@cartanova/qgrid-ai-sdk 2.3.2 → 2.3.4

package/dist/index.d.ts

@@ -5,22 +5,30 @@ import { TelemetrySettings } from "ai";
 type QgridProviderConfig = {
   serverUrl?: string;
   defaultEffort?: string;
+  /** request_logs.project_name. default: process.env.QGRID_PROJECT_NAME */
+  projectName?: string;
 };
 /**
- * codex app-server가 지원하는 OpenAI provider options.
- *
- * AI SDK의 openai provider options 중 codex가 처리할 수 있는 subset만 정의.
- * 여기에 없는 옵션(temperature, maxOutputTokens 등)은 codex가 무시합니다.
- *
- * @see https://ai-sdk.dev/providers/ai-sdk-providers/openai#provider-options
+ * qgrid provider options.
+ * providerOptions.qgrid 로 전달한다.
  */
-type QgridOpenAIProviderOptions = {
+type QgridProviderOptions = {
+  /**
+   * 멀티턴 시 codex thread reuse를 위한 대화 식별자, 호출자가 자기 도메인 ID(예: 게임 세션 ID) 하나만 넘기면
+   * provider가 같은 sessionKey 의 thread 좌표를 내부에서 회송해 prompt cache 를 적중시킨다
+   * 좌표는 sessionKey 별로 격리된다.
+   */
+  sessionKey?: string;
   /** reasoning 모델의 추론 깊이. 기본값은 qgrid config의 defaultEffort. */
-  reasoningEffort?: "none" | "minimal" | "low" | "medium" | "high" | "xhigh";
-  /** reasoning 모델의 추론 요약 출력 방식. Responses API 전용. */
+  effort?: "none" | "minimal" | "low" | "medium" | "high" | "xhigh";
+  /** 응답 텍스트의 상세도. OpenAI/codex route에서만 적용된다. */
+  verbosity?: "low" | "medium" | "high";
+  /** reasoning 모델의 추론 요약 출력 방식. OpenAI/codex route에서만 적용된다. */
   reasoningSummary?: "auto" | "concise" | "detailed" | "none";
-  /** 응답 텍스트의 상세도. */
-  textVerbosity?: "low" | "medium" | "high";
+  /** OpenAI/codex service tier. OpenAI/codex route에서만 적용된다. */
+  serviceTier?: string;
+  /** @todo 향후 qgrid 서버 fallback routing에 사용할 후보 모델 목록. */
+  fallbackModels?: string[];
 };
 /**
  * codex thread 좌표
@@ -28,18 +36,6 @@ type QgridOpenAIProviderOptions = {
  * provider 내부 threadCoordStore 가 sessionKey 별로 관리
  */
 
-/**
- * {@link QgridOpenAIProviderOptions}
- */
-type QgridProviderOptions = {
-  openai?: QgridOpenAIProviderOptions;
-  /**
-   * 멀티턴 대화 식별자. 호출자가 자기 도메인 ID(예: 게임 세션 ID) 하나만 넘기면, provider 가
-   * 같은 sessionKey 의 thread 좌표를 내부에서 회송해 prompt cache 를 적중시킨다.
-   * 좌표는 sessionKey 별로 격리된다.
