Restore from 8a81274 + apply: shorts @baodantri7941+@baosuckhoedoisongboyte, genk AI 30 articles+OG fallback, homepage genk_ai replaces cong-nghe

main.py

@@ -1 +1,512 @@
-Tôi cần thay thế hàm _yt_playlist và scrape_fpt_shorts
+"""VNEWS - FastAPI backend with livescore + xemlaibongda highlights + YouTube FPT shorts"""
+import hashlib, re, time, subprocess, json
+from datetime import datetime
+from concurrent.futures import ThreadPoolExecutor, as_completed
+from fastapi import FastAPI, Query, Request
+from fastapi.responses import HTMLResponse, JSONResponse, StreamingResponse, Response
+from fastapi.staticfiles import StaticFiles
+from urllib.parse import unquote, quote, urlencode
+import requests
+from bs4 import BeautifulSoup
+
+app = FastAPI()
+HEADERS = {"User-Agent":"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36","Accept-Language":"vi-VN,vi;q=0.9,en;q=0.8"}
+BONGDA_HEADERS = {"User-Agent":"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36","Accept-Language":"vi-VN,vi;q=0.9","Referer":"https://bongda.com.vn/lich-thi-dau","X-Requested-With":"XMLHttpRequest"}
+BASE_BDP = "https://bongdaplus.vn"
+SPACE_URL = "https://bep40-vnews.hf.space"
+_cache = {}
+_cache_ttl = 300
+_cache_ttl_live = 60
+_cache_ttl_yt = 1800
+PRIORITY_LEAGUES = ["Ngoại Hạng Anh","FA Cup","Champions League","LaLiga","Copa del Rey","Serie A","Bundesliga","Ligue 1","V-League"]
+LEAGUE_IDS = {"nha":27110,"laliga":27233,"seriea":27044,"bundesliga":26891,"ligue1":27212}
+HL_LEAGUES = {"premier-league":{"path":"anh/premier-league","name":"Premier League","emoji":"🏴󠁧󠁢󠁥󠁮󠁧󠁿"},"fa-cup":{"path":"anh/fa-cup","name":"FA Cup","emoji":"🏆"},"bundesliga":{"path":"duc/bundesliga","name":"Bundesliga","emoji":"🇩🇪"},"serie-a":{"path":"italy/serie-a","name":"Serie A","emoji":"🇮🇹"},"la-liga":{"path":"tay-ban-nha/la-liga","name":"La Liga","emoji":"🇪🇸"},"champions-league":{"path":"cup-chau-au/uefa-champions-league","name":"Champions League","emoji":"⭐"},"europa-league":{"path":"cup-chau-au/uefa-europa-league","name":"Europa League","emoji":"🟠"},"world-cup":{"path":"the-gioi/world-cup-qualifiers","name":"World Cup 2026","emoji":"🌍"}}
+def _cached(key, fn, ttl=None):
+    now=time.time();t=ttl or _cache_ttl
+    if key in _cache and now-_cache[key]["t"]<t:return _cache[key]["d"]
+    try:data=fn()
+    except:data=_cache.get(key,{}).get("d",[])
+    _cache[key]={"d":data,"t":now};return data
+def _get(url,headers=None):
+    h=headers or HEADERS;r=requests.get(url,headers=h,timeout=15);r.encoding="utf-8"
+    return BeautifulSoup(r.text,"lxml")
+def fetch_bongda_api(endpoint):
+    try:
+        r=requests.get(f"https://bongda.com.vn{endpoint}",headers=BONGDA_HEADERS,timeout=10)
+        if r.status_code==200:
+            data=r.json()
+            if data.get("status")=="success":return data.get("html","")
+        return ""
+    except:return ""
+def _parse_match_from_li(li, status_type="live"):
+    match_div=li.select_one("div.match")
+    if not match_div:return None
+    home_el=match_div.select_one(".home-team .name");away_el=match_div.select_one(".away-team .name")
+    if not home_el or not away_el:return None
+    status_el=match_div.select_one(".status a");league_el=li.find_previous("strong");time_el=match_div.select_one(".match-time")
