service"] = service
             if service_settings is not None:
                 __body["service_settings"] = service_settings
+            if chunking_settings is not None:
+                __body["chunking_settings"] = chunking_settings
         __headers = {"accept": "application/json", "content-type": "application/json"}
         return self.perform_request(  # type: ignore[return-value]
             "PUT",
@@ -2373,7 +2642,9 @@ class InferenceClient(NamespacedClient):
 
         :param inference_id: The inference Id
         :param input: Inference input. Either a string or an array of strings.
-        :param task_settings: Optional task settings
+        :param task_settings: Task settings for the individual inference request. These
+            settings are specific to the <task_type> you specified and override the task
+            settings specified when initializing the service.
         :param timeout: Specifies the amount of time to wait for the inference request
             to complete.
         """
@@ -2445,7 +2716,9 @@ class InferenceClient(NamespacedClient):
             to the relevant service-specific documentation for more info. > info > The
             `input_type` parameter specified on the root level of the request body will
             take precedence over the `input_type` parameter specified in `task_settings`.
-        :param task_settings: Optional task settings
+        :param task_settings: Task settings for the individual inference request. These
+            settings are specific to the <task_type> you specified and override the task
+            settings specified when initializing the service.
         :param timeout: Specifies the amount of time to wait for the inference request
             to complete.
         """

elasticsearch/_sync/client/ingest.py

@@ -580,6 +580,7 @@ class IngestClient(NamespacedClient):
         body_fields=(
             "deprecated",
             "description",
+            "field_access_pattern",
             "meta",
             "on_failure",
             "processors",
@@ -594,6 +595,9 @@ class IngestClient(NamespacedClient):
         deprecated: t.Optional[bool] = None,
         description: t.Optional[str] = None,
         error_trace: t.Optional[bool] = None,
+        field_access_pattern: t.Optional[
+            t.Union[str, t.Literal["classic", "flexible"]]
+        ] = None,
         filter_path: t.Optional[t.Union[str, t.Sequence[str]]] = None,
         human: t.Optional[bool] = None,
         if_version: t.Optional[int] = None,
@@ -621,6 +625,8 @@ class IngestClient(NamespacedClient):
             or updating a non-deprecated index template, Elasticsearch will emit a deprecation
             warning.
         :param description: Description of the ingest pipeline.
+        :param field_access_pattern: Controls how processors in this pipeline should
+            read and write data on a document's source.
         :param if_version: Required version for optimistic concurrency control for pipeline
             updates
         :param master_timeout: Period to wait for a connection to the master node. If
@@ -667,6 +673,8 @@ class IngestClient(NamespacedClient):
                 __body["deprecated"] = deprecated
             if description is not None:
                 __body["description"] = description
+            if field_access_pattern is not None:
+                __body["field_access_pattern"] = field_access_pattern
             if meta is not None:
                 __body["_meta"] = meta
             if on_failure is not None: