llama-stack-client 0.0.1-alpha.0

package/src/resources/evaluate/jobs/logs.ts

@@ -0,0 +1,42 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../../../resource';
+import * as Core from '../../../core';
+import * as LogsAPI from './logs';
+import * as JobsAPI from './jobs';
+
+export class Logs extends APIResource {
+  list(
+    params: LogListParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<JobsAPI.EvaluationJobLogStream> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...query } = params;
+    return this._client.get('/evaluate/job/logs', {
+      query,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+}
+
+export interface LogListParams {
+  /**
+   * Query param:
+   */
+  job_uuid: string;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace Logs {
+  export import LogListParams = LogsAPI.LogListParams;
+}

package/src/resources/evaluate/jobs/status.ts

@@ -0,0 +1,42 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../../../resource';
+import * as Core from '../../../core';
+import * as StatusAPI from './status';
+import * as JobsAPI from './jobs';
+
+export class Status extends APIResource {
+  list(
+    params: StatusListParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<JobsAPI.EvaluationJobStatus> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...query } = params;
+    return this._client.get('/evaluate/job/status', {
+      query,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+}
+
+export interface StatusListParams {
+  /**
+   * Query param:
+   */
+  job_uuid: string;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace Status {
+  export import StatusListParams = StatusAPI.StatusListParams;
+}

package/src/resources/evaluate/question-answering.ts

@@ -0,0 +1,42 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../../resource';
+import * as Core from '../../core';
+import * as QuestionAnsweringAPI from './question-answering';
+import * as EvaluateAPI from './evaluate';
+
+export class QuestionAnswering extends APIResource {
+  create(
+    params: QuestionAnsweringCreateParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<EvaluateAPI.EvaluationJob> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...body } = params;
+    return this._client.post('/evaluate/question_answering/', {
+      body,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+}
+
+export interface QuestionAnsweringCreateParams {
+  /**
+   * Body param:
+   */
+  metrics: Array<'em' | 'f1'>;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace QuestionAnswering {
+  export import QuestionAnsweringCreateParams = QuestionAnsweringAPI.QuestionAnsweringCreateParams;
+}

package/src/resources/evaluate/summarization.ts

@@ -0,0 +1,5 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../../resource';
+
+export class Summarization extends APIResource {}

package/src/resources/evaluate/text-generation.ts

@@ -0,0 +1,5 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../../resource';
+
+export class TextGeneration extends APIResource {}

package/src/resources/evaluations.ts

@@ -0,0 +1,73 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../resource';
+import * as Core from '../core';
+import * as EvaluationsAPI from './evaluations';
+import * as EvaluateAPI from './evaluate/evaluate';
+
+export class Evaluations extends APIResource {
+  summarization(
+    params: EvaluationSummarizationParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<EvaluateAPI.EvaluationJob> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...body } = params;
+    return this._client.post('/evaluate/summarization/', {
+      body,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+
+  textGeneration(
+    params: EvaluationTextGenerationParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<EvaluateAPI.EvaluationJob> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...body } = params;
+    return this._client.post('/evaluate/text_generation/', {
+      body,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+}
+
+export interface EvaluationSummarizationParams {
+  /**
+   * Body param:
+   */
+  metrics: Array<'rouge' | 'bleu'>;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export interface EvaluationTextGenerationParams {
+  /**
+   * Body param:
+   */
+  metrics: Array<'perplexity' | 'rouge' | 'bleu'>;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace Evaluations {
+  export import EvaluationSummarizationParams = EvaluationsAPI.EvaluationSummarizationParams;
+  export import EvaluationTextGenerationParams = EvaluationsAPI.EvaluationTextGenerationParams;
+}

