llama-stack-client 0.2.0 → 0.2.3-rc5

package/src/resources/chat/index.ts

@@ -0,0 +1,10 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+export { Chat, type ChatCompletionChunk } from './chat';
+export {
+  Completions,
+  type CompletionCreateResponse,
+  type CompletionCreateParams,
+  type CompletionCreateParamsNonStreaming,
+  type CompletionCreateParamsStreaming,
+} from './completions';

package/src/resources/chat.ts

@@ -0,0 +1,3 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+export * from './chat/index';

package/src/resources/completions.ts

@@ -0,0 +1,268 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../resource';
+import { APIPromise } from '../core';
+import * as Core from '../core';
+import * as CompletionsAPI from './completions';
+import { Stream } from '../streaming';
+
+export class Completions extends APIResource {
+  /**
+   * Generate an OpenAI-compatible completion for the given prompt using the
+   * specified model.
+   */
+  create(
+    body: CompletionCreateParamsNonStreaming,
+    options?: Core.RequestOptions,
+  ): APIPromise<CompletionCreateResponse>;
+  create(
+    body: CompletionCreateParamsStreaming,
+    options?: Core.RequestOptions,
+  ): APIPromise<Stream<CompletionCreateResponse>>;
+  create(
+    body: CompletionCreateParamsBase,
+    options?: Core.RequestOptions,
+  ): APIPromise<Stream<CompletionCreateResponse> | CompletionCreateResponse>;
+  create(
+    body: CompletionCreateParams,
+    options?: Core.RequestOptions,
+  ): APIPromise<CompletionCreateResponse> | APIPromise<Stream<CompletionCreateResponse>> {
+    return this._client.post('/v1/openai/v1/completions', {
+      body,
+      ...options,
+      stream: body.stream ?? false,
+    }) as APIPromise<CompletionCreateResponse> | APIPromise<Stream<CompletionCreateResponse>>;
+  }
+}
+
+/**
+ * Response from an OpenAI-compatible completion request.
+ */
+export interface CompletionCreateResponse {
+  id: string;
+
+  choices: Array<CompletionCreateResponse.Choice>;
+
+  created: number;
+
+  model: string;
+
+  object: 'text_completion';
+}
+
+export namespace CompletionCreateResponse {
+  /**
+   * A choice from an OpenAI-compatible completion response.
+   */
+  export interface Choice {
+    finish_reason: string;
+
+    index: number;
+
+    text: string;
+
+    /**
+     * The log probabilities for the tokens in the message from an OpenAI-compatible
+     * chat completion response.
+     */
+    logprobs?: Choice.Logprobs;
+  }
+
+  export namespace Choice {
+    /**
+     * The log probabilities for the tokens in the message from an OpenAI-compatible
+     * chat completion response.
+     */
+    export interface Logprobs {
+      /**
+       * (Optional) The log probabilities for the tokens in the message
+       */
+      content?: Array<Logprobs.Content>;
+
+      /**
+       * (Optional) The log probabilities for the tokens in the message
+       */
+      refusal?: Array<Logprobs.Refusal>;
+    }
+
+    export namespace Logprobs {
+      /**
+       * The log probability for a token from an OpenAI-compatible chat completion
+       * response.
+       */
+      export interface Content {
+        token: string;
+
+        logprob: number;
+
+        top_logprobs: Array<Content.TopLogprob>;
+
+        bytes?: Array<number>;
+      }
+
+      export namespace Content {
+        /**
+         * The top log probability for a token from an OpenAI-compatible chat completion
+         * response.
+         */
+        export interface TopLogprob {
