llama-stack-client 0.2.7 → 0.2.8-rc2

package/src/resources/chat/index.ts

@@ -4,7 +4,10 @@ export { Chat, type ChatCompletionChunk } from './chat';
 export {
   Completions,
   type CompletionCreateResponse,
+  type CompletionRetrieveResponse,
+  type CompletionListResponse,
   type CompletionCreateParams,
   type CompletionCreateParamsNonStreaming,
   type CompletionCreateParamsStreaming,
+  type CompletionListParams,
 } from './completions';

package/src/resources/completions.ts

@@ -155,86 +155,86 @@ export interface CompletionCreateParamsBase {
   model: string;
 
   /**
-   * The prompt to generate a completion for
+   * The prompt to generate a completion for.
    */
   prompt: string | Array<string> | Array<number> | Array<Array<number>>;
 
   /**
-   * (Optional) The number of completions to generate
+   * (Optional) The number of completions to generate.
    */
   best_of?: number;
 
   /**
-   * (Optional) Whether to echo the prompt
+   * (Optional) Whether to echo the prompt.
    */
   echo?: boolean;
 
   /**
-   * (Optional) The penalty for repeated tokens
+   * (Optional) The penalty for repeated tokens.
    */
   frequency_penalty?: number;
 
   guided_choice?: Array<string>;
 
   /**
-   * (Optional) The logit bias to use
+   * (Optional) The logit bias to use.
    */
   logit_bias?: Record<string, number>;
 
   /**
-   * (Optional) The log probabilities to use
+   * (Optional) The log probabilities to use.
    */
   logprobs?: boolean;
 
   /**
-   * (Optional) The maximum number of tokens to generate
+   * (Optional) The maximum number of tokens to generate.
    */
   max_tokens?: number;
 
   /**
-   * (Optional) The number of completions to generate
+   * (Optional) The number of completions to generate.
    */
   n?: number;
 
   /**
-   * (Optional) The penalty for repeated tokens
+   * (Optional) The penalty for repeated tokens.
    */
   presence_penalty?: number;
 
   prompt_logprobs?: number;
 
   /**
-   * (Optional) The seed to use
+   * (Optional) The seed to use.
    */
   seed?: number;
 
   /**
-   * (Optional) The stop tokens to use
+   * (Optional) The stop tokens to use.
    */
   stop?: string | Array<string>;
 
   /**
-   * (Optional) Whether to stream the response
+   * (Optional) Whether to stream the response.
    */
   stream?: boolean;
 
   /**
-   * (Optional) The stream options to use
+   * (Optional) The stream options to use.
    */
   stream_options?: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
   /**
-   * (Optional) The temperature to use
+   * (Optional) The temperature to use.
    */
   temperature?: number;
 
   /**
-   * (Optional) The top p to use
+   * (Optional) The top p to use.
    */
   top_p?: number;
 
   /**
-   * (Optional) The user to use
+   * (Optional) The user to use.
    */
   user?: string;
 }
@@ -246,14 +246,14 @@ export namespace CompletionCreateParams {
 
 export interface CompletionCreateParamsNonStreaming extends CompletionCreateParamsBase {
   /**
-   * (Optional) Whether to stream the response
+   * (Optional) Whether to stream the response.
    */
   stream?: false;
 }
 
 export interface CompletionCreateParamsStreaming extends CompletionCreateParamsBase {
   /**
-   * (Optional) Whether to stream the response
+   * (Optional) Whether to stream the response.
    */
   stream: true;
 }

package/src/resources/datasets.ts

@@ -5,10 +5,16 @@ import { isRequestOptions } from '../core';
 import * as Core from '../core';
 
 export class Datasets extends APIResource {
+  /**
+   * Get a dataset by its ID.
+   */
   retrieve(datasetId: string, options?: Core.RequestOptions): Core.APIPromise<DatasetRetrieveResponse> {
     return this._client.get(`/v1/datasets/${datasetId}`, options);
   }
 
