feat: add _extract_24h_video_url to shorts_carousel.py

shorts_carousel.py

@@ -1,4 +1,4 @@
-"""Self-contained 24h Shorts Carousel - scrapes video-tin-tuc page for actual video articles."""
+"""Self-contained 24h Shorts Carousel + video extraction."""
 import re
 import requests
 from bs4 import BeautifulSoup
@@ -12,7 +12,7 @@ BASE_24H = "https://www.24h.com.vn"
 
 def scrape_24h_news_shorts():
     """Scrape https://www.24h.com.vn/video/video-tin-tuc-cvd769.html
-    Returns only ACTUAL article links (pattern: -c{N}a{N}.html), not category pages."""
+    Returns only ACTUAL article links (pattern: -c{N}a{N}.html)."""
     try:
         url = f"{BASE_24H}/video/video-tin-tuc-cvd769.html"
         r = requests.get(url, headers=HEADERS, timeout=15)
@@ -21,7 +21,6 @@ def scrape_24h_news_shorts():
         articles, seen = [], set()
         for a in soup.find_all("a", href=True):
             href = a.get("href", "")
-            # Only match real article URLs: ends with -c{digits}a{digits}.html
             if not re.search(r'-c\d+a\d+\.html$', href):
                 continue
             if not href.startswith("http"):
@@ -29,7 +28,6 @@ def scrape_24h_news_shorts():
             if href in seen:
                 continue
             seen.add(href)
-            # Get title
             title = a.get("title", "") or ""
             img = a.find("img")
             if img and not title:
@@ -38,13 +36,11 @@ def scrape_24h_news_shorts():
                 title = a.get_text(strip=True)
             if not title or len(title) < 10:
                 continue
-            # Get image - priority: data-original > data-src > src
             img_src = None
             if img:
                 img_src = img.get("data-original") or img.get("data-src") or img.get("src") or ""
                 if "base64" in img_src or len(img_src) < 20:
                     img_src = None
-            # Try parent element for image if not found in link
             if not img_src:
                 parent = a.parent
                 if parent:
@@ -54,7 +50,7 @@ def scrape_24h_news_shorts():
                         if "base64" in img_src or len(img_src) < 20:
                             img_src = None
             if not img_src:
-                continue  # Skip items without thumbnail for visual carousel
+                continue
             articles.append({
                 "title": title, "link": href, "img": img_src,
                 "source": "24h", "is_video": True
@@ -64,6 +60,26 @@ def scrape_24h_news_shorts():
         return []
 
 
+def _extract_24h_video_url(article_url):
+    """Extract m3u8 video URL + poster from a 24h article page.
+    Returns dict {src, poster, vtype} or None."""
+    try:
+        r = requests.get(article_url, headers=HEADERS, timeout=10)
+        r.encoding = "utf-8"
+        m3u8s = re.findall(r'(https?://cdn\.24h\.com\.vn/[^\s"\'\\]+\.m3u8)', r.text)
+        full = [u for u in m3u8s if '_720p' not in u]
+        if not full:
+            full = m3u8s
+        if not full:
+            return None
+        soup = BeautifulSoup(r.text, "lxml")
+        og = soup.find("meta", property="og:image")
+        poster = og.get("content", "") if og else ""
+        return {"src": full[0], "poster": poster, "vtype": "hls"}
+    except:
+        return None
+
+
 # CSS for the 9:16 shorts slider
 SHORTS_CSS = """
 /* Shorts 9:16 Slider */
@@ -78,8 +94,7 @@ SHORTS_CSS = """
 
 def render_shorts_carousel(shorts, esc_fn, safe_url_fn, make_id_fn, slug_fn):
     """Render 9:16 vertical shorts carousel HTML.
-    Click opens the article directly (bdpOpen) so user sees the actual video.
-    """
+    Click opens the shorts article in TikTok fullscreen mode."""
     vids = [v for v in shorts if v.get("img")]
     if not vids:
         return ""
@@ -90,7 +105,7 @@ def render_shorts_carousel(shorts, esc_fn, safe_url_fn, make_id_fn, slug_fn):
         title = v.get("title", "")
         aid = make_id_fn(link)
         sl = slug_fn(title)
-        # Use bdpOpen to open article view with video player
+        # bdpOpen opens the article - which now renders as TikTok feed for 24h
         cjs = "window.bdpOpen('" + esc_fn(link) + "','" + aid + "','" + sl + "')"
         items.append(
             '<div class="vslide-shorts-item" onclick="' + cjs + '">'