+    home_logo=match_div.select_one(".home-team .logo img");away_logo=match_div.select_one(".away-team .logo img")
+    event_id=""
+    if status_el:
+        href=status_el.get("href","");m=re.search(r'/tran-dau/(\d+)/',href)
+        if m:event_id=m.group(1)
+    spans=status_el.find_all("span") if status_el else [];score="";minute=""
+    if len(spans)>=3:score=f"{spans[0].get_text(strip=True)} - {spans[2].get_text(strip=True)}"
+    if len(spans)>=4:minute=spans[3].get_text(strip=True)
+    if not score and status_el and status_el.select_one(".vs"):score="VS"
+    league=league_el.get_text(strip=True) if league_el else ""
+    return{"home":home_el.get_text(strip=True),"away":away_el.get_text(strip=True),"score":score or"VS","minute":minute,"league":league,"time":time_el.get_text(strip=True) if time_el else "","event_id":event_id,"home_logo":home_logo.get("src","") if home_logo else "","away_logo":away_logo.get("src","") if away_logo else "","status":status_type}
+
+# ===== VIDEO PROXY =====
+@app.get("/api/proxy/m3u8")
+def proxy_m3u8(url: str = Query(...)):
+    try:
+        r = requests.get(url, headers=HEADERS, timeout=15)
+        if r.status_code != 200:return Response(status_code=502, content="upstream error")
+        lines = r.text.strip().split('\n');rewritten = []
+        for line in lines:
+            if line.startswith('#') or not line.strip():rewritten.append(line)
+            else:rewritten.append("/api/proxy/seg?url=" + quote(line.strip(), safe=""))
+        return Response(content='\n'.join(rewritten).encode('utf-8'),media_type="application/vnd.apple.mpegurl",headers={"Access-Control-Allow-Origin":"*","Cache-Control":"public, max-age=300"})
+    except:return Response(status_code=502, content="proxy error")
+
+@app.get("/api/proxy/seg")
+def proxy_segment(url: str = Query(...)):
+    try:
+        r = requests.get(url, headers=HEADERS, timeout=30)
+        if r.status_code != 200:return Response(status_code=502, content="upstream error")
+        data = r.content
+        if len(data) > 188 and data[0:4] == b'\x89PNG' and data[188] == 0x47:data = data[188:]
+        return Response(content=data,media_type="video/mp2t",headers={"Access-Control-Allow-Origin":"*","Cache-Control":"public, max-age=3600"})
+    except:return Response(status_code=502, content="proxy error")
+
+@app.get("/api/proxy/video")
+def proxy_video(url: str = Query(...), request: Request = None):
+    try:
+        req_headers = dict(HEADERS)
+        if request and request.headers.get("range"):req_headers["Range"] = request.headers["range"]
+        r = requests.get(url, headers=req_headers, timeout=30, stream=True)
+        resp_headers = {"Access-Control-Allow-Origin":"*","Accept-Ranges":"bytes","Content-Type":r.headers.get("Content-Type","video/mp4")}
+        if "Content-Range" in r.headers:resp_headers["Content-Range"] = r.headers["Content-Range"]
+        if "Content-Length" in r.headers:resp_headers["Content-Length"] = r.headers["Content-Length"]
+        return StreamingResponse(r.iter_content(chunk_size=256*1024),status_code=r.status_code,headers=resp_headers)
+    except:return Response(status_code=502, content="proxy error")
+
+@app.get("/api/proxy/img")
+def proxy_img(url: str = Query(...)):
+    """Proxy images from sources that block hotlinking (DanTri CDN)."""