package/src/resources/index.ts

@@ -0,0 +1,84 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+export * from './shared';
+export {
+  BatchChatCompletion,
+  BatchInferenceChatCompletionParams,
+  BatchInferenceCompletionParams,
+  BatchInference,
+} from './batch-inference';
+export {
+  ChatCompletionStreamChunk,
+  CompletionStreamChunk,
+  TokenLogProbs,
+  InferenceChatCompletionResponse,
+  InferenceCompletionResponse,
+  InferenceChatCompletionParams,
+  InferenceChatCompletionParamsNonStreaming,
+  InferenceChatCompletionParamsStreaming,
+  InferenceCompletionParams,
+  Inference,
+} from './inference/inference';
+export { EvaluationJob, Evaluate } from './evaluate/evaluate';
+export { EvaluationSummarizationParams, EvaluationTextGenerationParams, Evaluations } from './evaluations';
+export {
+  InferenceStep,
+  MemoryRetrievalStep,
+  RestAPIExecutionConfig,
+  ShieldCallStep,
+  ToolExecutionStep,
+  ToolParamDefinition,
+  AgentCreateResponse,
+  AgentCreateParams,
+  AgentDeleteParams,
+  Agents,
+} from './agents/agents';
+export { MemoryBankSpec, MemoryBankListParams, MemoryBankGetParams, MemoryBanks } from './memory-banks';
+export { ModelServingSpec, ModelListParams, ModelGetParams, Models } from './models';
+export {
+  PostTrainingJob,
+  PostTrainingPreferenceOptimizeParams,
+  PostTrainingSupervisedFineTuneParams,
+  PostTraining,
+} from './post-training/post-training';
+export {
+  QueryDocuments,
+  MemoryCreateResponse,
+  MemoryRetrieveResponse,
+  MemoryListResponse,
+  MemoryDropResponse,
+  MemoryCreateParams,
+  MemoryRetrieveParams,
+  MemoryUpdateParams,
+  MemoryListParams,
+  MemoryDropParams,
+  MemoryInsertParams,
+  MemoryQueryParams,
+  Memory,
+} from './memory/memory';
+export {
+  RewardScoring,
+  ScoredDialogGenerations,
+  RewardScoringScoreParams,
+  RewardScoringResource,
+} from './reward-scoring';
+export { RunSheidResponse, SafetyRunShieldParams, Safety } from './safety';
+export { ShieldSpec, ShieldListParams, ShieldGetParams, Shields } from './shields';
+export {
+  SyntheticDataGeneration,
+  SyntheticDataGenerationGenerateParams,
+  SyntheticDataGenerationResource,
+} from './synthetic-data-generation';
+export {
+  TelemetryGetTraceResponse,
+  TelemetryGetTraceParams,
+  TelemetryLogParams,
+  Telemetry,
+} from './telemetry';
+export {
+  TrainEvalDataset,
+  DatasetCreateParams,
+  DatasetDeleteParams,
+  DatasetGetParams,
+  Datasets,
+} from './datasets';

package/src/resources/inference/embeddings.ts

@@ -0,0 +1,48 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+import { APIResource } from '../../resource';
+import * as Core from '../../core';
+import * as EmbeddingsAPI from './embeddings';
+
+export class Embeddings extends APIResource {
+  create(params: EmbeddingCreateParams, options?: Core.RequestOptions): Core.APIPromise<Embeddings> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...body } = params;
+    return this._client.post('/inference/embeddings', {
+      body,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+}
+
+export interface Embeddings {
+  embeddings: Array<Array<number>>;
+}
+
+export interface EmbeddingCreateParams {
+  /**
+   * Body param:
+   */
+  contents: Array<string | Array<string>>;
+
+  /**
+   * Body param:
+   */
+  model: string;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace Embeddings {
+  export import Embeddings = EmbeddingsAPI.Embeddings;
+  export import EmbeddingCreateParams = EmbeddingsAPI.EmbeddingCreateParams;
+}

package/src/resources/inference/index.ts

@@ -0,0 +1,15 @@
+// File generated from our OpenAPI spec by Stainless. See CONTRIBUTING.md for details.
+
+export {
+  ChatCompletionStreamChunk,
+  CompletionStreamChunk,
+  TokenLogProbs,
+  InferenceChatCompletionResponse,
+  InferenceCompletionResponse,
+  InferenceChatCompletionParams,
+  InferenceChatCompletionParamsNonStreaming,
+  InferenceChatCompletionParamsStreaming,
+  InferenceCompletionParams,
+  Inference,
+} from './inference';
+export { Embeddings, EmbeddingCreateParams } from './embeddings';