-   */
-  sessionKey?: string;
-};
 type QgridSupportedModel = "openai/gpt-5.5" | "openai/gpt-5.4" | "openai/gpt-5.2" | "openai/gpt-5.4-mini" | "openai/gpt-5.3-codex" | "openai/gpt-5.3-codex-spark" | "anthropic/claude-haiku-4-5" | "anthropic/claude-sonnet-4" | "anthropic/claude-sonnet-4-5" | "anthropic/claude-sonnet-4-6" | "anthropic/claude-sonnet-4-7" | "anthropic/claude-opus-4" | "anthropic/claude-opus-4-1" | "anthropic/claude-opus-4-5" | "anthropic/claude-opus-4-6" | "anthropic/claude-opus-4-7" | "anthropic/claude-opus-4-8";
 type QgridLoggerConfig = {
   /** qgrid 서버 주소. default: process.env.QGRID_URL, 없으면 기본값은 "http://localhost:44900" */

package/dist/index.js

@@ -1,3 +1,10 @@
+//#region src/qgrid.constant.ts
+const QGRID_PROVIDER_NAME = "qgrid";
+const THREAD_COORD_TTL_MS = 10 * 6e4;
+const TEXT_STREAM_ID_PREFIX = "text";
+const TOP_LEVEL_SCHEMA_WARNING = "OpenAI structured output requires \"object\". Falling back to client-side parsing.";
+const PENDING_TOOL_RESULTS_WARNING = "[qgrid] pending tool results not found in prompt, clearing client run state";
+//#endregion
 //#region src/utils.ts
 async function createRun(serverUrl, body) {
 	return (await fetch(`${serverUrl}/api/qgrid/createRun`, {
@@ -240,7 +247,6 @@ function timedKeySet() {
 function createQgridLogger(config = {}) {
 	const serverUrl = config.serverUrl ?? process.env.QGRID_URL ?? "http://localhost:44900";
 	const projectName = config.projectName ?? process.env.QGRID_PROJECT_NAME;
-	config.tokenName ?? process.env.QGRID_TOKEN_NAME;
 	const onLogError = config.onLogError ?? ((e) => console.warn(`[qgrid-logger] ${e.message}`));
 	const runs = /* @__PURE__ */ new Map();
 	const keyTtl = typeof config.staleRunTimeoutMs === "number" && config.staleRunTimeoutMs > 0 ? config.staleRunTimeoutMs : DEFAULT_STALE_RUN_TIMEOUT_MS;
@@ -459,9 +465,6 @@ function createQgridLogger(config = {}) {
 }
 //#endregion
 //#region src/index.ts
-const DEFAULT_SERVER_URL = "http://localhost:44900";
-const DEFAULT_EFFORT = "low";
-const THREAD_COORD_TTL_MS = 10 * 6e4;
 const threadCoordStore = /* @__PURE__ */ new Map();
 function getThreadCoord(sessionKey) {
 	const entry = threadCoordStore.get(sessionKey);
@@ -478,6 +481,12 @@ function setThreadCoord(sessionKey, coord) {
 		expiresAt: Date.now() + THREAD_COORD_TTL_MS
 	});
 }
+function reusableSessionKey(modelId, sessionKey) {
+	return modelId.startsWith("anthropic/") ? void 0 : sessionKey;
+}
+function getQgridProviderOptions(options) {
+	return options.providerOptions?.qgrid ?? {};
+}
 function toAiSdkUsage(usage) {
 	const cacheRead = usage.cache_read_input_tokens;
 	const cacheWrite = usage.cache_creation_input_tokens;
@@ -496,26 +505,29 @@ function toAiSdkUsage(usage) {
 	};
 }
 function qgrid(modelId, config) {
-	const serverUrl = config?.serverUrl ?? process.env.QGRID_URL ?? DEFAULT_SERVER_URL;
-	const effort = config?.defaultEffort ?? DEFAULT_EFFORT;
+	const serverUrl = config?.serverUrl ?? process.env.QGRID_URL ?? "http://localhost:44900";
+	const effort = config?.defaultEffort ?? "low";
+	const projectName = config?.projectName ?? process.env.QGRID_PROJECT_NAME;
 	let clientRun = null;
 	return {
 		specificationVersion: "v3",
-		provider: "qgrid",
+		provider: QGRID_PROVIDER_NAME,
 		modelId,
 		supportedUrls: {},
 		async doGenerate(options) {
 			const tools = options.tools?.filter((t) => t.type === "function");
 			const hasTools = tools && tools.length > 0;
 			const { prompt, system, history } = extractPromptAndHistory(options.prompt);
-			const openaiOpts = options.providerOptions?.openai;