+          token: string;
+
+          logprob: number;
+
+          bytes?: Array<number>;
+        }
+      }
+
+      /**
+       * The log probability for a token from an OpenAI-compatible chat completion
+       * response.
+       */
+      export interface Refusal {
+        token: string;
+
+        logprob: number;
+
+        top_logprobs: Array<Refusal.TopLogprob>;
+
+        bytes?: Array<number>;
+      }
+
+      export namespace Refusal {
+        /**
+         * The top log probability for a token from an OpenAI-compatible chat completion
+         * response.
+         */
+        export interface TopLogprob {
+          token: string;
+
+          logprob: number;
+
+          bytes?: Array<number>;
+        }
+      }
+    }
+  }
+}
+
+export type CompletionCreateParams = CompletionCreateParamsNonStreaming | CompletionCreateParamsStreaming;
+
+export interface CompletionCreateParamsBase {
+  /**
+   * The identifier of the model to use. The model must be registered with Llama
+   * Stack and available via the /models endpoint.
+   */
+  model: string;
+
+  /**
+   * The prompt to generate a completion for
+   */
+  prompt: string | Array<string> | Array<number> | Array<Array<number>>;
+
+  /**
+   * (Optional) The number of completions to generate
+   */
+  best_of?: number;
+
+  /**
+   * (Optional) Whether to echo the prompt
+   */
+  echo?: boolean;
+
+  /**
+   * (Optional) The penalty for repeated tokens
+   */
+  frequency_penalty?: number;
+
+  guided_choice?: Array<string>;
+
+  /**
+   * (Optional) The logit bias to use
+   */
+  logit_bias?: Record<string, number>;
+
+  /**
+   * (Optional) The log probabilities to use
+   */
+  logprobs?: boolean;
+
+  /**
+   * (Optional) The maximum number of tokens to generate
+   */
+  max_tokens?: number;
+
+  /**
+   * (Optional) The number of completions to generate
+   */
+  n?: number;
+
+  /**
+   * (Optional) The penalty for repeated tokens
+   */
+  presence_penalty?: number;
+
+  prompt_logprobs?: number;
+
+  /**
+   * (Optional) The seed to use
+   */
+  seed?: number;
+
+  /**
+   * (Optional) The stop tokens to use
+   */
+  stop?: string | Array<string>;
+
+  /**
+   * (Optional) Whether to stream the response
+   */
+  stream?: boolean;
+
+  /**
+   * (Optional) The stream options to use
+   */
+  stream_options?: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
+
+  /**
+   * (Optional) The temperature to use
+   */
+  temperature?: number;
+
+  /**
+   * (Optional) The top p to use
+   */
+  top_p?: number;
+
+  /**
+   * (Optional) The user to use
+   */
+  user?: string;
+}
+
+export namespace CompletionCreateParams {
+  export type CompletionCreateParamsNonStreaming = CompletionsAPI.CompletionCreateParamsNonStreaming;
+  export type CompletionCreateParamsStreaming = CompletionsAPI.CompletionCreateParamsStreaming;
+}
+
+export interface CompletionCreateParamsNonStreaming extends CompletionCreateParamsBase {
+  /**
+   * (Optional) Whether to stream the response
+   */
+  stream?: false;
+}
+
+export interface CompletionCreateParamsStreaming extends CompletionCreateParamsBase {
+  /**
+   * (Optional) Whether to stream the response
+   */
+  stream: true;
+}
+
+export declare namespace Completions {
+  export {
+    type CompletionCreateResponse as CompletionCreateResponse,
+    type CompletionCreateParams as CompletionCreateParams,
+    type CompletionCreateParamsNonStreaming as CompletionCreateParamsNonStreaming,
+    type CompletionCreateParamsStreaming as CompletionCreateParamsStreaming,
+  };
+}

