Restrict hashtags and topic sources to selected Vietnamese sources

ai_runtime_final6.py

@@ -1004,3 +1004,140 @@ async def index_final6_clean_links():
     html=f5.f4.f3.f2.f1._load_index_html()
     body=getattr(rt.old,'PATCH_INJECT','')+f5.f4.f3.f2.f1.FINAL_INJECT+f5.f4.f3.FINAL3_INJECT+f5.f4.FINAL4_INJECT+f5.FINAL5_INJECT+FINAL6_INJECT+globals().get('FINAL6_FAST_HOME_INJECT','')+globals().get('FINAL6E_INJECT','')+FINAL6F_INJECT
     return HTMLResponse(html.replace('</body>',body+'\n</body>') if '</body>' in html else html+body)
+
+
+# ===== FINAL6G: SELECTED FAST SOURCES =====
+# Restrict hot hashtags + topic articles to the requested sources only:
+# Thethaovanhoa, Dantri, VTV, VnExpress, Vatvostudio, GenK, VNReview.
+SELECTED_SOURCE_FEEDS=[
+    ('VnExpress','https://vnexpress.net/rss/tin-moi-nhat.rss'),
+    ('VnExpress Thời sự','https://vnexpress.net/rss/thoi-su.rss'),
+    ('VnExpress Thế giới','https://vnexpress.net/rss/the-gioi.rss'),
+    ('VnExpress Kinh doanh','https://vnexpress.net/rss/kinh-doanh.rss'),
+    ('VnExpress Công nghệ','https://vnexpress.net/rss/so-hoa.rss'),
+    ('VnExpress Thể thao','https://vnexpress.net/rss/the-thao.rss'),
+    ('Dân trí','https://dantri.com.vn/rss/home.rss'),
+    ('Dân trí Xã hội','https://dantri.com.vn/rss/xa-hoi.rss'),
+    ('Dân trí Kinh doanh','https://dantri.com.vn/rss/kinh-doanh.rss'),
+    ('Dân trí Thể thao','https://dantri.com.vn/rss/the-thao.rss'),
+    ('Dân trí Công nghệ','https://dantri.com.vn/rss/suc-manh-so.rss'),
+    ('VTV','https://vtv.vn/rss/trang-chu.rss'),
+    ('VTV Thời sự','https://vtv.vn/rss/thoi-su.rss'),
+    ('VTV Công nghệ','https://vtv.vn/rss/cong-nghe.rss'),
+    ('Thể thao văn hóa','https://thethaovanhoa.vn/rss/home.rss'),
+    ('Thể thao văn hóa Bóng đá','https://thethaovanhoa.vn/rss/bong-da.rss'),
+    ('GenK','https://genk.vn/home.rss'),
+    ('GenK AI','https://genk.vn/ai.rss'),
+    ('VNReview','https://vnreview.vn/rss/home.rss'),
+    ('VNReview Công nghệ','https://vnreview.vn/rss/cong-nghe.rss'),
+]
+SELECTED_HOMEPAGES=[
+    ('VTV','https://vtv.vn/'),
+    ('Thể thao văn hóa','https://thethaovanhoa.vn/'),
+    ('GenK','https://genk.vn/'),
+    ('VNReview','https://vnreview.vn/'),
+    ('Vatvostudio','https://vatvostudio.vn/'),
+]
+SELECTED_DOMAINS=['vnexpress.net','dantri.com.vn','vtv.vn','thethaovanhoa.vn','genk.vn','vnreview.vn','vatvostudio.vn']
+
+def _selected_fetch_rss(feed_name, feed_url, max_items=10):
+    items=[]
+    try:
+        r=requests.get(feed_url,headers=UA,timeout=4);r.encoding='utf-8'
+        if r.status_code>=400:return []
+        soup=BeautifulSoup(r.text,'xml')
+        for it in soup.find_all('item')[:max_items]:
+            title=clean(it.find('title').get_text(' ',strip=True) if it.find('title') else '')
+            link=clean(it.find('link').get_text(strip=True) if it.find('link') else '')
+            desc_raw=it.find('description').get_text(' ',strip=True) if it.find('description') else ''
+            ds=BeautifulSoup(desc_raw,'lxml')
+            desc=clean(ds.get_text(' ',strip=True))
+            if title and link:
+                items.append({'title':title,'url':link,'source':feed_name,'snippet':desc})
+    except Exception:pass
+    return items
+
+def _selected_scrape_homepage(name, url, max_items=10):
+    items=[];seen=set()
+    try:
+        r=requests.get(url,headers=UA,timeout=4);r.encoding='utf-8'
+        if r.status_code>=400:return []
+        soup=BeautifulSoup(r.text,'lxml')
+        base=url.rstrip('/')
+        for a in soup.find_all('a',href=True):
+            href=a.get('href','').strip();title=clean(a.get('title','') or a.get_text(' ',strip=True))
+            if not href or not title or len(title)<18:continue
+            if href.startswith('/'):
+                p=urlparse(url); href=f'{p.scheme}://{p.netloc}{href}'
+            if not href.startswith('http') or href in seen:continue
+            dom=_domain(href)
+            if not any(d in dom for d in SELECTED_DOMAINS):continue
+            if any(x in href.lower() for x in ['#','javascript:','facebook','youtube','tiktok']):continue
+            seen.add(href)
+            items.append({'title':title,'url':href,'source':name,'snippet':''})
+            if len(items)>=max_items:break
+    except Exception:pass
+    return items
+
+def _fast_rss_pool():
+    now=time.time();key='selected_fast_pool'
+    if key in _FAST_TOPIC_CACHE and now-_FAST_TOPIC_CACHE[key]['t']<600:return _FAST_TOPIC_CACHE[key]['d']
+    pool=[];seen=set()
+    # RSS first: fast and reliable.