+  /**
+   * List all datasets.
+   */
   list(options?: Core.RequestOptions): Core.APIPromise<DatasetListResponse> {
     return (
       this._client.get('/v1/datasets', options) as Core.APIPromise<{ data: DatasetListResponse }>
@@ -23,8 +29,8 @@ export class Datasets extends APIResource {
    *
    * The response includes:
    *
-   * - data: List of items for the current page
-   * - has_more: Whether there are more items available after this set
+   * - data: List of items for the current page.
+   * - has_more: Whether there are more items available after this set.
    */
   iterrows(
     datasetId: string,
@@ -53,6 +59,9 @@ export class Datasets extends APIResource {
     return this._client.post('/v1/datasets', { body, ...options });
   }
 
+  /**
+   * Unregister a dataset by its ID.
+   */
   unregister(datasetId: string, options?: Core.RequestOptions): Core.APIPromise<void> {
     return this._client.delete(`/v1/datasets/${datasetId}`, {
       ...options,
@@ -72,8 +81,6 @@ export interface DatasetRetrieveResponse {
 
   provider_id: string;
 
-  provider_resource_id: string;
-
   /**
    * Purpose of the dataset. Each purpose has a required input data schema.
    */
@@ -85,6 +92,8 @@ export interface DatasetRetrieveResponse {
   source: DatasetRetrieveResponse.UriDataSource | DatasetRetrieveResponse.RowsDataSource;
 
   type: 'dataset';
+
+  provider_resource_id?: string;
 }
 
 export namespace DatasetRetrieveResponse {
@@ -127,8 +136,6 @@ export namespace DatasetListResponse {
 
     provider_id: string;
 
-    provider_resource_id: string;
-
     /**
      * Purpose of the dataset. Each purpose has a required input data schema.
      */
@@ -140,6 +147,8 @@ export namespace DatasetListResponse {
     source: DatasetListResponseItem.UriDataSource | DatasetListResponseItem.RowsDataSource;
 
     type: 'dataset';
+
+    provider_resource_id?: string;
   }
 
   export namespace DatasetListResponseItem {
@@ -195,8 +204,6 @@ export interface DatasetRegisterResponse {
 
   provider_id: string;
 
-  provider_resource_id: string;
-
   /**
    * Purpose of the dataset. Each purpose has a required input data schema.
    */
@@ -208,6 +215,8 @@ export interface DatasetRegisterResponse {
   source: DatasetRegisterResponse.UriDataSource | DatasetRegisterResponse.RowsDataSource;
 
   type: 'dataset';
+
+  provider_resource_id?: string;
 }
 
 export namespace DatasetRegisterResponse {
@@ -254,7 +263,7 @@ export interface DatasetIterrowsParams {
 
 export interface DatasetRegisterParams {
   /**
-   * The purpose of the dataset. One of - "post-training/messages": The dataset
+   * The purpose of the dataset. One of: - "post-training/messages": The dataset
    * contains a messages column with list of messages for post-training. {
    * "messages": [ {"role": "user", "content": "Hello, world!"}, {"role":
    * "assistant", "content": "Hello, world!"}, ] } - "eval/question-answer": The
@@ -286,7 +295,7 @@ export interface DatasetRegisterParams {
   dataset_id?: string;
 
   /**
-   * The metadata for the dataset. - E.g. {"description": "My dataset"}
+   * The metadata for the dataset. - E.g. {"description": "My dataset"}.
    */
   metadata?: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 }

package/src/resources/index.ts

@@ -81,6 +81,16 @@ export {
   type PostTrainingSupervisedFineTuneParams,
 } from './post-training/post-training';
 export { Providers, type ListProvidersResponse, type ProviderListResponse } from './providers';
+export {
+  Responses,
+  type ResponseObject,
+  type ResponseObjectStream,
+  type ResponseListResponse,
+  type ResponseCreateParams,
+  type ResponseCreateParamsNonStreaming,
+  type ResponseCreateParamsStreaming,
+  type ResponseListParams,
+} from './responses';
 export { Routes, type ListRoutesResponse, type RouteListResponse } from './routes';
 export { Safety, type RunShieldResponse, type SafetyRunShieldParams } from './safety';
 export {

package/src/resources/inference.ts

@@ -8,6 +8,9 @@ import * as Shared from './shared';
 import { Stream } from '../streaming';
 
 export class Inference extends APIResource {
+  /**
+   * Generate chat completions for a batch of messages using the specified model.
+   */
   batchChatCompletion(
     body: InferenceBatchChatCompletionParams,
     options?: Core.RequestOptions,
@@ -15,6 +18,9 @@ export class Inference extends APIResource {
     return this._client.post('/v1/inference/batch-chat-completion', { body, ...options });
   }
 
