Upload main.py

main.py

@@ -202,17 +202,52 @@ def proxy_video(url: str = Query(...), request: Request = None):
 @app.get("/api/proxy/img")
 def proxy_img(url: str = Query(...)):
     try:
-        r = requests.get(url, headers={**HEADERS, "Referer": "https://dantri.com.vn/"}, timeout=10)
+        r = requests.get(url, headers={**HEADERS, "Referer": "https://xemlaibongda.top/"}, timeout=10)
         if r.status_code != 200: return Response(status_code=502)
         ct = r.headers.get("Content-Type", "image/jpeg")
         return Response(content=r.content, media_type=ct, headers={"Cache-Control": "public, max-age=86400", "Access-Control-Allow-Origin": "*"})
     except: return Response(status_code=502)
 
 # ===== XEMLAIBONGDA HIGHLIGHTS =====
+def _fetch_video_thumbnail(url):
+    """Fetch og:image from a video detail page."""
+    try:
+        r = requests.get(url, headers=HEADERS, timeout=8)
+        if r.status_code != 200:
+            return ""
+        r.encoding = "utf-8"
+        soup = BeautifulSoup(r.text, "lxml")
+        # Try og:image
+        og = soup.find("meta", property="og:image")
+        if og:
+            src = og.get("content", "")
+            if src:
+                if src.startswith("//"): src = "https:" + src
+                elif src.startswith("/"): src = "https://xemlaibongda.top" + src
+                return src
+        # Try twitter:image
+        tw = soup.find("meta", attrs={"name": "twitter:image"})
+        if tw:
+            src = tw.get("content", "")
+            if src:
+                if src.startswith("//"): src = "https:" + src
+                elif src.startswith("/"): src = "https://xemlaibongda.top" + src
+                return src
+        # Try first significant img
+        for img in soup.find_all("img"):
+            src = img.get("src", "") or img.get("data-src", "")
+            if src and "logo" not in src.lower() and "avatar" not in src.lower() and "icon" not in src.lower():
+                if src.startswith("//"): src = "https:" + src
+                elif src.startswith("/"): src = "https://xemlaibongda.top" + src
+                if src.startswith("http"):
+                    return src
+    except:
+        pass
+    return ""
+
 def _scrape_xemlaibongda_page(page_path, limit=20):
     """
-    Scrape video từ xemlaibongda.top - Simple & Reliable
-    Dùng logic cũ đã test, không fetch từng trang (tránh timeout)
+    Scrape video từ xemlaibongda.top - with proper image extraction
     """
     try:
         url = f"https://xemlaibongda.top/{page_path}" if page_path else "https://xemlaibongda.top/"
@@ -296,10 +331,13 @@ def _scrape_xemlaibongda_page(page_path, limit=20):
             if not title or len(title) < 3:
                 continue
             
-            # Fallback thumbnail từ slug
+            # ===== FIX: Fetch thumbnail from page if not found =====
             if not img_src:
-                slug = clean_href.split("/video/")[-1].rstrip("/").split("/xem-lai/")[-1].rstrip("/")
-                img_src = f"https://xemlaibongda.top/uploads/thumb/{slug}.jpg"
+                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='')
             
             videos.append({
                 "title": title[:100],
@@ -664,9 +702,6 @@ def api_wc2026_tab(tab: str):
     
     return JSONResponse(_cached(f"wc2026_{tab}", _fetch_tab, ttl=_cache_ttl))
 
-# Note: WC functions (scrape_wc_news, scrape_fixtures, scrape_stats, scrape_standings) 
-# are imported from wc2026_scraper.py at the top of this file
-
 @app.get("/api/video_url")
 def api_video_url(url:str=Query(...)):
     if "youtube.com" in url or "youtu.be" in url:
@@ -794,4 +829,4 @@ def api_hot_topics():
 @app.get("/", response_class=HTMLResponse)
 async def root():
     return HTMLResponse("<h1>VNEWS</h1><p>Running</p>")
-# v14 rebuild 2026-06-18T12:21:41.503230
+# v15 rebuild 2026-06-27 - fix highlight images + AI rewrite line-by-line