Update app/fetch.py

app/fetch.py

@@ -1,14 +1,14 @@
 import random
 import json
+import asyncio
 from fastapi import HTTPException
-from playwright.async_api import async_playwright
+from playwright.async_api import async_playwright, TimeoutError
 
 
 UAS = [
     # Windows Chrome
     "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
     "(KHTML, like Gecko) Chrome/121.0.0.0 Safari/537.36",
-
     "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
     "(KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36",
 
@@ -40,11 +40,11 @@ class Fetcher:
         if self.pw:
             await self.pw.stop()
 
-    async def fetch(self, url: str):
-        ua = random.choice(UAS)
+    # =============================
 
-        context = await self.browser.new_context(
-            user_agent=ua,
+    async def _new_context(self):
+        return await self.browser.new_context(
+            user_agent=random.choice(UAS),
             locale="en-US",
             timezone_id="Asia/Ho_Chi_Minh",
             extra_http_headers={
@@ -53,17 +53,20 @@ class Fetcher:
             },
         )
 
+    async def fetch(self, url: str, retry: int = 1):
+        context = await self._new_context()
         page = await context.new_page()
 
         try:
-            # warm homepage (bắt buộc với ihentai.dog)
+            # ===== CF WARM UP =====
             await page.goto(
                 "https://ihentai.dog/",
                 wait_until="domcontentloaded",
                 timeout=15000,
             )
+            await page.wait_for_timeout(1200)  # CF settle
 
-            # load API như browser thật
+            # ===== REAL NAVIGATION =====
             resp = await page.goto(
                 url,
                 wait_until="domcontentloaded",
@@ -74,15 +77,38 @@ class Fetcher:
                 raise HTTPException(500, "No response")
 
             status = resp.status
-            text = await resp.text()
+            headers = resp.headers
+            body = await resp.text()
+
+            # ===== STATUS HANDLE =====
+            if status >= 400:
+                raise HTTPException(status, "Upstream error")
 
-            if status != 200:
-                raise HTTPException(status, "Fetch failed")
+            if not body:
+                return {
+                    "status": status,
+                    "headers": headers,
+                    "data": None,
+                }
 
+            # ===== JSON SAFE PARSE =====
             try:
-                return json.loads(text)
+                data = json.loads(body)
             except Exception:
-                return {"raw": text}
+                data = body
+
+            return {
+                "status": status,
+                "headers": headers,
+                "data": data,
+            }
+
+        except TimeoutError:
+            if retry > 0:
+                await page.close()
+                await context.close()
+                return await self.fetch(url, retry=retry - 1)
+            raise HTTPException(504, "Timeout")
 
         finally:
             await page.close()