{"version":3,"file":"openai-responses.d.ts","sourceRoot":"","sources":["../../src/providers/openai-responses.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,6BAA6B,EAAE,MAAM,yCAAyC,CAAC;AAG7F,OAAO,KAAK,EAOX,mBAAmB,EACnB,cAAc,EACd,aAAa,EAEb,MAAM,aAAa,CAAC;AAuCrB,MAAM,WAAW,sBAAuB,SAAQ,aAAa;IAC5D,eAAe,CAAC,EAAE,SAAS,GAAG,KAAK,GAAG,QAAQ,GAAG,MAAM,GAAG,OAAO,CAAC;IAClE,gBAAgB,CAAC,EAAE,MAAM,GAAG,UAAU,GAAG,SAAS,GAAG,IAAI,CAAC;IAC1D,WAAW,CAAC,EAAE,6BAA6B,CAAC,cAAc,CAAC,CAAC;CAC5D;AAED;;GAEG;AACH,eAAO,MAAM,qBAAqB,EAAE,cAAc,CAAC,kBAAkB,EAAE,sBAAsB,CA4E5F,CAAC;AAEF,eAAO,MAAM,2BAA2B,EAAE,cAAc,CAAC,kBAAkB,EAAE,mBAAmB,CAkB/F,CAAC","sourcesContent":["import OpenAI from \"openai\";\nimport type { ResponseCreateParamsStreaming } from \"openai/resources/responses/responses.js\";\nimport { getEnvApiKey } from \"../env-api-keys.js\";\nimport { clampThinkingLevel } from \"../models.js\";\nimport type {\n\tApi,\n\tAssistantMessage,\n\tCacheRetention,\n\tContext,\n\tModel,\n\tOpenAIResponsesCompat,\n\tSimpleStreamOptions,\n\tStreamFunction,\n\tStreamOptions,\n\tUsage,\n} from \"../types.js\";\nimport { AssistantMessageEventStream } from \"../utils/event-stream.js\";\nimport { headersToRecord } from \"../utils/headers.js\";\nimport { isCloudflareProvider, resolveCloudflareBaseUrl } from \"./cloudflare.js\";\nimport { buildCopilotDynamicHeaders, hasCopilotVisionInput } from \"./github-copilot-headers.js\";\nimport { convertResponsesMessages, convertResponsesTools, processResponsesStream } from \"./openai-responses-shared.js\";\nimport { buildBaseOptions } from \"./simple-options.js\";\n\nconst OPENAI_TOOL_CALL_PROVIDERS = new Set([\"openai\", \"openai-codex\", \"opencode\"]);\n\n/**\n * Resolve cache retention preference.\n * Defaults to \"short\" and uses PI_CACHE_RETENTION for backward compatibility.\n */\nfunction resolveCacheRetention(cacheRetention?: CacheRetention): CacheRetention {\n\tif (cacheRetention) {\n\t\treturn cacheRetention;\n\t}\n\tif (typeof process !== \"undefined\" && process.env.PI_CACHE_RETENTION === \"long\") {\n\t\treturn \"long\";\n\t}\n\treturn \"short\";\n}\n\nfunction getCompat(model: Model<\"openai-responses\">): Required<OpenAIResponsesCompat> {\n\treturn {\n\t\tsendSessionIdHeader: model.compat?.sendSessionIdHeader ?? true,\n\t\tsupportsLongCacheRetention: model.compat?.supportsLongCacheRetention ?? true,\n\t};\n}\n\nfunction getPromptCacheRetention(\n\tcompat: Required<OpenAIResponsesCompat>,\n\tcacheRetention: CacheRetention,\n): \"24h\" | undefined {\n\treturn cacheRetention === \"long\" && compat.supportsLongCacheRetention ? \"24h\" : undefined;\n}\n\n// OpenAI Responses-specific options\nexport interface OpenAIResponsesOptions extends StreamOptions {\n\treasoningEffort?: \"minimal\" | \"low\" | \"medium\" | \"high\" | \"xhigh\";\n\treasoningSummary?: \"auto\" | \"detailed\" | \"concise\" | null;\n\tserviceTier?: ResponseCreateParamsStreaming[\"service_tier\"];\n}\n\n/**\n * Generate function for OpenAI Responses API\n */\nexport const streamOpenAIResponses: StreamFunction<\"openai-responses\", OpenAIResponsesOptions> = (\n\tmodel: Model<\"openai-responses\">,\n\tcontext: Context,\n\toptions?: OpenAIResponsesOptions,\n): AssistantMessageEventStream => {\n\tconst stream = new AssistantMessageEventStream();\n\n\t// Start async processing\n\t(async () => {\n\t\tconst output: AssistantMessage = {\n\t\t\trole: \"assistant\",\n\t\t\tcontent: [],\n\t\t\tapi: model.api as Api,\n\t\t\tprovider: model.provider,\n\t\t\tmodel: model.id,\n\t\t\tusage: {\n\t\t\t\tinput: 0,\n\t\t\t\toutput: 0,\n\t\t\t\tcacheRead: 0,\n\t\t\t\tcacheWrite: 0,\n\t\t\t\ttotalTokens: 0,\n\t\t\t\tcost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 },\n\t\t\t},\n\t\t\tstopReason: \"stop\",\n\t\t\ttimestamp: Date.now(),\n\t\t};\n\n\t\ttry {\n\t\t\t// Create OpenAI client\n\t\t\tconst apiKey = options?.apiKey || getEnvApiKey(model.provider) || \"\";\n\t\t\tconst cacheRetention = resolveCacheRetention(options?.cacheRetention);\n\t\t\tconst cacheSessionId = cacheRetention === \"none\" ? undefined : options?.sessionId;\n\t\t\tconst client = createClient(model, context, apiKey, options?.headers, cacheSessionId);\n\t\t\tlet params = buildParams(model, context, options);\n\t\t\tconst nextParams = await options?.onPayload?.