cf逻辑复用

app/services/surugaya_client.py

@@ -91,21 +91,7 @@ class SurugayaClient:
         logger.debug("开始抓取搜索页：payload=%s", json.dumps(payload.model_dump(), ensure_ascii=False))
         url = str(payload.search_url) if payload.search_url else self._build_search_url(payload)
 
-        session = await self._session_manager.get_verified_session(url)
-        if session.html is not None:
-            # 新创建的会话已携带页面 HTML，直接使用，避免二次加载
-            html = session.html
-        else:
-            async with self._browser_pool.page_session(storage_state=session.storage_state) as (_, page):
-                response = await page.goto(
-                    url,
-                    wait_until="domcontentloaded",
-                    timeout=int(self._settings.request_timeout_seconds * 1000),
-                )
-                status = response.status if response is not None else None
-                if status != 200:
-                    await self._raise_upstream_error(page=page, response=response, url=url)
-                html = await page.content()
+        html, session = await self._session_manager.fetch_html(url)
 
         tree = HTMLParser(html)
 
@@ -165,21 +151,8 @@ class SurugayaClient:
         logger.info("开始抓取详情页：goods_id=%s", goods_id)
         logger.debug("详情页URL：url=%s", url)
 
-        # 抓取页面 HTML
-        session = await self._session_manager.get_verified_session(url)
-        if session.html is not None:
-            html = session.html
-        else:
-            async with self._browser_pool.page_session(storage_state=session.storage_state) as (_, page):
-                response = await page.goto(
-                    url,
-                    wait_until="domcontentloaded",
-                    timeout=int(self._settings.request_timeout_seconds * 1000),
-                )
-                status = response.status if response is not None else None
-                if status != 200:
-                    await self._raise_upstream_error(page=page, response=response, url=url)
-                html = await page.content()
+        # 抓取页面 HTML（会话复用失败时由 fetch_html 内部自动重建，避免成功/失败交替）
+        html, _ = await self._session_manager.fetch_html(url)
 
         tree = HTMLParser(html)