-			const effectiveEffort = openaiOpts?.reasoningEffort ?? effort;
-			const verbosity = openaiOpts?.textVerbosity ?? openaiOpts?.verbosity;
-			const reasoningSummary = openaiOpts?.reasoningSummary;
-			const sessionKey = (options.providerOptions?.qgrid)?.sessionKey;
+			const qgridOptions = getQgridProviderOptions(options);
+			const effectiveEffort = qgridOptions.effort ?? effort;
+			const verbosity = qgridOptions.verbosity;
+			const reasoningSummary = qgridOptions.reasoningSummary;
+			const serviceTier = qgridOptions.serviceTier;
+			const sessionKey = qgridOptions.sessionKey;
+			const reuseSessionKey = reusableSessionKey(modelId, sessionKey);
 			const rawSchema = options.responseFormat?.type === "json" ? options.responseFormat.schema : void 0;
 			const schemaType = rawSchema ? rawSchema.type : void 0;
-			if (rawSchema && !hasTools && schemaType !== "object") console.warn(`[qgrid] responseFormat.schema top-level type is "${schemaType ?? "unknown"}". OpenAI structured output requires "object". Falling back to client-side parsing.`);
+			if (rawSchema && !hasTools && schemaType !== "object") console.warn(`[qgrid] responseFormat.schema top-level type is "${schemaType ?? "unknown"}". ${TOP_LEVEL_SCHEMA_WARNING}`);
 			const jsonSchema = !hasTools && rawSchema && schemaType === "object" ? JSON.stringify(rawSchema) : void 0;
 			let runContext;
 			let toolResultsPayload;
@@ -531,11 +543,11 @@ function qgrid(modelId, config) {
 					}));
 					logMode = "run";
 				} else {
-					console.warn("[qgrid] pending tool results not found in prompt, clearing client run state");
+					console.warn(PENDING_TOOL_RESULTS_WARNING);
 					clientRun = null;
 				}
 			}
-			const storedCoord = sessionKey ? getThreadCoord(sessionKey) : void 0;
+			const storedCoord = reuseSessionKey ? getThreadCoord(reuseSessionKey) : void 0;
 			if (!runContext && storedCoord) runContext = { threadCoord: storedCoord };
 			if (!logMode && hasTools) logMode = "run";
 			const data = await fetch(`${serverUrl}/api/qgrid/query`, {
@@ -548,9 +560,11 @@ function qgrid(modelId, config) {
 					effort: effectiveEffort,
 					...verbosity ? { verbosity } : {},
 					...reasoningSummary ? { reasoningSummary } : {},
+					...serviceTier ? { serviceTier } : {},
 					...hasTools ? { tools: tools.map(toQgridTool) } : {},
 					...jsonSchema ? { jsonSchema } : {},
 					...history.length > 0 ? { history: JSON.stringify(history) } : {},
+					...projectName ? { projectName } : {},
 					...logMode ? { logMode } : {},
 					...runContext ? { runContext } : {},
 					...toolResultsPayload ? { toolResults: toolResultsPayload } : {}
@@ -592,7 +606,7 @@ function qgrid(modelId, config) {
 				pendingToolCallIds: new Set(content.filter((c) => c.type === "tool-call").map((c) => c.toolCallId))
 			};
 			else clientRun = null;
-			if (sessionKey && data.runContext?.threadCoord) setThreadCoord(sessionKey, data.runContext.threadCoord);
+			if (reuseSessionKey && data.runContext?.threadCoord) setThreadCoord(reuseSessionKey, data.runContext.threadCoord);
 			return {
 				content,
 				finishReason,
@@ -611,11 +625,13 @@ function qgrid(modelId, config) {
 			const tools = options.tools?.filter((t) => t.type === "function");
 			const hasTools = tools && tools.length > 0;
 			const { prompt, system, history } = extractPromptAndHistory(options.prompt);
-			const openaiOpts = options.providerOptions?.openai;
-			const effectiveEffort = openaiOpts?.reasoningEffort ?? effort;
-			const verbosity = openaiOpts?.textVerbosity ?? openaiOpts?.verbosity;
