Auto scheduler: 3 topics HOT khác nhau ko trùng, mỗi topic 1 bài

app_v2_entry.py

@@ -1326,10 +1326,16 @@ async def _auto_fetch_short(post_id):
     except: pass
     return False
 
-async def _auto_rewrite_one(topic, slot_label):
-    """Rewrite one topic: find articles, summarize, post to wall, trigger short."""
+async def _auto_rewrite_one(topic, slot_label, used_urls=None):
+    """Rewrite one topic: find articles, summarize, post to wall, trigger short.
+    used_urls: shared set to avoid duplicate articles across topics."""
     from urllib.parse import quote as _q
     items = _search_all(topic, limit=6)
+    # Skip URLs already used by another topic
+    if used_urls is not None:
+        filtered = [it for it in items if it.get('url') not in used_urls]
+        if filtered:
+            items = filtered
     if not items:
         # fallback: Google RSS
         try:
@@ -1348,6 +1354,8 @@ async def _auto_rewrite_one(topic, slot_label):
     item = items[0]  # best match
     url = item.get('url', '')
     title = item.get('title', topic)
+    if url and used_urls is not None:
+        used_urls.add(url)
     if not url.startswith('http'):
         return False
     
@@ -1399,34 +1407,46 @@ async def _auto_rewrite_one(topic, slot_label):
     return True
 
 async def _do_scheduled_run(slot_label):
-    """Main scheduled run: top 3 hot topics + latest articles."""
+    """Main scheduled run: exactly 3 posts from 3 different HOT topics, no duplicates."""
     print(f"[auto] Starting scheduled rewrite for {slot_label}")
     
-    # 1) Get top 3 hot topics
-    topics = _get_hot_topics()[:3]
-    job_topics = [t['topic'] for t in topics if t.get('topic')]
+    # Get top hot topics, skip duplicates
+    all_topics = _get_hot_topics()
+    seen_topics = set()
+    unique_topics = []
+    for t in all_topics:
+        kw = t.get('topic', '').lower().strip()
+        if kw and len(kw) > 5 and kw not in seen_topics:
+            is_dup = False
+            for s in seen_topics:
+                # Check if one topic is substring of another
+                if kw in s or s in kw:
+                    is_dup = True
+                    break
+            if not is_dup:
+                seen_topics.add(kw)
+                unique_topics.append(t)
+        if len(unique_topics) >= 3:
+            break
     
-    # 2) Add latest articles from RSS
-    for feed_url in ['https://vnexpress.net/rss/tin-moi-nhat.rss',
-                     'https://dantri.com.vn/rss/home.rss',
-                     'https://vietnamnet.vn/rss/tin-moi-nhat.rss']:
-        try:
-            r = req.get(feed_url, headers={'User-Agent':'Mozilla/5.0'}, timeout=5)
-            r.encoding = 'utf-8'
-            soup = BeautifulSoup(r.text, 'xml')
-            for it in soup.find_all('item')[:3]:
-                t = _clean(it.find('title').get_text(' ',strip=True) if it.find('title') else '')
-                if t and len(t) > 10 and t[:50] not in ' '.join(job_topics):
-                    job_topics.append(t[:60])
-        except: pass
+    job_topics = [t['topic'] for t in unique_topics[:3] if t.get('topic')]
+    if not job_topics:
+        print(f"[auto] No hot topics found, skipping")
+        return
+    
+    print(f"[auto] Running 3 topics: {job_topics}")
     
-    # 3) Execute jobs
+    # Track used URLs to avoid cross-topic duplicates
+    _used_urls = set()
     results = []
-    for jt in job_topics[:6]:
+    
+    for jt in job_topics:
         try:
-            ok = await asyncio.wait_for(_auto_rewrite_one(jt, slot_label), timeout=90)
+            ok = await asyncio.wait_for(_auto_rewrite_one(jt, slot_label, _used_urls), timeout=120)
             results.append((jt, ok))
-        except: results.append((jt, False))
+        except Exception as e:
+            print(f"[auto] Error on '{jt}': {e}")
+            results.append((jt, False))
         await asyncio.sleep(2)
     
     # Log