Class OrchestrationClient

Client for the SAP AI Core Orchestration Service.

Remarks

The configuration (model, parameters, tools, etc.) is fixed at construction time. To use different model parameters or tools for a specific request, create a new OrchestrationClient instance with the desired configuration.

Index

Constructors

constructor

Methods

chatCompletion stream

Constructors

constructor

new OrchestrationClient(
    config: OrchestrationModuleConfig,
    deploymentConfig?: DeploymentIdConfig | ResourceGroupConfig,
    destination?: HttpDestinationOrFetchOptions,
): OrchestrationClient
Creates an instance of the orchestration client with an inline module configuration.
Parameters
- config: OrchestrationModuleConfig
  A single orchestration module configuration.
- OptionaldeploymentConfig: DeploymentIdConfig | ResourceGroupConfig
  Deployment configuration.
- Optionaldestination: HttpDestinationOrFetchOptions
  The destination to use for the request.
Returns OrchestrationClient
- Defined in packages/orchestration/dist/orchestration-client.d.ts:26
new OrchestrationClient(
    configRef: OrchestrationConfigRefById,
    deploymentConfig?: DeploymentIdConfig | ResourceGroupConfig,
    destination?: HttpDestinationOrFetchOptions,
): OrchestrationClient
Creates an instance of the orchestration client referencing a stored configuration by its unique ID.
Parameters
- configRef: OrchestrationConfigRefById
  An object with the configuration id.
- OptionaldeploymentConfig: DeploymentIdConfig | ResourceGroupConfig
  Deployment configuration.
- Optionaldestination: HttpDestinationOrFetchOptions
  The destination to use for the request.
Returns OrchestrationClient
- Defined in packages/orchestration/dist/orchestration-client.d.ts:33
new OrchestrationClient(
    configRef: OrchestrationConfigRefByName,
    deploymentConfig?: DeploymentIdConfig | ResourceGroupConfig,
    destination?: HttpDestinationOrFetchOptions,
): OrchestrationClient
Creates an instance of the orchestration client referencing a stored configuration by scenario, name and version.
Parameters
- configRef: OrchestrationConfigRefByName
  An object with scenario, name, and version.
- OptionaldeploymentConfig: DeploymentIdConfig | ResourceGroupConfig
  Deployment configuration.
- Optionaldestination: HttpDestinationOrFetchOptions
  The destination to use for the request.
Returns OrchestrationClient
- Defined in packages/orchestration/dist/orchestration-client.d.ts:40
new OrchestrationClient(
    config: string,
    deploymentConfig?: DeploymentIdConfig | ResourceGroupConfig,
    destination?: HttpDestinationOrFetchOptions,
): OrchestrationClient
Creates an instance of the orchestration client with a JSON string configuration.
Parameters
- config: string
  A JSON string obtained from AI Launchpad.
- OptionaldeploymentConfig: DeploymentIdConfig | ResourceGroupConfig
  Deployment configuration.
- Optionaldestination: HttpDestinationOrFetchOptions
  The destination to use for the request.
Returns OrchestrationClient
- Defined in packages/orchestration/dist/orchestration-client.d.ts:47
new OrchestrationClient(
    configs: OrchestrationModuleConfigList,
    deploymentConfig?: DeploymentIdConfig | ResourceGroupConfig,
    destination?: HttpDestinationOrFetchOptions,
): OrchestrationClient
Creates an instance of the orchestration client with an inline module configuration.
Parameters
- configs: OrchestrationModuleConfigList
- OptionaldeploymentConfig: DeploymentIdConfig | ResourceGroupConfig
  Deployment configuration.
- Optionaldestination: HttpDestinationOrFetchOptions
  The destination to use for the request.
Returns OrchestrationClient
- Defined in packages/orchestration/dist/orchestration-client.d.ts:54
new OrchestrationClient(
    config:
        | string
        | OrchestrationModuleConfigList
        | Xor<
            OrchestrationConfigRefById,
            Xor<OrchestrationConfigRefByName, OrchestrationModuleConfig>,
        >,
    deploymentConfig?: DeploymentIdConfig | ResourceGroupConfig,
    destination?: HttpDestinationOrFetchOptions,
): OrchestrationClient
Parameters
- config:
      | string
      | OrchestrationModuleConfigList
      | Xor<
          OrchestrationConfigRefById,
          Xor<OrchestrationConfigRefByName, OrchestrationModuleConfig>,
      >
- OptionaldeploymentConfig: DeploymentIdConfig | ResourceGroupConfig
- Optionaldestination: HttpDestinationOrFetchOptions
Returns OrchestrationClient
- Defined in packages/orchestration/dist/orchestration-client.d.ts:55