+  /**
+   * Generate completions for a batch of content using the specified model.
+   */
   batchCompletion(
     body: InferenceBatchCompletionParams,
     options?: Core.RequestOptions,
@@ -192,31 +198,49 @@ export interface InferenceBatchChatCompletionResponse {
 }
 
 export interface InferenceBatchChatCompletionParams {
+  /**
+   * The messages to generate completions for.
+   */
   messages_batch: Array<Array<Shared.Message>>;
 
+  /**
+   * The identifier of the model to use. The model must be registered with Llama
+   * Stack and available via the /models endpoint.
+   */
   model_id: string;
 
+  /**
+   * (Optional) If specified, log probabilities for each token position will be
+   * returned.
+   */
   logprobs?: InferenceBatchChatCompletionParams.Logprobs;
 
   /**
-   * Configuration for JSON schema-guided response generation.
+   * (Optional) Grammar specification for guided (structured) decoding.
    */
   response_format?: Shared.ResponseFormat;
 
   /**
-   * Sampling parameters.
+   * (Optional) Parameters to control the sampling strategy.
    */
   sampling_params?: Shared.SamplingParams;
 
   /**
-   * Configuration for tool use.
+   * (Optional) Configuration for tool use.
    */
   tool_config?: InferenceBatchChatCompletionParams.ToolConfig;
 
+  /**
+   * (Optional) List of tool definitions available to the model.
+   */
   tools?: Array<InferenceBatchChatCompletionParams.Tool>;
 }
 
 export namespace InferenceBatchChatCompletionParams {
+  /**
+   * (Optional) If specified, log probabilities for each token position will be
+   * returned.
+   */
   export interface Logprobs {
     /**
      * How many tokens (for each position) to return log probabilities for.
@@ -225,7 +249,7 @@ export namespace InferenceBatchChatCompletionParams {
   }
 
   /**
-   * Configuration for tool use.
+   * (Optional) Configuration for tool use.
    */
   export interface ToolConfig {
     /**
@@ -265,24 +289,39 @@ export namespace InferenceBatchChatCompletionParams {
 }
 
 export interface InferenceBatchCompletionParams {
+  /**
+   * The content to generate completions for.
+   */
   content_batch: Array<Shared.InterleavedContent>;
 
+  /**
+   * The identifier of the model to use. The model must be registered with Llama
+   * Stack and available via the /models endpoint.
+   */
   model_id: string;
 
+  /**
+   * (Optional) If specified, log probabilities for each token position will be
+   * returned.
+   */
   logprobs?: InferenceBatchCompletionParams.Logprobs;
 
   /**
-   * Configuration for JSON schema-guided response generation.
+   * (Optional) Grammar specification for guided (structured) decoding.
    */
   response_format?: Shared.ResponseFormat;
 
   /**
-   * Sampling parameters.
+   * (Optional) Parameters to control the sampling strategy.
    */
   sampling_params?: Shared.SamplingParams;
 }
 
 export namespace InferenceBatchCompletionParams {
+  /**
+   * (Optional) If specified, log probabilities for each token position will be
+   * returned.
+   */
   export interface Logprobs {
     /**
      * How many tokens (for each position) to return log probabilities for.
@@ -297,7 +336,7 @@ export type InferenceChatCompletionParams =
 
 export interface InferenceChatCompletionParamsBase {
   /**
-   * List of messages in the conversation
+   * List of messages in the conversation.
    */
   messages: Array<Shared.Message>;
 
@@ -322,7 +361,7 @@ export interface InferenceChatCompletionParamsBase {
   response_format?: Shared.ResponseFormat;
 
   /**
-   * Parameters to control the sampling strategy
+   * Parameters to control the sampling strategy.
    */
   sampling_params?: Shared.SamplingParams;
 
@@ -355,7 +394,7 @@ export interface InferenceChatCompletionParamsBase {
   tool_prompt_format?: 'json' | 'function_tag' | 'python_list';
 