+    try:
+        r = requests.get(url, headers={**HEADERS, "Referer": "https://dantri.com.vn/"}, 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 _scrape_xemlaibongda_page(page_path, limit=20):
+    try:
+        url = f"https://xemlaibongda.top/{page_path}" if page_path else "https://xemlaibongda.top/"
+        r=requests.get(url,headers=HEADERS,timeout=15)
+        if r.status_code!=200:return[]
+        r.encoding="utf-8";soup=BeautifulSoup(r.text,"lxml");videos=[];seen=set()
+        for a in soup.find_all("a",href=True):
+            href=a.get("href","")
+            if"/video/" not in href:continue
+            if not href.startswith("http"):href="https://xemlaibongda.top"+href
+            if href in seen:continue
+            seen.add(href);slug=href.split("/video/")[-1].rstrip("/")
+            title=slug.replace("-"," ").title()
+            title=re.sub(r'\d{4}\s*\d{2}\s*\d{2}$','',title).strip()
+            title=re.sub(r'\s+V\s+',' vs ',title);title=re.sub(r'\s+Vs\s+',' vs ',title)
+            img=a.find("img") or (a.parent.find("img") if a.parent else None)
+            img_src=""
+            if img:img_src=img.get("data-src","") or img.get("src","") or img.get("data-lazy","")
+            if not img_src:img_src=f"https://img.refooty.com/thumbnail/{slug}.webp"
+            videos.append({"title":title,"link":href,"img":img_src,"source":"xemlaibongda"})
+            if len(videos)>=limit:break
+        return videos
+    except:return[]
+
+def scrape_xemlaibongda():return _scrape_xemlaibongda_page("",20)
+def scrape_highlights_by_league(league_key):
+    if league_key not in HL_LEAGUES:return[]
+    return _scrape_xemlaibongda_page(HL_LEAGUES[league_key]["path"],20)
+
+def scrape_all_league_highlights():
+    results = {}
+    def _fetch(key):return key, scrape_highlights_by_league(key)
+    with ThreadPoolExecutor(8) as ex:
+        futs = [ex.submit(_fetch, k) for k in HL_LEAGUES]
+        for f in as_completed(futs):
+            try:
+                key, vids = f.result()
+                if vids:results[key] = vids
+            except:pass
+    return results
+
+def extract_xemlaibongda_video(url):
+    try:
+        r=requests.get(url,headers=HEADERS,timeout=15)
+        if r.status_code!=200:return None
+        r.encoding="utf-8";soup=BeautifulSoup(r.text,"lxml");video=soup.find("video")
+        if video:
+            src=video.get("src","");poster=video.get("poster","")
+            if not src:
+                source=video.find("source")
+                if source:src=source.get("src","")
+            if src:return{"src":src,"poster":poster,"type":"hls" if".m3u8" in src else"video"}
+        m3u8s=re.findall(r'(https?://[^\s"\'<>]+\.m3u8)',r.text)
+        if m3u8s:
+            og=soup.find("meta",property="og:image");poster=og.get("content","") if og else ""
+            return{"src":m3u8s[0],"poster":poster,"type":"hls"}
+        return None
+    except:return None
+
+# ===== YOUTUBE SHORTS =====
+def _yt_channel_shorts(channel, count=15):
+    try:
+        result=subprocess.run(["yt-dlp","--flat-playlist","--dump-json","--no-download","--playlist-end",str(count),f"https://www.youtube.com/@{channel}/shorts"],capture_output=True,text=True,timeout=60)
+        videos=[]
+        for line in result.stdout.strip().split('\n'):
+            if not line:continue
+            try:
+                info=json.loads(line);vid=info.get("id","");title=info.get("title","")
+                if not vid or not title:continue
+                videos.append({"title":title,"link":f"https://www.youtube.com/watch?v={vid}","img":f"https://i.ytimg.com/vi/{vid}/hqdefault.jpg","source":"yt","id":vid})
+            except:pass
+        return videos
+    except:return[]
+def scrape_shorts():
+    """Fetch shorts from multiple channels"""
+    vids=[]
+    with ThreadPoolExecutor(2) as ex:
+        futs=[ex.submit(_yt_channel_shorts,ch,10) for ch in ["baodantri7941","baosuckhoedoisongboyte"]]
+        for f in as_completed(futs):
+            try:vids.extend(f.result())
