Fix VTV: stable HLS config, EPG API, proxy m3u8 fix

vtv_api.py

@@ -1,11 +1,13 @@
 """
 VTV Channels API - Backend endpoints for VTV1-VTV10 + VTVPrime
 Fetches stream URLs from hd.xemtv.net PHP endpoints and VTVPrime
+EPG schedule scraped from vtv.vn
 """
 import re, time, threading
 import requests
 from fastapi import APIRouter, Query
 from fastapi.responses import JSONResponse, Response
+from bs4 import BeautifulSoup
 
 router = APIRouter()
 
@@ -15,6 +17,102 @@ UA = {
     "Referer": "https://hd.xemtv.net/",
 }
 
+# ===== EPG: VTV.vn schedule scraping =====
+VTV_SCHEDULE_SLUGS = {
+    "vtv1": "vtv1", "vtv2": "vtv2", "vtv3": "vtv3", "vtv4": "vtv4",
+    "vtv5": "vtv5", "vtv6": "vtv6", "vtv7": "vtv7", "vtv8": "vtv8",
+    "vtv9": "vtv9", "vtv10": "vtv-can-tho", "vtvprime": "vtvprime",
+}
+
+_epg_cache = {}
+_epg_lock = threading.Lock()
+_EPG_CACHE_TTL = 3600  # 1 hour
+
+
+def _get_cached_epg(ch_id):
+    with _epg_lock:
+        if ch_id in _epg_cache and time.time() - _epg_cache[ch_id]['t'] < _EPG_CACHE_TTL:
+            return _epg_cache[ch_id]['d']
+    return None
+
+
+def _set_epg_cache(ch_id, data):
+    with _epg_lock:
+        _epg_cache[ch_id] = {'t': time.time(), 'd': data}
+
+
+def fetch_vtv_epg(channel_id):
+    """Fetch EPG schedule for a VTV channel from vtv.vn."""
+    cached = _get_cached_epg(channel_id)
+    if cached is not None:
+        return cached
+
+    slug = VTV_SCHEDULE_SLUGS.get(channel_id, channel_id)
+    urls = [
+        f"https://vtv.vn/lich-phat-song/{slug}.htm",
+        f"https://vtv.vn/lich-phat-song.htm",
+    ]
+
+    for url in urls:
+        try:
+            r = requests.get(url, headers={
+                "User-Agent": UA["User-Agent"],
+                "Accept-Language": "vi-VN,vi;q=0.9",
+            }, timeout=10, allow_redirects=True)
+            if r.status_code != 200:
+                continue
+            r.encoding = "utf-8"
+            soup = BeautifulSoup(r.text, "lxml")
+            schedule = []
+
+            # Method 1: table rows with time + program
+            for table in soup.find_all("table"):
+                for row in table.find_all("tr"):
+                    cells = row.find_all(["td", "th"])
+                    if len(cells) >= 2:
+                        time_text = cells[0].get_text(strip=True)
+                        prog_text = cells[1].get_text(strip=True)
+                        if re.match(r'\d{1,2}:\d{2}', time_text) and prog_text and len(prog_text) > 2:
+                            schedule.append({"t": time_text, "n": prog_text[:80]})
+
+            if schedule:
+                _set_epg_cache(channel_id, schedule)
+                return schedule
+
+            # Method 2: list items with time dash program
+            for ul in soup.find_all("ul"):
+                for li in ul.find_all("li"):
+                    text = li.get_text(" ", strip=True)
+                    m = re.match(r'(\d{1,2}:\d{2})\s*[–\-:]\s*(.+)', text)
+                    if m and len(m.group(2)) > 2:
+                        schedule.append({"t": m.group(1), "n": m.group(2).strip()[:80]})
+
+            if schedule:
+                _set_epg_cache(channel_id, schedule)
+                return schedule
+
+            # Method 3: divs with time/program children
+            time_els = soup.find_all(text=re.compile(r'^\d{1,2}:\d{2}$'))
+            for te in time_els:
+                parent = te.parent
+                if parent:
+                    prog_el = parent.find_next_sibling() or parent.parent
+                    if prog_el:
+                        prog_text = prog_el.get_text(strip=True)
+                        if prog_text and len(prog_text) > 2:
+                            schedule.append({"t": te.strip(), "n": prog_text[:80]})
+
+            if schedule:
+                _set_epg_cache(channel_id, schedule)
+                return schedule
+
+        except Exception:
+            continue
+
+    _set_epg_cache(channel_id, None)
+    return None
+
+
 # Channel ID -> xemtv.net PHP endpoint mapping
 XEMTV_PHP_ENDPOINTS = {
     "vtv1":  "https://hd.xemtv.net/kenh/vtv1.php",
@@ -58,7 +156,6 @@ VTVGO_FAILOVER = {
 }
 
