From da07bdc8c856c6f54eab312e77b3a8eda0040af8 Mon Sep 17 00:00:00 2001
From: Mina Choi <mnchoi@o2o.kr>
Date: Thu, 21 May 2026 08:41:32 +0900
Subject: [PATCH] =?UTF-8?q?feat(report):=20=3Frun=5Fid=3D=20=EC=BF=BC?=
 =?UTF-8?q?=EB=A6=AC=EB=A1=9C=20=EB=A1=9C=EB=94=A9=ED=8E=98=EC=9D=B4?=
 =?UTF-8?q?=EC=A7=80=20resume=20+=20ky=20=ED=83=80=EC=9E=84=EC=95=84?=
 =?UTF-8?q?=EC=9B=83=20=ED=95=B4=EC=A0=9C?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- AnalysisLoadingPage 가 /report/loading?run_id=xxx 쿼리도 path param 과 동일하게
  resume 트리거로 인식. createClinic·startAnalysis 재호출 없이 폴링만 재개.
- startAnalysis 성공 후 replaceState 를 path 가 아닌 ?run_id= 쿼리로 갱신해
  새로고침·공유·뒤로가기 모두 동일한 resume 흐름을 탄다.
- ky 의 60s 글로벌 timeout 을 끔. createClinic 처럼 백엔드가 외부 크롤·LLM 을
  동기로 호출하는 엔드포인트는 60s 안에 끝난다는 보장이 없어 클라이언트에서
  timeout 으로 끊으면 안 됨 (폴링은 어차피 짧음).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 .../report/hooks/useAnalysisPipeline.ts       | 23 ++++++++++++-------
 src/shared/api/api.ts                         |  4 +++-
 2 files changed, 18 insertions(+), 9 deletions(-)

diff --git a/src/features/report/hooks/useAnalysisPipeline.ts b/src/features/report/hooks/useAnalysisPipeline.ts
index d188c3f..0808ba7 100644
--- a/src/features/report/hooks/useAnalysisPipeline.ts
+++ b/src/features/report/hooks/useAnalysisPipeline.ts
@@ -82,6 +82,11 @@ export function useAnalysisPipeline(): UseAnalysisPipelineResult {
   const location = useLocation();
   const { reportId: urlReportId } = useParams<{ reportId?: string }>();
 
+  // ?run_id= 쿼리로 진입하면 createClinic/startAnalysis 건너뛰고 폴링부터 시작.
+  // path 파라미터(/report/loading/:reportId) 와 동일 의미로 취급.
+  const queryRunId = new URLSearchParams(location.search).get('run_id');
+  const resumeRunId = urlReportId || queryRunId || undefined;
+
   const locState = (location.state as { url?: string; manualChannels?: ManualChannels }) ?? {};
   const url = locState.url;
   const manualChannels = locState.manualChannels;
@@ -137,7 +142,9 @@ export function useAnalysisPipeline(): UseAnalysisPipelineResult {
           reportId = runId; // 백엔드: analysis_run_id 가 report 조회 키와 동일
 
           saveSession({ reportId, clinicId, runId, url: startUrl });
-          window.history.replaceState(null, '', `/report/loading/${reportId}`);
+          // 새로고침·공유 시 createClinic/startAnalysis 재호출 없이 폴링만 재개되도록
+          // run_id 를 쿼리 파라미터로 URL 에 박아둔다.
+          window.history.replaceState(null, '', `/report/loading?run_id=${reportId}`);
         }
 
         if (!runId) throw new Error('runId 없음');
@@ -203,17 +210,17 @@ export function useAnalysisPipeline(): UseAnalysisPipelineResult {
     if (hasStarted.current) return;
     hasStarted.current = true;
 
-    // 1. URL 파라미터 resume (예: /report/loading/abc-123)
-    if (urlReportId) {
+    // 1. URL 파라미터/쿼리 resume (예: /report/loading/abc-123 또는 /report/loading?run_id=abc-123)
+    if (resumeRunId) {
       setPhase('resuming');
       const session = loadSession();
-      const runId = session.runId || urlReportId;
+      const runId = session.runId || resumeRunId;
       getAnalysisStatus(runId)
         .then((res) => {
           if (res.status !== 200) throw new Error('status fetch failed');
           const status = res.data;
           if (status.status === AnalysisStatus.completed) {
-            navigate(`/report/${urlReportId}`, { replace: true });
+            navigate(`/report/${resumeRunId}`, { replace: true });
             return;
           }
           if (status.status === AnalysisStatus.failed) {
@@ -222,13 +229,13 @@ export function useAnalysisPipeline(): UseAnalysisPipelineResult {
             return;
           }
           saveSession({
-            reportId: urlReportId,
+            reportId: resumeRunId,
             clinicId: session.clinicId || undefined,
             runId,
             url: session.url || undefined,
           });
           runPipeline(undefined, {
-            reportId: urlReportId,
+            reportId: resumeRunId,
             clinicId: session.clinicId || undefined,
             runId,
           });
@@ -274,7 +281,7 @@ export function useAnalysisPipeline(): UseAnalysisPipelineResult {
 
     // 4. 신규 분석 시작
     runPipeline(url);
-  }, [url, urlReportId, navigate, runPipeline]);
+  }, [url, resumeRunId, navigate, runPipeline]);
 
   return {
     phase,
diff --git a/src/shared/api/api.ts b/src/shared/api/api.ts
index fd77738..1f2ffd0 100644
--- a/src/shared/api/api.ts
+++ b/src/shared/api/api.ts
@@ -14,7 +14,9 @@ import ky, { type KyInstance } from 'ky'
 const API_BASE_URL = (__VITE_API_BASE_URL__ ?? '').replace(/\/$/, '')
 
 export const kyInstance: KyInstance = ky.create({
-  timeout: 60_000,
+  // createClinic·startAnalysis 처럼 백엔드가 외부 크롤·LLM 을 호출하는 동기 엔드포인트는
+  // 60s 안에 끝난다는 보장이 없어 클라이언트 타임아웃은 끈다. 폴링 호출은 어차피 짧다.
+  timeout: false,
   retry: 1,
   // orval generated 타입이 4xx/5xx 응답도 data로 받아오므로 throw 비활성
   throwHttpErrors: false,