(params, model);\n\t\t\tif (nextParams !== undefined) {\n\t\t\t\tparams = nextParams as ResponseCreateParamsStreaming;\n\t\t\t}\n\t\t\tconst requestOptions = {\n\t\t\t\t...(options?.signal ? { signal: options.signal } : {}),\n\t\t\t\t...(options?.timeoutMs !== undefined ? { timeout: options.timeoutMs } : {}),\n\t\t\t\t...(options?.maxRetries !== undefined ? { maxRetries: options.maxRetries } : {}),\n\t\t\t};\n\t\t\tconst { data: openaiStream, response } = await client.responses.create(params, requestOptions).withResponse();\n\t\t\tawait options?.onResponse?.({ status: response.status, headers: headersToRecord(response.headers) }, model);\n\t\t\tstream.push({ type: \"start\", partial: output });\n\n\t\t\tawait processResponsesStream(openaiStream, output, stream, model, {\n\t\t\t\tserviceTier: options?.serviceTier,\n\t\t\t\tapplyServiceTierPricing: (usage, serviceTier) => applyServiceTierPricing(usage, serviceTier, model),\n\t\t\t});\n\n\t\t\tif (options?.signal?.aborted) {\n\t\t\t\tthrow new Error(\"Request was aborted\");\n\t\t\t}\n\n\t\t\tif (output.stopReason === \"aborted\" || output.stopReason === \"error\") {\n\t\t\t\tthrow new Error(\"An unknown error occurred\");\n\t\t\t}\n\n\t\t\tstream.push({ type: \"done\", reason: output.stopReason, message: output });\n\t\t\tstream.end();\n\t\t} catch (error) {\n\t\t\tfor (const block of output.content) {\n\t\t\t\tdelete (block as { index?: number }).index;\n\t\t\t\t// partialJson is only a streaming scratch buffer; never persist it.\n\t\t\t\tdelete (block as { partialJson?: string }).partialJson;\n\t\t\t}\n\t\t\toutput.stopReason = options?.signal?.aborted ? \"aborted\" : \"error\";\n\t\t\toutput.errorMessage = error instanceof Error ? error.message : JSON.stringify(error);\n\t\t\tstream.push({ type: \"error\", reason: output.stopReason, error: output });\n\t\t\tstream.end();\n\t\t}\n\t})();\n\n\treturn stream;\n};\n\nexport const streamSimpleOpenAIResponses: StreamFunction<\"openai-responses\", SimpleStreamOptions> = (\n\tmodel: Model<\"openai-responses\">,\n\tcontext: Context,\n\toptions?: SimpleStreamOptions,\n): AssistantMessageEventStream => {\n\tconst apiKey = options?.apiKey || getEnvApiKey(model.provider);\n\tif (!apiKey) {\n\t\tthrow new Error(`No API key for provider: ${model.provider}`);\n\t}\n\n\tconst base = buildBaseOptions(model, options, apiKey);\n\tconst clampedReasoning = options?.reasoning ? clampThinkingLevel(model, options.reasoning) : undefined;\n\tconst reasoningEffort = clampedReasoning === \"off\" ? undefined : clampedReasoning;\n\n\treturn streamOpenAIResponses(model, context, {\n\t\t...base,\n\t\treasoningEffort,\n\t} satisfies OpenAIResponsesOptions);\n};\n\nfunction createClient(\n\tmodel: Model<\"openai-responses\">,\n\tcontext: Context,\n\tapiKey?: string,\n\toptionsHeaders?: Record<string, string>,\n\tsessionId?: string,\n) {\n\tif (!apiKey) {\n\t\tif (!process.env.OPENAI_API_KEY) {\n\t\t\tthrow new Error(\n\t\t\t\t\"OpenAI API key is required. Set OPENAI_API_KEY environment variable or pass it as an argument.