Upload main.py

main.py

@@ -209,10 +209,19 @@ def proxy_img(url: str = Query(...)):
     except: return Response(status_code=502)
 
 # ===== XEMLAIBONGDA HIGHLIGHTS =====
+# Thread-local session for connection pooling
+_thread_local = threading.local()
+def _get_session():
+    if not hasattr(_thread_local, 'session'):
+        _thread_local.session = requests.Session()
+        _thread_local.session.headers.update(HEADERS)
+    return _thread_local.session
+
 def _fetch_video_thumbnail(url):
     """Fetch og:image from a video detail page."""
     try:
-        r = requests.get(url, headers=HEADERS, timeout=8)
+        s = _get_session()
+        r = s.get(url, timeout=6)
         if r.status_code != 200:
             return ""
         r.encoding = "utf-8"
@@ -245,9 +254,16 @@ def _fetch_video_thumbnail(url):
         pass
     return ""
 
+def _proxy_img_url(src):
+    """Proxy image URL through our server if needed."""
+    if src and ("xemlaibongda.top" in src or "cdnphoto.dantri" in src):
+        return '/api/proxy/img?url=' + quote(src, safe='')
+    return src
+
 def _scrape_xemlaibongda_page(page_path, limit=20):
     """
     Scrape video từ xemlaibongda.top - with proper image extraction
+    Uses parallel fetching for thumbnails to avoid timeout
     """
     try:
         url = f"https://xemlaibongda.top/{page_path}" if page_path else "https://xemlaibongda.top/"
@@ -331,24 +347,36 @@ def _scrape_xemlaibongda_page(page_path, limit=20):
             if not title or len(title) < 3:
                 continue
             
-            # ===== FIX: Fetch thumbnail from page if not found =====
-            if not img_src:
-                img_src = _fetch_video_thumbnail(clean_href)
-            
-            # Proxy images through our server to avoid referer/CORS issues
-            if img_src and "xemlaibongda.top" in img_src:
-                img_src = '/api/proxy/img?url=' + quote(img_src, safe='')
+            # Proxy image if needed
+            if img_src:
+                img_src = _proxy_img_url(img_src)
             
             videos.append({
                 "title": title[:100],
                 "link": clean_href,
                 "img": img_src,
-                "source": "xemlaibongda"
+                "source": "xemlaibongda",
+                "_needs_thumb": not bool(img_src)  # flag for later fetch
             })
             
             if len(videos) >= limit:
                 break
         
+        # Fetch missing thumbnails in parallel (max 5 at a time to avoid timeout)
+        missing = [v for v in videos if v.get("_needs_thumb")]
+        if missing:
+            def _fetch_thumb(v):
+                thumb = _fetch_video_thumbnail(v["link"])
+                if thumb:
+                    v["img"] = _proxy_img_url(thumb)
+                v.pop("_needs_thumb", None)
+            
+            with ThreadPoolExecutor(5) as ex:
+                list(ex.map(_fetch_thumb, missing))
+        else:
+            for v in videos:
+                v.pop("_needs_thumb", None)
+        
         return videos
     except Exception as e:
         print(f"[xemlaibongda] Error: {e}")
@@ -658,17 +686,14 @@ def api_video_url(url:str=Query(...)):
     return JSONResponse({"error":"not found"})
 
 # ===== WORLD CUP 2026 API =====
-# Rate limiting cho WC API
 _wc_request_times = []
 _wc_rate_limit_lock = threading.Lock()
-_WC_RATE_LIMIT = 10  # Max 10 requests per minute
+_WC_RATE_LIMIT = 10
 
 def _wc_rate_limit():
-    """Kiểm tra rate limit cho WC API"""
     global _wc_request_times
     with _wc_rate_limit_lock:
         now = time.time()
-        # Xóa các request cũ hơn 60 giây
         _wc_request_times = [t for t in _wc_request_times if now - t < 60]
         if len(_wc_request_times) >= _WC_RATE_LIMIT:
             return False
@@ -677,12 +702,10 @@ def _wc_rate_limit():
 
 @app.get("/api/wc2026")
 def api_wc2026():
-    """Trả về tất cả dữ liệu World Cup 2026"""
     return JSONResponse(_cached("wc2026", get_wc2026_all, ttl=_cache_ttl))
 
 @app.get("/api/wc2026/{tab}")
 def api_wc2026_tab(tab: str):
-    """Trả về từng tab của World Cup"""
     valid_tabs = ["news", "fixtures", "standings", "stats", "highlights"]
     if tab not in valid_tabs:
         return JSONResponse({"error": "invalid tab"}, status_code=400)
@@ -829,4 +852,4 @@ def api_hot_topics():
 @app.get("/", response_class=HTMLResponse)
 async def root():
     return HTMLResponse("<h1>VNEWS</h1><p>Running</p>")
-# v15 rebuild 2026-06-27 - fix highlight images + AI rewrite line-by-line
+# v16 rebuild 2026-06-27 - parallel thumbnail fetch + AI rewrite per line