 # ===== Channels that should use xemtv.net scraping (not VTVGo failover) =====
-# VTV6, VTV10: xemtv.net scraping (returns fptplay CDN URLs that need proxying)
 XEMTV_ONLY_CHANNELS = {"vtv6", "vtv10"}
 
 # ===== LAST RESORT: fptplay CDN URLs (need proxy for referer) =====
@@ -170,7 +267,6 @@ def fetch_vtv_stream(channel_id):
     # --- VTV10: xemtv returns vtvcantho (403), skip to fptplay CDN directly ---
     if channel_id in XEMTV_ONLY_CHANNELS:
         if channel_id == "vtv10":
-            # VTV10: xemtv.net returns dead vtvcantho URL (403), use fptplay CDN directly
             fpt_url = FPTPLAY_URLS.get(channel_id)
             if fpt_url:
                 _set_cache(channel_id, fpt_url)
@@ -242,6 +338,16 @@ def api_vtv_stream(channel_id: str):
     return JSONResponse({"error": "stream not found", "status": "offline"}, status_code=404)
 
 
+@router.get("/api/vtv/epg/{channel_id}")
+def api_vtv_epg(channel_id: str):
+    """Get EPG schedule for a specific VTV channel."""
+    channel_id = channel_id.lower().strip()
+    epg = fetch_vtv_epg(channel_id)
+    if epg:
+        return JSONResponse({"channel_id": channel_id, "schedule": epg})
+    return JSONResponse({"channel_id": channel_id, "schedule": [], "source": "none"})
+
+
 @router.get("/api/proxy/page")
 def proxy_page(url: str = Query(...)):
     """Proxy a web page."""
@@ -263,7 +369,12 @@ def proxy_page(url: str = Query(...)):
 
 @router.get("/api/proxy/m3u8/vtv")
 def proxy_vtv_m3u8(url: str = Query(...)):
-    """Proxy an m3u8 stream URL with proper headers for fptplay CDN."""
+    """Proxy an m3u8 stream URL with proper headers for fptplay CDN.
+    
+    CRITICAL FIX: Only proxy segment URLs (.ts, .aac, etc.), NOT sub-manifest URLs (.m3u8).
+    Sub-manifests must be fetched directly by the client (HLS.js) to avoid
+    rewriting issues with relative paths inside nested playlists.
+    """
     try:
         headers = {"User-Agent": UA["User-Agent"], "Accept": "*/*"}
         if "fptplay" in url:
@@ -286,7 +397,12 @@ def proxy_vtv_m3u8(url: str = Query(...)):
                 seg_url = line
                 if not seg_url.startswith('http'):
                     seg_url = base_url + seg_url
-                rewritten.append("/api/proxy/seg/vtv?url=" + requests.utils.quote(seg_url, safe=""))
+                # Only proxy segment files (.ts, .aac, .mp4, .webvtt, .jpg, .png etc.)
+                # Do NOT proxy .m3u8 sub-manifests — let HLS.js fetch them directly
+                if seg_url.endswith('.m3u8') or seg_url.endswith('.m3u'):
+                    rewritten.append(seg_url)
+                else:
+                    rewritten.append("/api/proxy/seg/vtv?url=" + requests.utils.quote(seg_url, safe=""))
         return Response(
             content='\n'.join(rewritten).encode('utf-8'),
             media_type="application/vnd.apple.mpegurl",