+            except:pass
+    return vids[:20]
+
+# ===== LIVESCORE =====
+@app.get("/api/livescore/live")
+def api_livescore_live():return JSONResponse({"html":_cached("ls_live",lambda:fetch_bongda_api("/api/fixtures/live"),ttl=_cache_ttl_live)})
+@app.get("/api/livescore/incoming")
+def api_livescore_incoming():return JSONResponse({"html":_cached("ls_incoming",lambda:fetch_bongda_api("/api/fixtures/incoming"),ttl=_cache_ttl_live)})
+@app.get("/api/livescore/today")
+def api_livescore_today():
+    today=datetime.now().strftime("%Y-%m-%d");return JSONResponse({"html":_cached("ls_today",lambda:fetch_bongda_api(f"/api/fixtures/get-by-date?date={today}"),ttl=_cache_ttl)})
+@app.get("/api/livescore/results")
+def api_livescore_results():
+    today=datetime.now().strftime("%Y-%m-%d");return JSONResponse({"html":_cached("ls_results",lambda:fetch_bongda_api(f"/api/fixtures/get-by-date?date={today}&status=finished"),ttl=_cache_ttl)})
+@app.get("/api/livescore/standings/{league}")
+def api_livescore_standings(league:str):
+    tid=LEAGUE_IDS.get(league,27110);return JSONResponse({"html":_cached(f"ls_bxh_{league}",lambda:fetch_bongda_api(f"/api/league-table/home?tournament_id={tid}&is_detail=True"),ttl=_cache_ttl)})
+@app.get("/api/livescore/date/{date}")
+def api_livescore_date(date:str):return JSONResponse({"html":fetch_bongda_api(f"/api/fixtures/get-by-date?date={date}")})
+@app.get("/api/match/{event_id}/commentaries")
+def api_match_commentaries(event_id:int):return JSONResponse({"html":fetch_bongda_api(f"/api/fixtures/commentaries?event_id={event_id}")})
+@app.get("/api/match/{event_id}/stats")
+def api_match_stats(event_id:int):return JSONResponse({"html":fetch_bongda_api(f"/api/event-standing/player-performance?event_id={event_id}")})
+@app.get("/api/livescore/featured")
+def api_livescore_featured():
+    def _f():
+        sources=[("/api/fixtures/live","live"),("/api/fixtures/get-by-date?date="+datetime.now().strftime("%Y-%m-%d"),"today"),("/api/fixtures/incoming","upcoming")]
+        for endpoint, stype in sources:
+            html=fetch_bongda_api(endpoint)
+            if not html or len(html)<100:continue
+            soup=BeautifulSoup(html,"lxml");all_matches=[]
+            for li in soup.select("li.match-detail"):
+                match=_parse_match_from_li(li, stype)
+                if not match or not match["event_id"]:continue
+                if stype=="today" and "KT" in match.get("minute",""):continue
+                all_matches.append(match)
+            if not all_matches:continue
+            for pl in PRIORITY_LEAGUES:
+                for match in all_matches:
+                    if pl in match["league"]:return match
+            return all_matches[0]
+        return None
+    return JSONResponse(_cached("ls_featured",_f,ttl=30))
+
+# ===== VIDEO APIs =====
+@app.get("/api/shorts")
+def api_shorts():return JSONResponse(_cached("yt_shorts",scrape_shorts,ttl=_cache_ttl_yt))
+@app.get("/api/highlights")
+def api_highlights():return JSONResponse(_cached("xemlaibongda_hl",scrape_xemlaibongda,ttl=_cache_ttl))
+@app.get("/api/highlights/leagues")
+def api_highlights_leagues():return JSONResponse(_cached("hl_leagues",scrape_all_league_highlights,ttl=_cache_ttl))
+@app.get("/api/highlights/{league}")
+def api_highlights_league(league:str):
+    if league not in HL_LEAGUES:return JSONResponse({"error":"league not found"})
+    return JSONResponse(_cached(f"hl_{league}",lambda:scrape_highlights_by_league(league),ttl=_cache_ttl))
+@app.get("/api/highlights_config")
+def api_highlights_config():return JSONResponse(HL_LEAGUES)
+@app.get("/api/video_url")
+def api_video_url(url:str=Query(...)):
+    if "youtube.com" in url or "youtu.be" in url:
+        m=re.search(r'(?:v=|shorts/|youtu\.be/)([a-zA-Z0-9_-]{11})',url)
+        if m:vid=m.group(1);return JSONResponse({"src":f"https://www.youtube.com/embed/{vid}?autoplay=1&rel=0&enablejsapi=1","poster":f"https://i.ytimg.com/vi/{vid}/hqdefault.jpg","type":"youtube"})
+    if "xemlaibongda.top" in url:
+        v=extract_xemlaibongda_video(url)