package/src/resources/datasets.ts

@@ -16,7 +16,15 @@ export class Datasets extends APIResource {
   }
 
   /**
-   * Get a paginated list of rows from a dataset. Uses cursor-based pagination.
+   * Get a paginated list of rows from a dataset. Uses offset-based pagination where:
+   *
+   * - start_index: The starting index (0-based). If None, starts from beginning.
+   * - limit: Number of items to return. If None or -1, returns all items.
+   *
+   * The response includes:
+   *
+   * - data: List of items for the current page
+   * - has_more: Whether there are more items available after this set
    */
   iterrows(
     datasetId: string,
@@ -166,19 +174,18 @@ export namespace DatasetListResponse {
 }
 
 /**
- * A paginated list of rows from a dataset.
+ * A generic paginated response that follows a simple format.
  */
 export interface DatasetIterrowsResponse {
   /**
-   * The rows in the current page.
+   * The list of items for the current page
    */
   data: Array<Record<string, boolean | number | string | Array<unknown> | unknown | null>>;
 
   /**
-   * Index into dataset for the first row in the next page. None if there are no more
-   * rows.
+   * Whether there are more items available after this set
    */
-  next_start_index?: number;
+  has_more: boolean;
 }
 
 export interface DatasetRegisterResponse {

package/src/resources/eval/eval.ts

@@ -132,7 +132,7 @@ export interface EvaluateResponse {
 export interface Job {
   job_id: string;
 
-  status: 'completed' | 'in_progress' | 'failed' | 'scheduled';
+  status: 'completed' | 'in_progress' | 'failed' | 'scheduled' | 'cancelled';
 }
 
 export interface EvalEvaluateRowsParams {

package/src/resources/index.ts

@@ -11,12 +11,6 @@ export {
   type AgentCreateResponse,
   type AgentCreateParams,
 } from './agents/agents';
-export {
-  BatchInference,
-  type BatchInferenceChatCompletionResponse,
-  type BatchInferenceChatCompletionParams,
-  type BatchInferenceCompletionParams,
-} from './batch-inference';
 export {
   Benchmarks,
   type Benchmark,
@@ -24,6 +18,14 @@ export {
   type BenchmarkListResponse,
   type BenchmarkRegisterParams,
 } from './benchmarks';
+export { Chat, type ChatCompletionChunk } from './chat/chat';
+export {
+  Completions,
+  type CompletionCreateResponse,
+  type CompletionCreateParams,
+  type CompletionCreateParamsNonStreaming,
+  type CompletionCreateParamsStreaming,
+} from './completions';
 export {
   Datasets,
   type ListDatasetsResponse,
@@ -51,6 +53,9 @@ export {
   type CompletionResponse,
   type EmbeddingsResponse,
   type TokenLogProbs,
+  type InferenceBatchChatCompletionResponse,
+  type InferenceBatchChatCompletionParams,
+  type InferenceBatchCompletionParams,
   type InferenceChatCompletionParams,
   type InferenceChatCompletionParamsNonStreaming,
   type InferenceChatCompletionParamsStreaming,
@@ -126,6 +131,7 @@ export {
   ToolRuntime,
   type ToolDef,
   type ToolInvocationResult,
+  type ToolRuntimeListToolsResponse,
   type ToolRuntimeInvokeToolParams,
   type ToolRuntimeListToolsParams,
 } from './tool-runtime/tool-runtime';

package/src/resources/inference.ts

@@ -8,6 +8,20 @@ import * as Shared from './shared';
 import { Stream } from '../streaming';
 
 export class Inference extends APIResource {
+  batchChatCompletion(
+    body: InferenceBatchChatCompletionParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<InferenceBatchChatCompletionResponse> {
+    return this._client.post('/v1/inference/batch-chat-completion', { body, ...options });
+  }
+
+  batchCompletion(
+    body: InferenceBatchCompletionParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<Shared.BatchCompletion> {
+    return this._client.post('/v1/inference/batch-completion', { body, ...options });
+  }
+
   /**
    * Generate a chat completion for the given messages using the specified model.
    */
@@ -173,6 +187,110 @@ export interface TokenLogProbs {
   logprobs_by_token: Record<string, number>;
 }
 