-			const reasoningSummary = openaiOpts?.reasoningSummary;
-			const sessionKey = (options.providerOptions?.qgrid)?.sessionKey;
+			const qgridOptions = getQgridProviderOptions(options);
+			const effectiveEffort = qgridOptions.effort ?? effort;
+			const verbosity = qgridOptions.verbosity;
+			const reasoningSummary = qgridOptions.reasoningSummary;
+			const serviceTier = qgridOptions.serviceTier;
+			const sessionKey = qgridOptions.sessionKey;
+			const reuseSessionKey = reusableSessionKey(modelId, sessionKey);
 			const rawSchema = options.responseFormat?.type === "json" ? options.responseFormat.schema : void 0;
 			const jsonSchema = !hasTools && rawSchema && rawSchema.type === "object" ? JSON.stringify(rawSchema) : void 0;
 			let runContext;
@@ -632,11 +648,11 @@ function qgrid(modelId, config) {
 					}));
 					logMode = "run";
 				} else {
-					console.warn("[qgrid] pending tool results not found in prompt, clearing client run state");
+					console.warn(PENDING_TOOL_RESULTS_WARNING);
 					clientRun = null;
 				}
 			}
-			const storedCoord = sessionKey ? getThreadCoord(sessionKey) : void 0;
+			const storedCoord = reuseSessionKey ? getThreadCoord(reuseSessionKey) : void 0;
 			if (!runContext && storedCoord) runContext = { threadCoord: storedCoord };
 			if (!logMode && hasTools) logMode = "run";
 			const prepRes = await fetch(`${serverUrl}/api/qgrid/prepareStream`, {
@@ -649,9 +665,11 @@ function qgrid(modelId, config) {
 					effort: effectiveEffort,
 					...verbosity ? { verbosity } : {},
 					...reasoningSummary ? { reasoningSummary } : {},
+					...serviceTier ? { serviceTier } : {},
 					...hasTools ? { tools: tools.map(toQgridTool) } : {},
 					...jsonSchema ? { jsonSchema } : {},
 					...history.length > 0 ? { history: JSON.stringify(history) } : {},
+					...projectName ? { projectName } : {},
 					...logMode ? { logMode } : {},
 					...runContext ? { runContext } : {},
 					...toolResultsPayload ? { toolResults: toolResultsPayload } : {}
@@ -668,7 +686,7 @@ function qgrid(modelId, config) {
 				const text = await streamRes.text().catch(() => "");
 				throw new Error(`qgrid stream ${streamRes.status}: ${text}`);
 			}
-			const textId = `text_${Math.random().toString(36).slice(2, 10)}`;
+			const textId = `${TEXT_STREAM_ID_PREFIX}_${Math.random().toString(36).slice(2, 10)}`;
 			let textStarted = false;
 			let deltaTextEmitted = false;
 			return { stream: new ReadableStream({ async start(controller) {
@@ -704,10 +722,10 @@ function qgrid(modelId, config) {
 							pendingToolCallIds: new Set((done.content ?? []).filter((c) => c.type === "tool-call").map((c) => c.toolCallId))
 						};
 						else clientRun = null;
-						if (sessionKey && done.runContext?.threadCoord) setThreadCoord(sessionKey, done.runContext.threadCoord);
+						if (reuseSessionKey && done.runContext?.threadCoord) setThreadCoord(reuseSessionKey, done.runContext.threadCoord);
 						if (done.content) {
 							for (const item of done.content) if (item.type === "text" && !deltaTextEmitted) {
-								const tid = `text_${Math.random().toString(36).slice(2, 10)}`;
+								const tid = `${TEXT_STREAM_ID_PREFIX}_${Math.random().toString(36).slice(2, 10)}`;
 								controller.enqueue({
 									type: "text-start",
 									id: tid

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cartanova/qgrid-ai-sdk",
-  "version": "2.3.2",
+  "version": "2.3.4",
   "description": "AI SDK LanguageModelV3 provider for qgrid",
   "repository": {
     "type": "git",