+        if v:
+            if v["type"]=="hls":v["src"]="/api/proxy/m3u8?url="+quote(v["src"],safe="")
+            return JSONResponse(v)
+    if "bongdaplus.vn" in url:
+        try:
+            m=re.search(r'-(\d{6,})\.html',url)
+            if m:
+                r=requests.get(f"{BASE_BDP}/video-embed/{m.group(1)}.html",headers=HEADERS,timeout=10);r.encoding="utf-8"
+                soup=BeautifulSoup(r.text,"lxml");video=soup.select_one("video#videoPlayer")
+                if video:
+                    source=video.find("source");src=source.get("src","") if source else "";poster=video.get("poster","")
+                    if src:return JSONResponse({"src":"/api/proxy/video?url="+quote(src,safe=""),"poster":poster,"type":"video"})
+        except:pass
+    return JSONResponse({"error":"not found"})
+@app.get("/api/bdp_videos")
+def api_bdp_videos():
+    def _f():
+        try:
+            soup=_get(f"{BASE_BDP}/video");arts=[];seen=set()
+            for a in soup.find_all("a",href=True):
+                href=a.get("href","")
+                if"/video/" not in href or href in("/video/","/video/ban-thang-dep","/video/highlight"):continue
+                if not href.startswith("http"):href=BASE_BDP+href
+                if href in seen:continue
+                title=re.sub(r'^\d{2}:\d{2}','',a.get_text(strip=True)).strip()
+                if not title or len(title)<5:continue
+                img_tag=a.find("img") or(a.parent.find("img") if a.parent else None)
+                img=(img_tag.get("data-src") or img_tag.get("src","")) if img_tag else ""
+                seen.add(href);arts.append({"title":title,"link":href,"img":img,"source":"bdp"})
+            return arts[:20]
+        except:return[]
+    return JSONResponse(_cached("bdp_videos",_f))
+# ===== NEWS =====
+def scrape_vne(cat_url):
+    try:
+        soup=_get(cat_url);arts=[]
+        for it in soup.select("article.item-news")[:15]:
+            a=it.select_one("h2.title-news a") or it.select_one("h3.title-news a")
+            if not a:continue
+            t=a.get("title","") or a.get_text(strip=True);lk=a.get("href","")
+            if not t or not lk:continue
+            im=it.find("img");img=(im.get("data-src") or im.get("src","")) if im else ""
+            if img and'blank'in img:
+                src=it.find("source")
+                if src:img=src.get("srcset","").split(",")[0].strip().split(" ")[0]
+            arts.append({"title":t,"link":lk,"img":img,"source":"vne"})
+        return arts
+    except:return[]
+def scrape_vne_article(url):
+    try:
+        soup=_get(url);h1=soup.select_one("h1.title-detail");desc=soup.select_one("p.description")
+        og=soup.find("meta",property="og:image");og_img=og.get("content","") if og else ""
+        cd=soup.select_one("article.fck_detail");body=[]
+        if cd:
+            for ch in cd.children:
+                if not hasattr(ch,'name') or not ch.name:continue
+                if ch.name=="p":t=ch.get_text(strip=True);(body.append({"type":"p","text":t}) if t else None)
+                elif ch.name=="figure":
+                    im=ch.find("img")
+                    if im:s=im.get("data-src") or im.get("src","");body.append({"type":"img","src":s})
+                elif ch.name in("h2","h3"):body.append({"type":"heading","text":ch.get_text(strip=True)})
+        return{"title":h1.get_text(strip=True) if h1 else "","summary":desc.get_text(strip=True) if desc else "","og_image":og_img,"body":body,"source":"vne","url":url}
+    except:return None
+def _scrape_dantri_homepage(cat_filter=None):
+    try:
+        soup=_get("https://dantri.com.vn/");arts=[];seen=set()
+        for a in soup.find_all("a",href=True):
+            href=a.get("href","");title=a.get("title","") or a.get_text(strip=True)
+            if not title or len(title)<15 or"javascript:" in href:continue
+            if not href.startswith("http"):href="https://dantri.com.vn"+href
+            if href in seen or not href.endswith(".htm"):continue