+export interface InferenceBatchChatCompletionResponse {
+  batch: Array<Shared.ChatCompletionResponse>;
+}
+
+export interface InferenceBatchChatCompletionParams {
+  messages_batch: Array<Array<Shared.Message>>;
+
+  model_id: string;
+
+  logprobs?: InferenceBatchChatCompletionParams.Logprobs;
+
+  /**
+   * Configuration for JSON schema-guided response generation.
+   */
+  response_format?: Shared.ResponseFormat;
+
+  /**
+   * Sampling parameters.
+   */
+  sampling_params?: Shared.SamplingParams;
+
+  /**
+   * Configuration for tool use.
+   */
+  tool_config?: InferenceBatchChatCompletionParams.ToolConfig;
+
+  tools?: Array<InferenceBatchChatCompletionParams.Tool>;
+}
+
+export namespace InferenceBatchChatCompletionParams {
+  export interface Logprobs {
+    /**
+     * How many tokens (for each position) to return log probabilities for.
+     */
+    top_k?: number;
+  }
+
+  /**
+   * Configuration for tool use.
+   */
+  export interface ToolConfig {
+    /**
+     * (Optional) Config for how to override the default system prompt. -
+     * `SystemMessageBehavior.append`: Appends the provided system message to the
+     * default system prompt. - `SystemMessageBehavior.replace`: Replaces the default
+     * system prompt with the provided system message. The system message can include
+     * the string '{{function_definitions}}' to indicate where the function definitions
+     * should be inserted.
+     */
+    system_message_behavior?: 'append' | 'replace';
+
+    /**
+     * (Optional) Whether tool use is automatic, required, or none. Can also specify a
+     * tool name to use a specific tool. Defaults to ToolChoice.auto.
+     */
+    tool_choice?: 'auto' | 'required' | 'none' | (string & {});
+
+    /**
+     * (Optional) Instructs the model how to format tool calls. By default, Llama Stack
+     * will attempt to use a format that is best adapted to the model. -
+     * `ToolPromptFormat.json`: The tool calls are formatted as a JSON object. -
+     * `ToolPromptFormat.function_tag`: The tool calls are enclosed in a
+     * <function=function_name> tag. - `ToolPromptFormat.python_list`: The tool calls
+     * are output as Python syntax -- a list of function calls.
+     */
+    tool_prompt_format?: 'json' | 'function_tag' | 'python_list';
+  }
+
+  export interface Tool {
+    tool_name: 'brave_search' | 'wolfram_alpha' | 'photogen' | 'code_interpreter' | (string & {});
+
+    description?: string;
+
+    parameters?: Record<string, Shared.ToolParamDefinition>;
+  }
+}
+
+export interface InferenceBatchCompletionParams {
+  content_batch: Array<Shared.InterleavedContent>;
+
+  model_id: string;
+
+  logprobs?: InferenceBatchCompletionParams.Logprobs;
+
+  /**
+   * Configuration for JSON schema-guided response generation.
+   */
+  response_format?: Shared.ResponseFormat;
+
+  /**
+   * Sampling parameters.
+   */
+  sampling_params?: Shared.SamplingParams;
+}
+
+export namespace InferenceBatchCompletionParams {
+  export interface Logprobs {
+    /**
+     * How many tokens (for each position) to return log probabilities for.
+     */
+    top_k?: number;
+  }
+}
+
 export type InferenceChatCompletionParams =
   | InferenceChatCompletionParamsNonStreaming
   | InferenceChatCompletionParamsStreaming;
@@ -424,6 +542,9 @@ export declare namespace Inference {
     type CompletionResponse as CompletionResponse,
     type EmbeddingsResponse as EmbeddingsResponse,
     type TokenLogProbs as TokenLogProbs,
+    type InferenceBatchChatCompletionResponse as InferenceBatchChatCompletionResponse,
+    type InferenceBatchChatCompletionParams as InferenceBatchChatCompletionParams,
+    type InferenceBatchCompletionParams as InferenceBatchCompletionParams,
     type InferenceChatCompletionParams as InferenceChatCompletionParams,
     type InferenceChatCompletionParamsNonStreaming as InferenceChatCompletionParamsNonStreaming,
     type InferenceChatCompletionParamsStreaming as InferenceChatCompletionParamsStreaming,

package/src/resources/inspect.ts

@@ -14,7 +14,7 @@ export class Inspect extends APIResource {
 }
 
 export interface HealthInfo {
-  status: string;
+  status: 'OK' | 'Error' | 'Not Implemented';
 }
 
 export interface ProviderInfo {
@@ -22,6 +22,8 @@ export interface ProviderInfo {
 
   config: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
+  health: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
+
   provider_id: string;
 
   provider_type: string;