\",\n\t\t\t);\n\t\t}\n\t\tapiKey = process.env.OPENAI_API_KEY;\n\t}\n\n\tconst compat = getCompat(model);\n\tconst headers = { ...model.headers };\n\tif (model.provider === \"github-copilot\") {\n\t\tconst hasImages = hasCopilotVisionInput(context.messages);\n\t\tconst copilotHeaders = buildCopilotDynamicHeaders({\n\t\t\tmessages: context.messages,\n\t\t\thasImages,\n\t\t});\n\t\tObject.assign(headers, copilotHeaders);\n\t}\n\n\tif (sessionId) {\n\t\tif (compat.sendSessionIdHeader) {\n\t\t\theaders.session_id = sessionId;\n\t\t}\n\t\theaders[\"x-client-request-id\"] = sessionId;\n\t}\n\n\t// Merge options headers last so they can override defaults\n\tif (optionsHeaders) {\n\t\tObject.assign(headers, optionsHeaders);\n\t}\n\n\tconst defaultHeaders =\n\t\tmodel.provider === \"cloudflare-ai-gateway\"\n\t\t\t? {\n\t\t\t\t\t...headers,\n\t\t\t\t\tAuthorization: headers.Authorization ?? null,\n\t\t\t\t\t\"cf-aig-authorization\": `Bearer ${apiKey}`,\n\t\t\t\t}\n\t\t\t: headers;\n\n\treturn new OpenAI({\n\t\tapiKey,\n\t\tbaseURL: isCloudflareProvider(model.provider) ? resolveCloudflareBaseUrl(model) : model.baseUrl,\n\t\tdangerouslyAllowBrowser: true,\n\t\tdefaultHeaders,\n\t});\n}\n\nfunction buildParams(model: Model<\"openai-responses\">, context: Context, options?: OpenAIResponsesOptions) {\n\tconst messages = convertResponsesMessages(model, context, OPENAI_TOOL_CALL_PROVIDERS);\n\n\tconst cacheRetention = resolveCacheRetention(options?.cacheRetention);\n\tconst compat = getCompat(model);\n\tconst params: ResponseCreateParamsStreaming = {\n\t\tmodel: model.id,\n\t\tinput: messages,\n\t\tstream: true,\n\t\tprompt_cache_key: cacheRetention === \"none\" ? undefined : options?.sessionId,\n\t\tprompt_cache_retention: getPromptCacheRetention(compat, cacheRetention),\n\t\tstore: false,\n\t};\n\n\tif (options?.maxTokens) {\n\t\tparams.max_output_tokens = options?.maxTokens;\n\t}\n\n\tif (options?.temperature !== undefined) {\n\t\tparams.temperature = options?.temperature;\n\t}\n\n\tif (options?.serviceTier !== undefined) {\n\t\tparams.service_tier = options.serviceTier;\n\t}\n\n\tif (context.tools && context.tools.length > 0) {\n\t\tparams.tools = convertResponsesTools(context.tools);\n\t}\n\n\tif (model.reasoning) {\n\t\tif (options?.reasoningEffort || options?.reasoningSummary) {\n\t\t\tconst effort = options?.reasoningEffort\n\t\t\t\t? (model.thinkingLevelMap?.[options.reasoningEffort] ?? options.reasoningEffort)\n\t\t\t\t: \"medium\";\n\t\t\tparams.reasoning = {\n\t\t\t\teffort: effort as NonNullable<typeof params.reasoning>[\"effort\"],\n\t\t\t\tsummary: options?.reasoningSummary || \"auto\",\n\t\t\t};\n\t\t\tparams.include = [\"reasoning.encrypted_content\"];\n\t\t} else if (model.provider !== \"github-copilot\" && model.thinkingLevelMap?.off !== null) {\n\t\t\tparams.reasoning = {\n\t\t\t\teffort: (model.thinkingLevelMap?.off ?? \"none\") as NonNullable<typeof params.reasoning>[\"effort\"],\n\t\t\t};\n\t\t}\n\t}\n\n\treturn params;\n}\n\nfunction getServiceTierCostMultiplier(\n\tmodel: Pick<Model<\"openai-responses\">, \"id\">,\n\tserviceTier: ResponseCreateParamsStreaming[\"service_tier\"] | undefined,\n): number {\n\tswitch (serviceTier) {\n\t\tcase \"flex\":\n\t\t\treturn 0.5;\n\t\tcase \"priority\":\n\t\t\treturn model.id === \"gpt-5.5\" ? 2.5 : 2;\n\t\tdefault:\n\t\t\treturn 1;\n\t}\n}\n\nfunction applyServiceTierPricing(\n\tusage: Usage,\n\tserviceTier: ResponseCreateParamsStreaming[\"service_tier\"] | undefined,\n\tmodel: Pick<Model<\"openai-responses\">, \"id\">,\n) {\n\tconst multiplier = getServiceTierCostMultiplier(model, serviceTier);\n\tif (multiplier === 1) return;\n\n\tusage.cost.input *= multiplier;\n\tusage.cost.output *= multiplier;\n\tusage.cost.cacheRead *= multiplier;\n\tusage.cost.cacheWrite *= multiplier;\n\tusage.cost.total = usage.cost.input + usage.cost.output + usage.cost.cacheRead + usage.cost.cacheWrite;\n}\n"]}