+            if cat_filter and f"/{cat_filter}/" not in href:continue
+            img_tag=a.find("img")
+            if not img_tag and a.parent:img_tag=a.parent.find("img")
+            img_src=""
+            if img_tag:img_src=img_tag.get("data-src","") or img_tag.get("src","")
+            if not img_src or "cdn" not in img_src:continue
+            proxied_img="/api/proxy/img?url="+quote(img_src,safe="")
+            seen.add(href);arts.append({"title":title,"link":href,"img":proxied_img,"source":"dantri"})
+            if len(arts)>=15:break
+        return arts
+    except:return[]
+def scrape_dantri_hot():return _scrape_dantri_homepage()
+def scrape_dantri_congnghe():
+    try:
+        soup=_get("https://dantri.com.vn/");arts=[];seen=set()
+        for a in soup.find_all("a",href=True):
+            href=a.get("href","");title=a.get("title","") or a.get_text(strip=True)
+            if not title or len(title)<15 or"javascript:" in href:continue
+            if not href.startswith("http"):href="https://dantri.com.vn"+href
+            if href in seen or not href.endswith(".htm"):continue
+            if"/cong-nghe/" not in href:continue
+            img_tag=a.find("img")
+            if not img_tag and a.parent:img_tag=a.parent.find("img")
+            img_src=""
+            if img_tag:img_src=img_tag.get("data-src","") or img_tag.get("src","")
+            if img_src and "cdn" in img_src:img_src="/api/proxy/img?url="+quote(img_src,safe="")
+            else:img_src=""
+            seen.add(href);arts.append({"title":title,"link":href,"img":img_src,"source":"dantri"})
+            if len(arts)>=15:break
+        return arts
+    except:return[]
+def scrape_genk_ai():
+    """Scrape AI articles from genk.vn - readable in-app"""
+    try:
+        r=requests.get("https://genk.vn/ai.chn",headers=HEADERS,timeout=15)
+        if r.status_code!=200:return[]
+        r.encoding="utf-8";soup=BeautifulSoup(r.text,"lxml")
+        articles=[];seen=set()
+        for a in soup.find_all("a",href=True):
+            href=a.get("href","")
+            if not href.endswith(".chn") or href=="/ai.chn":continue
+            if href.startswith("/"):href="https://genk.vn"+href
+            if href in seen or "genk.vn" not in href:continue
+            title=a.get("title","") or a.get_text(strip=True)
+            if not title or len(title)<20:continue
+            container=a.parent;img_src=""
+            for _ in range(6):
+                if container is None:break
+                for img in container.find_all("img"):
+                    s=img.get("data-src","") or img.get("src","")
+                    if s and "mediacdn" in s and "avatar" not in s and "logo" not in s:
+                        img_src=s;break
+                if img_src:break
+                container=container.parent
+            seen.add(href)
+            if not img_src:
+                try:
+                    og_r=requests.get(href,headers=HEADERS,timeout=8);og_r.encoding="utf-8"
+                    og_soup=BeautifulSoup(og_r.text,"lxml");og_tag=og_soup.find("meta",property="og:image")
+                    if og_tag:img_src=og_tag.get("content","")
+                except:pass
+            articles.append({"title":title,"link":href,"img":img_src,"source":"genk"})
+            if len(articles)>=30:break
+        return articles
+    except:return[]
+
+def scrape_dantri_article(url):
+    try:
+        r=requests.get(url,headers=HEADERS,timeout=15);r.encoding="utf-8";soup=BeautifulSoup(r.text,"lxml")
+        for tag in soup.find_all(["script","style","nav","footer","aside"]):tag.decompose()
+        h1=soup.find("h1");og=soup.find("meta",property="og:image");og_img=og.get("content","") if og else ""
+        if og_img and "cdnphoto.dantri" in og_img:og_img="/api/proxy/img?url="+quote(og_img,safe="")
+        content=soup.select_one("main") or soup.select_one("div.singular-content") or soup.select_one("article");body=[]
+        if content:
+            for el in content.find_all(["p","h2","h3","figure","img"],recursive=True):