   /**
-   * (Optional) List of tool definitions available to the model
+   * (Optional) List of tool definitions available to the model.
    */
   tools?: Array<InferenceChatCompletionParams.Tool>;
 }
@@ -438,7 +477,7 @@ export type InferenceCompletionParams =
 
 export interface InferenceCompletionParamsBase {
   /**
-   * The content to generate a completion for
+   * The content to generate a completion for.
    */
   content: Shared.InterleavedContent;
 
@@ -455,12 +494,12 @@ export interface InferenceCompletionParamsBase {
   logprobs?: InferenceCompletionParams.Logprobs;
 
   /**
-   * (Optional) Grammar specification for guided (structured) decoding
+   * (Optional) Grammar specification for guided (structured) decoding.
    */
   response_format?: Shared.ResponseFormat;
 
   /**
-   * (Optional) Parameters to control the sampling strategy
+   * (Optional) Parameters to control the sampling strategy.
    */
   sampling_params?: Shared.SamplingParams;
 

package/src/resources/inspect.ts

@@ -4,10 +4,16 @@ import { APIResource } from '../resource';
 import * as Core from '../core';
 
 export class Inspect extends APIResource {
+  /**
+   * Get the health of the service.
+   */
   health(options?: Core.RequestOptions): Core.APIPromise<HealthInfo> {
     return this._client.get('/v1/health', options);
   }
 
+  /**
+   * Get the version of the service.
+   */
   version(options?: Core.RequestOptions): Core.APIPromise<VersionInfo> {
     return this._client.get('/v1/version', options);
   }

package/src/resources/models.ts

@@ -4,20 +4,32 @@ import { APIResource } from '../resource';
 import * as Core from '../core';
 
 export class Models extends APIResource {
+  /**
+   * Get a model by its identifier.
+   */
   retrieve(modelId: string, options?: Core.RequestOptions): Core.APIPromise<Model> {
     return this._client.get(`/v1/models/${modelId}`, options);
   }
 
+  /**
+   * List all models.
+   */
   list(options?: Core.RequestOptions): Core.APIPromise<ModelListResponse> {
     return (
       this._client.get('/v1/models', options) as Core.APIPromise<{ data: ModelListResponse }>
     )._thenUnwrap((obj) => obj.data);
   }
 
+  /**
+   * Register a model.
+   */
   register(body: ModelRegisterParams, options?: Core.RequestOptions): Core.APIPromise<Model> {
     return this._client.post('/v1/models', { body, ...options });
   }
 
+  /**
+   * Unregister a model.
+   */
   unregister(modelId: string, options?: Core.RequestOptions): Core.APIPromise<void> {
     return this._client.delete(`/v1/models/${modelId}`, {
       ...options,
@@ -39,22 +51,37 @@ export interface Model {
 
   provider_id: string;
 
-  provider_resource_id: string;
-
   type: 'model';
+
+  provider_resource_id?: string;
 }
 
 export type ModelListResponse = Array<Model>;
 
 export interface ModelRegisterParams {
+  /**
+   * The identifier of the model to register.
+   */
   model_id: string;
 
+  /**
+   * Any additional metadata for this model.
+   */
   metadata?: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
+  /**
+   * The type of model to register.
+   */
   model_type?: 'llm' | 'embedding';
 
+  /**
+   * The identifier of the provider.
+   */
   provider_id?: string;
 
+  /**
+   * The identifier of the model in the provider.
+   */
   provider_model_id?: string;
 }
 

package/src/resources/post-training/job.ts

@@ -6,6 +6,9 @@ import { ListPostTrainingJobsResponse } from './post-training';
 import * as PostTrainingAPI from './post-training';
 
 export class Job extends APIResource {
+  /**
+   * Get all training jobs.
+   */
   list(
     options?: Core.RequestOptions,
   ): Core.APIPromise<Array<PostTrainingAPI.ListPostTrainingJobsResponse.Data>> {
@@ -16,10 +19,16 @@ export class Job extends APIResource {
     )._thenUnwrap((obj) => obj.data);
   }
 
+  /**
+   * Get the artifacts of a training job.
+   */
   artifacts(query: JobArtifactsParams, options?: Core.RequestOptions): Core.APIPromise<JobArtifactsResponse> {
     return this._client.get('/v1/post-training/job/artifacts', { query, ...options });
   }
 