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

@@ -58,7 +58,7 @@ export interface JobStatusResponse {
 
   job_uuid: string;
 
-  status: 'completed' | 'in_progress' | 'failed' | 'scheduled';
+  status: 'completed' | 'in_progress' | 'failed' | 'scheduled' | 'cancelled';
 
   completed_at?: string;
 

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

@@ -101,21 +101,21 @@ export namespace PostTrainingPreferenceOptimizeParams {
   }
 
   export interface TrainingConfig {
-    data_config: TrainingConfig.DataConfig;
-
     gradient_accumulation_steps: number;
 
     max_steps_per_epoch: number;
 
-    max_validation_steps: number;
-
     n_epochs: number;
 
-    optimizer_config: TrainingConfig.OptimizerConfig;
+    data_config?: TrainingConfig.DataConfig;
 
     dtype?: string;
 
     efficiency_config?: TrainingConfig.EfficiencyConfig;
+
+    max_validation_steps?: number;
+
+    optimizer_config?: TrainingConfig.OptimizerConfig;
   }
 
   export namespace TrainingConfig {
@@ -135,16 +135,6 @@ export namespace PostTrainingPreferenceOptimizeParams {
       validation_dataset_id?: string;
     }
 
-    export interface OptimizerConfig {
-      lr: number;
-
-      num_warmup_steps: number;
-
-      optimizer_type: 'adam' | 'adamw' | 'sgd';
-
-      weight_decay: number;
-    }
-
     export interface EfficiencyConfig {
       enable_activation_checkpointing?: boolean;
 
@@ -154,6 +144,16 @@ export namespace PostTrainingPreferenceOptimizeParams {
 
       memory_efficient_fsdp_wrap?: boolean;
     }
+
+    export interface OptimizerConfig {
+      lr: number;
+
+      num_warmup_steps: number;
+
+      optimizer_type: 'adam' | 'adamw' | 'sgd';
+
+      weight_decay: number;
+    }
   }
 }
 
@@ -164,32 +164,32 @@ export interface PostTrainingSupervisedFineTuneParams {
 
   logger_config: Record<string, boolean | number | string | Array<unknown> | unknown | null>;
 
-  model: string;
-
   training_config: PostTrainingSupervisedFineTuneParams.TrainingConfig;
 
   algorithm_config?: AlgorithmConfig;
 
   checkpoint_dir?: string;
+
+  model?: string;
 }
 
 export namespace PostTrainingSupervisedFineTuneParams {
   export interface TrainingConfig {
-    data_config: TrainingConfig.DataConfig;
-
     gradient_accumulation_steps: number;
 
     max_steps_per_epoch: number;
 
-    max_validation_steps: number;
-
     n_epochs: number;
 
-    optimizer_config: TrainingConfig.OptimizerConfig;
+    data_config?: TrainingConfig.DataConfig;
 
     dtype?: string;
 
     efficiency_config?: TrainingConfig.EfficiencyConfig;
+
+    max_validation_steps?: number;
+
+    optimizer_config?: TrainingConfig.OptimizerConfig;
   }
 
   export namespace TrainingConfig {
@@ -209,16 +209,6 @@ export namespace PostTrainingSupervisedFineTuneParams {
       validation_dataset_id?: string;
     }
 
-    export interface OptimizerConfig {
-      lr: number;
-
-      num_warmup_steps: number;
-
-      optimizer_type: 'adam' | 'adamw' | 'sgd';
-
-      weight_decay: number;
-    }
-
     export interface EfficiencyConfig {
       enable_activation_checkpointing?: boolean;
 
@@ -228,6 +218,16 @@ export namespace PostTrainingSupervisedFineTuneParams {
 
       memory_efficient_fsdp_wrap?: boolean;
     }
+
+    export interface OptimizerConfig {
+      lr: number;
+
+      num_warmup_steps: number;
+
+      optimizer_type: 'adam' | 'adamw' | 'sgd';
+
+      weight_decay: number;
+    }
   }
 }