+                if el.name=="p":t=el.get_text(strip=True);(body.append({"type":"p","text":t}) if t and len(t)>15 else None)
+                elif el.name in("h2","h3"):t=el.get_text(strip=True);(body.append({"type":"heading","text":t}) if t else None)
+                elif el.name in("figure","img"):
+                    im=el if el.name=="img" else el.find("img")
+                    if im:
+                        s=im.get("data-src") or im.get("src","")
+                        if s and"base64" not in s:
+                            if "cdnphoto.dantri" in s:s="/api/proxy/img?url="+quote(s,safe="")
+                            body.append({"type":"img","src":s})
+        desc="";sapo=soup.select_one("h2.singular-sapo") or soup.select_one("h2[class*=sapo]")
+        if not sapo:
+            og_desc=soup.find("meta",property="og:description")
+            if og_desc:desc=og_desc.get("content","")
+        else:desc=sapo.get_text(strip=True)
+        return{"title":h1.get_text(strip=True) if h1 else "","summary":desc,"og_image":og_img,"body":body,"source":"dantri","url":url}
+    except:return None
+def scrape_bbc_vietnamese():
+    try:
+        r=requests.get("https://www.bbc.com/vietnamese",headers={"User-Agent":"Mozilla/5.0","Accept-Language":"en-GB"},timeout=15);r.encoding="utf-8"
+        soup=BeautifulSoup(r.text,"lxml");arts=[];seen=set()
+        for a in soup.select("a[href*='/vietnamese/']"):
+            href=a.get("href","")
+            if not href or href=="/vietnamese" or href.count("/")<3:continue
+            if not href.startswith("http"):href="https://www.bbc.com"+href
+            if href in seen:continue
+            title=a.get_text(strip=True)
+            if not title or len(title)<15 or any(x in title.lower() for x in["đăng nhập","trang chủ","bbc news"]):continue
+            img="";container=a.parent
+            for _ in range(3):
+                if container:
+                    im=container.find("img")
+                    if im:img=im.get("src","") or im.get("data-src","");break
+                    container=container.parent
+            seen.add(href);arts.append({"title":title,"link":href,"img":img,"source":"bbc"})
+            if len(arts)>=15:break
+        return arts
+    except:return[]
+def scrape_bbc_article(url):
+    try:
+        r=requests.get(url,headers={"User-Agent":"Mozilla/5.0","Accept-Language":"en-GB"},timeout=15);r.encoding="utf-8"
+        soup=BeautifulSoup(r.text,"lxml");h1=soup.find("h1")
+        og=soup.find("meta",property="og:image");og_img=og.get("content","") if og else ""
+        body=[]
+        for p in soup.select("[data-component='text-block'] p, article p, main p"):
+            t=p.get_text(strip=True)
+            if t and len(t)>20:body.append({"type":"p","text":t})
+        return{"title":h1.get_text(strip=True) if h1 else "","summary":"","og_image":og_img,"body":body,"source":"bbc","url":url}
+    except:return None
+VNE_CATS={"thoi-su":("https://vnexpress.net/thoi-su","Thời Sự"),"the-gioi":("https://vnexpress.net/the-gioi","Thế Giới"),"kinh-doanh":("https://vnexpress.net/kinh-doanh","Kinh Doanh"),"the-thao":("https://vnexpress.net/the-thao","Thể Thao"),"giai-tri":("https://vnexpress.net/giai-tri","Giải Trí"),"suc-khoe":("https://vnexpress.net/suc-khoe","Sức Khỏe"),"phap-luat":("https://vnexpress.net/phap-luat","Pháp Luật"),"giao-duc":("https://vnexpress.net/giao-duc","Giáo Dục"),"du-lich":("https://vnexpress.net/du-lich","Du Lịch"),"doi-song":("https://vnexpress.net/doi-song","Đời Sống")}
+@app.get("/api/homepage")
+def api_homepage():
+    def _f():
+        articles=[]
+        with ThreadPoolExecutor(12) as ex:
+            futs={ex.submit(scrape_vne,VNE_CATS[k][0]):VNE_CATS[k][1] for k in["thoi-su","the-gioi","kinh-doanh","the-thao","giai-tri","phap-luat","giao-duc","du-lich","doi-song"]}
+            futs[ex.submit(scrape_bbc_vietnamese)]="BBC"
+            futs[ex.submit(scrape_genk_ai)]="Công Nghệ"
+            for f in as_completed(futs):
+                try:
