llama-stack-client 0.2.2 → 0.2.3

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/index.ts

@@ -18,6 +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,

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

@@ -2,6 +2,7 @@
 
 import { APIResource } from '../../resource';
 import * as Core from '../../core';
+import { ListPostTrainingJobsResponse } from './post-training';
 import * as PostTrainingAPI from './post-training';
 
 export class Job extends APIResource {

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;
+    }
   }
 }
 

package/src/resources/shared.ts

@@ -4,23 +4,40 @@ import * as Shared from './shared';
 import * as InferenceAPI from './inference';
 import * as ToolRuntimeAPI from './tool-runtime/tool-runtime';
 
+/**
+ * Configuration for an agent.
+ */
 export interface AgentConfig {
+  /**
+   * The system instructions for the agent
+   */
   instructions: string;
 
+  /**
+   * The model identifier to use for the agent
+   */
   model: string;
 
   client_tools?: Array<ToolRuntimeAPI.ToolDef>;
 
+  /**
+   * Optional flag indicating whether session data has to be persisted
+   */
   enable_session_persistence?: boolean;
 
   input_shields?: Array<string>;
 
   max_infer_iters?: number;
 
+  /**
+   * Optional name for the agent, used in telemetry and identification
+   */
+  name?: string;
+
   output_shields?: Array<string>;
 
   /**
-   * Configuration for JSON schema-guided response generation.
+   * Optional response format configuration
    */
   response_format?: ResponseFormat;
 

package/src/version.ts

@@ -1 +1 @@
-export const VERSION = '0.2.2';
+export const VERSION = '0.2.3';

package/version.d.ts

@@ -1,2 +1,2 @@
-export declare const VERSION = "0.2.2";
+export declare const VERSION = "0.2.3";
 //# sourceMappingURL=version.d.ts.map

package/version.js

@@ -1,5 +1,5 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.VERSION = void 0;
-exports.VERSION = '0.2.2';
+exports.VERSION = '0.2.3';
 //# sourceMappingURL=version.js.map

package/version.mjs

@@ -1,2 +1,2 @@
-export const VERSION = '0.2.2';
+export const VERSION = '0.2.3';
 //# sourceMappingURL=version.mjs.map