package/src/resources/inference/inference.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 InferenceAPI from './inference';
+import * as Shared from '../shared';
+import * as AgentsAPI from '../agents/agents';
+import * as EmbeddingsAPI from './embeddings';
+import { Stream } from '../../streaming';
+
+export class Inference extends APIResource {
+  embeddings: EmbeddingsAPI.Embeddings = new EmbeddingsAPI.Embeddings(this._client);
+
+  chatCompletion(
+    params: InferenceChatCompletionParamsNonStreaming,
+    options?: Core.RequestOptions,
+  ): APIPromise<InferenceChatCompletionResponse>;
+  chatCompletion(
+    params: InferenceChatCompletionParamsStreaming,
+    options?: Core.RequestOptions,
+  ): APIPromise<Stream<InferenceChatCompletionResponse>>;
+  chatCompletion(
+    params: InferenceChatCompletionParamsBase,
+    options?: Core.RequestOptions,
+  ): APIPromise<Stream<InferenceChatCompletionResponse> | InferenceChatCompletionResponse>;
+  chatCompletion(
+    params: InferenceChatCompletionParams,
+    options?: Core.RequestOptions,
+  ): APIPromise<InferenceChatCompletionResponse> | APIPromise<Stream<InferenceChatCompletionResponse>> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...body } = params;
+    return this._client.post('/inference/chat_completion', {
+      body,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+      stream: params.stream ?? false,
+    }) as APIPromise<InferenceChatCompletionResponse> | APIPromise<Stream<InferenceChatCompletionResponse>>;
+  }
+
+  completion(
+    params: InferenceCompletionParams,
+    options?: Core.RequestOptions,
+  ): Core.APIPromise<InferenceCompletionResponse> {
+    const { 'X-LlamaStack-ProviderData': xLlamaStackProviderData, ...body } = params;
+    return this._client.post('/inference/completion', {
+      body,
+      ...options,
+      headers: {
+        ...(xLlamaStackProviderData != null ?
+          { 'X-LlamaStack-ProviderData': xLlamaStackProviderData }
+        : undefined),
+        ...options?.headers,
+      },
+    });
+  }
+}
+
+export interface ChatCompletionStreamChunk {
+  event: ChatCompletionStreamChunk.Event;
+}
+
+export namespace ChatCompletionStreamChunk {
+  export interface Event {
+    delta: string | Event.ToolCallDelta;
+
+    event_type: 'start' | 'complete' | 'progress';
+
+    logprobs?: Array<InferenceAPI.TokenLogProbs>;
+
+    stop_reason?: 'end_of_turn' | 'end_of_message' | 'out_of_tokens';
+  }
+
+  export namespace Event {
+    export interface ToolCallDelta {
+      content: string | Shared.ToolCall;
+
+      parse_status: 'started' | 'in_progress' | 'failure' | 'success';
+    }
+  }
+}
+
+export interface CompletionStreamChunk {
+  delta: string;
+
+  logprobs?: Array<TokenLogProbs>;
+
+  stop_reason?: 'end_of_turn' | 'end_of_message' | 'out_of_tokens';
+}
+
+export interface TokenLogProbs {
+  logprobs_by_token: Record<string, number>;
+}
+
+export type InferenceChatCompletionResponse =
+  | InferenceChatCompletionResponse.ChatCompletionResponse
+  | ChatCompletionStreamChunk;
+
+export namespace InferenceChatCompletionResponse {
+  export interface ChatCompletionResponse {
+    completion_message: Shared.CompletionMessage;
+
+    logprobs?: Array<InferenceAPI.TokenLogProbs>;
+  }
+}
+
+export type InferenceCompletionResponse =
+  | InferenceCompletionResponse.CompletionResponse
+  | CompletionStreamChunk;
+
+export namespace InferenceCompletionResponse {
+  export interface CompletionResponse {
+    completion_message: Shared.CompletionMessage;
+
+    logprobs?: Array<InferenceAPI.TokenLogProbs>;
+  }
+}
+
+export type InferenceChatCompletionParams =
+  | InferenceChatCompletionParamsNonStreaming
+  | InferenceChatCompletionParamsStreaming;
+
+export interface InferenceChatCompletionParamsBase {