Methods

chatCompletion

chatCompletion(
    request?: ChatCompletionRequest,
    requestConfig?: Pick<
        HttpRequestConfig,
        | "middleware"
        | "maxContentLength"
        | "proxy"
        | "httpAgent"
        | "httpsAgent"
        | "parameterEncoder"
        | "signal",
    > & Record<string, any> & { headers?: OrchestrationRequestHeaders },
): Promise<OrchestrationResponse>
Send a chat completion request to the orchestration service.
Parameters
- Optionalrequest: ChatCompletionRequest
  Request containing messages, placeholder values, and message history.
- OptionalrequestConfig: Pick<
      HttpRequestConfig,
      | "middleware"
      | "maxContentLength"
      | "proxy"
      | "httpAgent"
      | "httpsAgent"
      | "parameterEncoder"
      | "signal",
  > & Record<string, any> & { headers?: OrchestrationRequestHeaders }
  Additional request configuration. Use requestConfig.headers to pass service-specific headers:
  
  AI-Object-Store-Secret-Name: Name of the object store secret used by the feedback service.
Returns Promise<OrchestrationResponse>
The orchestration service response.
- Defined in packages/orchestration/dist/orchestration-client.d.ts:63

stream

stream(
    request?: ChatCompletionRequest,
    signal?: AbortSignal,
    options?: StreamOptions,
    requestConfig?: Pick<
        HttpRequestConfig,
        | "middleware"
        | "maxContentLength"
        | "proxy"
        | "httpAgent"
        | "httpsAgent"
        | "parameterEncoder"
        | "signal",
    > & Record<string, any> & { headers?: OrchestrationRequestHeaders },
): Promise<OrchestrationStreamResponse<OrchestrationStreamChunkResponse>>
Create a streaming chat completion request to the orchestration service.
Parameters
- Optionalrequest: ChatCompletionRequest
  Request containing messages, placeholder values, and message history.
- Optionalsignal: AbortSignal
  An abort signal to cancel the request.
- Optionaloptions: StreamOptions
  Streaming options, e.g., for input/output filtering.
- OptionalrequestConfig: Pick<
      HttpRequestConfig,
      | "middleware"
      | "maxContentLength"
      | "proxy"
      | "httpAgent"
      | "httpsAgent"
      | "parameterEncoder"
      | "signal",
  > & Record<string, any> & { headers?: OrchestrationRequestHeaders }
  Additional request configuration. Use requestConfig.headers to pass service-specific headers:
  
  AI-Object-Store-Secret-Name: Name of the object store secret used by the feedback service.
Returns Promise<OrchestrationStreamResponse<OrchestrationStreamChunkResponse>>
The orchestration stream response.
- Defined in packages/orchestration/dist/orchestration-client.d.ts:75

Class OrchestrationClient

Remarks

Index

Constructors

Methods

Constructors

constructor

Parameters

Returns OrchestrationClient

Parameters

Returns OrchestrationClient

Parameters

Returns OrchestrationClient

Parameters

Returns OrchestrationClient

Parameters

Returns OrchestrationClient

Parameters

Returns OrchestrationClient

Methods

chatCompletion

Parameters

Returns Promise<OrchestrationResponse>

stream

Parameters

Returns Promise<OrchestrationStreamResponse<OrchestrationStreamChunkResponse>>

Settings

On This Page