+    for name,url in SELECTED_SOURCE_FEEDS:
+        for it in _selected_fetch_rss(name,url,10):
+            if it['url'] not in seen:
+                seen.add(it['url']);pool.append(it)
+    # Homepage fallback for sources with weak/no RSS, especially Vatvostudio.
+    for name,url in SELECTED_HOMEPAGES:
+        for it in _selected_scrape_homepage(name,url,10):
+            if it['url'] not in seen:
+                seen.add(it['url']);pool.append(it)
+    _FAST_TOPIC_CACHE[key]={'t':now,'d':pool}
+    return pool
+
+def _hot_topics():
+    now=time.time()
+    if _HOT_CACHE['d'] and now-_HOT_CACHE['t']<600:return _HOT_CACHE['d']
+    pool=_fast_rss_pool()
+    freq={};display={}
+    for it in pool[:220]:
+        title=re.sub(r'\s+-\s+.*$','',it.get('title',''))
+        kws=[]
+        for m in re.findall(r'([A-ZĐÀ-Ỹ][A-Za-zÀ-ỹ0-9]+(?:\s+[A-ZĐÀ-ỸA-Za-zÀ-ỹ0-9][A-Za-zÀ-ỹ0-9]+){1,4})',title):
+            if len(m)>=6:kws.append(m)
+        kws+=_keywords_from_title(title)
+        for kw in kws[:5]:
+            words=[w for w in clean(kw).split() if w.lower() not in STOP_WORDS]
+            if len(words)<2:continue
+            kw=' '.join(words[:5])
+            if len(kw)<6 or len(kw)>55:continue
+            key=kw.lower();freq[key]=freq.get(key,0)+1;display[key]=kw
+    topics=[];seen=set()
+    for key,_ in sorted(freq.items(),key=lambda x:x[1],reverse=True):
+        kw=display[key]
+        if key in seen:continue
+        seen.add(key);topics.append({'label':'#'+re.sub(r'\s+','',kw.title()),'topic':kw})
+        if len(topics)>=24:break
+    for kw in ['AI tại Việt Nam','Công nghệ Việt Nam','VTV thời sự','VnExpress kinh doanh','Dân trí xã hội','GenK AI','VNReview công nghệ','Vatvostudio smartphone','Thể thao văn hóa World Cup']:
+        if kw.lower() not in seen:topics.append({'label':'#'+re.sub(r'\s+','',kw.title()),'topic':kw})
+    _HOT_CACHE.update({'t':now,'d':topics[:24]})
+    return _HOT_CACHE['d']
+
+def _candidate_urls(topic):
+    seen=set();items=[]
+    scored=[]
+    for it in _fast_rss_pool():
+        sc=_fast_score(topic,it)
+        if sc>0:scored.append((sc,it))
+    for sc,it in sorted(scored,key=lambda x:(x[0],len(x[1].get('snippet',''))),reverse=True)[:14]:
+        if it['url'] not in seen:
+            seen.add(it['url']);items.append(it)
+    # Search only selected sources when RSS lacks matches.
+    if len(items)<6:
+        for dom in SELECTED_DOMAINS:
+            for it in _ddg_search(f'{topic} site:{dom}',4):
+                if it['url'] not in seen:
+                    seen.add(it['url']);items.append(it)
+            if len(items)>=12:break
+    return items[:20]