+  /**
+   * Body param:
+   */
+  messages: Array<
+    Shared.UserMessage | Shared.SystemMessage | Shared.ToolResponseMessage | Shared.CompletionMessage
+  >;
+
+  /**
+   * Body param:
+   */
+  model: string;
+
+  /**
+   * Body param:
+   */
+  logprobs?: InferenceChatCompletionParams.Logprobs;
+
+  /**
+   * Body param:
+   */
+  sampling_params?: Shared.SamplingParams;
+
+  /**
+   * Body param:
+   */
+  stream?: boolean;
+
+  /**
+   * Body param:
+   */
+  tool_choice?: 'auto' | 'required';
+
+  /**
+   * Body param: `json` -- Refers to the json format for calling tools. The json
+   * format takes the form like { "type": "function", "function" : { "name":
+   * "function_name", "description": "function_description", "parameters": {...} } }
+   *
+   * `function_tag` -- This is an example of how you could define your own user
+   * defined format for making tool calls. The function_tag format looks like this,
+   * <function=function_name>(parameters)</function>
+   *
+   * The detailed prompts for each of these formats are added to llama cli
+   */
+  tool_prompt_format?: 'json' | 'function_tag';
+
+  /**
+   * Body param:
+   */
+  tools?: Array<InferenceChatCompletionParams.Tool>;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace InferenceChatCompletionParams {
+  export interface Logprobs {
+    top_k?: number;
+  }
+
+  export interface Tool {
+    tool_name: 'brave_search' | 'wolfram_alpha' | 'photogen' | 'code_interpreter' | (string & {});
+
+    description?: string;
+
+    parameters?: Record<string, AgentsAPI.ToolParamDefinition>;
+  }
+
+  export type InferenceChatCompletionParamsNonStreaming =
+    InferenceAPI.InferenceChatCompletionParamsNonStreaming;
+  export type InferenceChatCompletionParamsStreaming = InferenceAPI.InferenceChatCompletionParamsStreaming;
+}
+
+export interface InferenceChatCompletionParamsNonStreaming extends InferenceChatCompletionParamsBase {
+  /**
+   * Body param:
+   */
+  stream?: false;
+}
+
+export interface InferenceChatCompletionParamsStreaming extends InferenceChatCompletionParamsBase {
+  /**
+   * Body param:
+   */
+  stream: true;
+}
+
+export interface InferenceCompletionParams {
+  /**
+   * Body param:
+   */
+  content: string | Array<string>;
+
+  /**
+   * Body param:
+   */
+  model: string;
+
+  /**
+   * Body param:
+   */
+  logprobs?: InferenceCompletionParams.Logprobs;
+
+  /**
+   * Body param:
+   */
+  sampling_params?: Shared.SamplingParams;
+
+  /**
+   * Body param:
+   */
+  stream?: boolean;
+
+  /**
+   * Header param: JSON-encoded provider data which will be made available to the
+   * adapter servicing the API
+   */
+  'X-LlamaStack-ProviderData'?: string;
+}
+
+export namespace InferenceCompletionParams {
+  export interface Logprobs {
+    top_k?: number;
+  }
+}
+
+export namespace Inference {
+  export import ChatCompletionStreamChunk = InferenceAPI.ChatCompletionStreamChunk;
+  export import CompletionStreamChunk = InferenceAPI.CompletionStreamChunk;
+  export import TokenLogProbs = InferenceAPI.TokenLogProbs;
+  export import InferenceChatCompletionResponse = InferenceAPI.InferenceChatCompletionResponse;
+  export import InferenceCompletionResponse = InferenceAPI.InferenceCompletionResponse;
+  export import InferenceChatCompletionParams = InferenceAPI.InferenceChatCompletionParams;
+  export import InferenceChatCompletionParamsNonStreaming = InferenceAPI.InferenceChatCompletionParamsNonStreaming;
+  export import InferenceChatCompletionParamsStreaming = InferenceAPI.InferenceChatCompletionParamsStreaming;
+  export import InferenceCompletionParams = InferenceAPI.InferenceCompletionParams;
+  export import Embeddings = EmbeddingsAPI.Embeddings;
+  export import EmbeddingCreateParams = EmbeddingsAPI.EmbeddingCreateParams;
+}