+                    for a in f.result():a["group"]=futs[f];articles.append(a)
+                except:pass
+        return articles
+    return JSONResponse(_cached("homepage",_f))
+@app.get("/api/category/{cat_id}")
+def api_category(cat_id:str):
+    def _f():
+        if cat_id=="bbc":return scrape_bbc_vietnamese()
+        if cat_id=="cong-nghe":return scrape_genk_ai()
+        if cat_id in VNE_CATS:arts=scrape_vne(VNE_CATS[cat_id][0]);[a.update({"group":VNE_CATS[cat_id][1]}) for a in arts];return arts
+        return[]
+    return JSONResponse(_cached(f"cat_{cat_id}",_f))
+@app.get("/api/categories")
+def api_categories():
+    cats=[{"id":"bbc","name":"BBC Tiếng Việt","source":"bbc"},{"id":"cong-nghe","name":"Công Nghệ","source":"genk"}]
+    for k,(u,n) in VNE_CATS.items():cats.append({"id":k,"name":n,"source":"vne"})
+    return JSONResponse(cats)
+@app.get("/api/dantri_hot")
+def api_dantri_hot():return JSONResponse(_cached("dantri_hot",scrape_dantri_hot))
+@app.get("/api/genk_ai")
+def api_genk_ai():return JSONResponse(_cached("genk_ai",scrape_genk_ai,ttl=_cache_ttl))
+def scrape_genk_article(url):
+    try:
+        r=requests.get(url,headers=HEADERS,timeout=15);r.encoding="utf-8";soup=BeautifulSoup(r.text,"lxml")
+        h1=soup.find("h1");og=soup.find("meta",property="og:image");og_img=og.get("content","") if og else ""
+        desc_el=soup.find("meta",property="og:description");desc=desc_el.get("content","") if desc_el else ""
+        cd=soup.select_one(".knc-content");body=[]
+        if cd:
+            for el in cd.find_all(["p","h2","h3","figure","img"],recursive=True):
+                if el.name=="p":t=el.get_text(strip=True);(body.append({"type":"p","text":t}) if t and len(t)>15 else None)
+                elif el.name in("h2","h3"):t=el.get_text(strip=True);(body.append({"type":"heading","text":t}) if t else None)
+                elif el.name in("figure","img"):
+                    im=el if el.name=="img" else el.find("img")
+                    if im:s=im.get("data-src") or im.get("src","");(body.append({"type":"img","src":s}) if s and"base64" not in s else None)
+        return{"title":h1.get_text(strip=True) if h1 else "","summary":desc,"og_image":og_img,"body":body,"source":"genk","url":url}
+    except:return None
+
+@app.get("/api/article")
+def api_article(url:str=Query(...)):
+    if"vnexpress.net" in url:data=scrape_vne_article(url)
+    elif"bbc.com" in url:data=scrape_bbc_article(url)
+    elif"dantri.com.vn" in url:data=scrape_dantri_article(url)
+    elif"genk.vn" in url:data=scrape_genk_article(url)
+    else:data=None
+    return JSONResponse(data if data else{"error":"not supported"})
+@app.get("/v")
+async def video_share(url:str=Query(default=""),title:str=Query(default="VNEWS Video"),img:str=Query(default=""),type:str=Query(default="highlights")):
+    decoded_url=unquote(url);decoded_title=unquote(title)
+    redirect_script=f'<script>localStorage.setItem("pending_video",JSON.stringify({{"url":"{decoded_url}","type":"{type}"}}));location.href="{SPACE_URL}";</script>' if decoded_url else f'<script>location.href="{SPACE_URL}";</script>'
+    return HTMLResponse(f'<!DOCTYPE html><html><head><meta charset="utf-8"><title>{decoded_title}</title></head><body style="background:#111;color:#fff;text-align:center;padding:40px"><p>⏳</p>{redirect_script}</body></html>')
+@app.get("/s")
+async def share_redirect(url:str=Query(default=""),title:str=Query(default="VNEWS"),img:str=Query(default="")):
+    decoded_url=unquote(url)
+    redirect_script=f'<script>localStorage.setItem("pending_article","{decoded_url}");location.href="{SPACE_URL}";</script>' if decoded_url else f'<script>location.href="{SPACE_URL}";</script>'
+    return HTMLResponse(f'<!DOCTYPE html><html><head><meta charset="utf-8"><title>{unquote(title)}</title></head><body>{redirect_script}</body></html>')
+@app.get("/")
+async def index():
+    with open("/app/static/index.html","r",encoding="utf-8") as f:return HTMLResponse(content=f.read())
+app.mount("/static",StaticFiles(directory="/app/static"),name="static")