+  /**
+   * Cancel a training job.
+   */
   cancel(body: JobCancelParams, options?: Core.RequestOptions): Core.APIPromise<void> {
     return this._client.post('/v1/post-training/job/cancel', {
       body,
@@ -28,6 +37,9 @@ export class Job extends APIResource {
     });
   }
 
+  /**
+   * Get the status of a training job.
+   */
   status(query: JobStatusParams, options?: Core.RequestOptions): Core.APIPromise<JobStatusResponse> {
     return this._client.get('/v1/post-training/job/status', { query, ...options });
   }
@@ -70,14 +82,23 @@ export interface JobStatusResponse {
 }
 
 export interface JobArtifactsParams {
+  /**
+   * The UUID of the job to get the artifacts of.
+   */
   job_uuid: string;
 }
 
 export interface JobCancelParams {
+  /**
+   * The UUID of the job to cancel.
+   */
   job_uuid: string;
 }
 
 export interface JobStatusParams {
+  /**
+   * The UUID of the job to get the status of.
+   */
   job_uuid: string;
 }
 

package/src/resources/post-training/post-training.ts

@@ -16,6 +16,9 @@ import {
 export class PostTraining extends APIResource {
   job: JobAPI.Job = new JobAPI.Job(this._client);
 
+  /**
+   * Run preference optimization of a model.
+   */
   preferenceOptimize(
     body: PostTrainingPreferenceOptimizeParams,
     options?: Core.RequestOptions,
@@ -23,6 +26,9 @@ export class PostTraining extends APIResource {
     return this._client.post('/v1/post-training/preference-optimize', { body, ...options });
   }
 
+  /**
+   * Run supervised fine-tuning of a model.
+   */
   supervisedFineTune(
     body: PostTrainingSupervisedFineTuneParams,
     options?: Core.RequestOptions,
@@ -76,20 +82,41 @@ export interface PostTrainingJob {
 }
 
 export interface PostTrainingPreferenceOptimizeParams {
+  /**
+   * The algorithm configuration.
+   */
   algorithm_config: PostTrainingPreferenceOptimizeParams.AlgorithmConfig;
 
+  /**
+   * The model to fine-tune.
+   */
   finetuned_model: string;
 
+  /**
+   * The hyperparam search configuration.
+   */
   hyperparam_search_config: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
+  /**
+   * The UUID of the job to create.
+   */
   job_uuid: string;
 
+  /**
+   * The logger configuration.
+   */
   logger_config: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
+  /**
+   * The training configuration.
+   */
   training_config: PostTrainingPreferenceOptimizeParams.TrainingConfig;
 }
 
 export namespace PostTrainingPreferenceOptimizeParams {
+  /**
+   * The algorithm configuration.
+   */
   export interface AlgorithmConfig {
     epsilon: number;
 
@@ -100,6 +127,9 @@ export namespace PostTrainingPreferenceOptimizeParams {
     reward_scale: number;
   }
 
+  /**
+   * The training configuration.
+   */
   export interface TrainingConfig {
     gradient_accumulation_steps: number;
 
@@ -158,22 +188,46 @@ export namespace PostTrainingPreferenceOptimizeParams {
 }
 
 export interface PostTrainingSupervisedFineTuneParams {
+  /**
+   * The hyperparam search configuration.
+   */
   hyperparam_search_config: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
+  /**
+   * The UUID of the job to create.
+   */
   job_uuid: string;
 
+  /**
+   * The logger configuration.
+   */
   logger_config: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
+  /**
+   * The training configuration.
+   */
   training_config: PostTrainingSupervisedFineTuneParams.TrainingConfig;
 
+  /**
+   * The algorithm configuration.
+   */
   algorithm_config?: AlgorithmConfig;
 
+  /**
+   * The directory to save checkpoint(s) to.
+   */
   checkpoint_dir?: string;
 
+  /**
+   * The model to fine-tune.
+   */
   model?: string;
 }
 
 export namespace PostTrainingSupervisedFineTuneParams {
+  /**
+   * The training configuration.
+   */
   export interface TrainingConfig {
     gradient_accumulation_steps: number;