Fix rewrite slide storage and JSON error handling (#46)

- Fix rewrite slide storage and JSON error handling (c80226615706151e20325d12881a6a8763664df2)

rewrite_slide.py

@@ -2,63 +2,88 @@
 from main import app
 from fastapi import Request
 from fastapi.responses import JSONResponse
-import requests, re, time, random, json, os
+import requests, re, time, random, json, os, threading
 from bs4 import BeautifulSoup
-from urllib.parse import quote
+from urllib.parse import quote, urlparse
 
 UA = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36', 'Accept-Language': 'vi-VN,vi;q=0.9'}
 
-try:
-    from main import _load_wall, _save_wall
-except:
-    _data_dir = "/data" if os.path.isdir("/data") else "/app/data"
-    _wall_file = os.path.join(_data_dir, "wall_posts.json")
-    def _load_wall():
-        try:
-            if os.path.exists(_wall_file):
-                with open(_wall_file, 'r', encoding='utf-8') as f: return json.load(f)
-        except: pass
-        return []
-    def _save_wall(posts):
-        try:
-            os.makedirs(os.path.dirname(_wall_file), exist_ok=True)
-            with open(_wall_file+'.tmp', 'w', encoding='utf-8') as f: json.dump(posts[:100], f, ensure_ascii=False)
-            os.replace(_wall_file+'.tmp', _wall_file)
-        except: pass
+_DATA_DIR = "/data" if os.path.isdir("/data") else os.path.join(os.path.dirname(os.path.abspath(__file__)), "data")
+_WALL_FILE = os.path.join(_DATA_DIR, "wall_posts.json")
+_WALL_LOCK = threading.Lock()
 
+def _safe_json_load(path, default):
+    try:
+        path = str(path)
+        with open(path, 'r', encoding='utf-8') as f:
+            return json.load(f)
+    except Exception:
+        return default
+
+def _safe_json_dump(path, data):
+    try:
+        path = str(path)
+        os.makedirs(os.path.dirname(path), exist_ok=True)
+        tmp = path + '.tmp'
+        with open(tmp, 'w', encoding='utf-8') as f:
+            json.dump(data, f, ensure_ascii=False)
+        os.replace(tmp, path)
+    except Exception:
+        pass
+
+def _load_wall():
+    with _WALL_LOCK:
+        posts = _safe_json_load(_WALL_FILE, [])
+        return posts if isinstance(posts, list) else []
+
+def _save_wall(posts):
+    with _WALL_LOCK:
+        if not isinstance(posts, list):
+            posts = []
+        _safe_json_dump(_WALL_FILE, posts[:100])
 
 def _clean(s): return re.sub(r'\s+', ' ', str(s or '')).strip()
 
 
 def _scrape_article_full(url):
-    """Scrape article: extract paragraphs + ALL images."""
+    """Scrape article: extract paragraphs + images from common Vietnamese news layouts."""
     try:
+        parsed = urlparse(url)
+        if not parsed.netloc:
+            return None
         r = requests.get(url, headers=UA, timeout=15, allow_redirects=True)
         r.encoding = 'utf-8'
+        if r.status_code != 200 or not r.text:
+            return None
         soup = BeautifulSoup(r.text, 'lxml')
-        for tag in soup.find_all(['script', 'style', 'nav', 'footer', 'aside', 'form']): tag.decompose()
+        for tag in soup.find_all(['script', 'style', 'nav', 'footer', 'aside', 'form', 'noscript', 'iframe', '.ads', '.ad', '.banner-ads', '.fb-comments', '.fb-root', '.social-share', '.related-news', '.breadcrumb']):
+            tag.decompose()
         
         # Title
         h1 = soup.find('h1')
         ogt = soup.find('meta', property='og:title')
-        title = (h1.get_text(strip=True) if h1 else '') or (ogt.get('content', '') if ogt else '')
+        title = _clean((h1.get_text(strip=True) if h1 else '') or (ogt.get('content', '') if ogt else ''))
         
         # OG image
         ogi = soup.find('meta', property='og:image')
-        og_img = ogi.get('content', '') if ogi else ''
+        og_img = _clean(ogi.get('content', '') if ogi else '')
         if og_img and og_img.startswith('//'): og_img = 'https:' + og_img
         
         # Find content block
         block = None
-        for sel in ['article', '.singular-content', '.detail-content', '.fck_detail', '.content-detail', '.knc-content', 'main', '.cms-body', '.article__body']:
+        for sel in ['article', '.singular-content', '.detail-content', '.fck_detail', '.content-detail', '.knc-content', 'main', '.cms-body', '.article__body', '.post-content', '.entry-content']:
             el = soup.select_one(sel)
-            if el and len(el.find_all('p')) >= 2: block = el; break
-        if not block: block = soup.body or soup
+            if el and len(el.find_all('p')) >= 2:
+                block = el
+                break
+        if not block:
+            block = soup.body or soup
         
         # Extract paragraphs and images IN ORDER
         paragraphs = []
         images = []
         seen_imgs = set()
+        seen_text = set()
         
         if og_img and og_img not in seen_imgs:
             images.append(og_img)
@@ -67,20 +92,22 @@ def _scrape_article_full(url):
         for el in block.find_all(['p', 'h2', 'h3', 'figure', 'img'], recursive=True):
             if el.name == 'p':
                 t = _clean(el.get_text(strip=True))
-                if t and len(t) > 40:
+                if t and len(t) > 40 and t not in seen_text:
                     paragraphs.append(t)
+                    seen_text.add(t)
             elif el.name in ('figure', 'img'):
                 im = el if el.name == 'img' else el.find('img')
                 if im:
-                    src = im.get('data-src') or im.get('src') or im.get('data-original') or ''
+                    src = im.get('data-src') or im.get('src') or im.get('data-original') or im.get('data-lazy') or ''
                     if src and 'base64' not in src:
                         if src.startswith('//'): src = 'https:' + src
+                        if src.startswith('/'): src = parsed.scheme + '://' + parsed.netloc + src
                         if src not in seen_imgs:
                             images.append(src)
                             seen_imgs.add(src)
         
-        return {'title': _clean(title), 'paragraphs': paragraphs, 'images': images, 'og_img': og_img}
-    except Exception as e:
+        return {'title': title, 'paragraphs': paragraphs, 'images': images, 'og_img': og_img}
+    except Exception:
         return None
 
 
@@ -123,7 +150,6 @@ def _extract_key_points(paragraphs, max_points=5):
     
     return points
 
-
 @app.post("/api/rewrite_slide")
 async def api_rewrite_slide(request: Request):
     """
@@ -131,10 +157,13 @@ async def api_rewrite_slide(request: Request):
     - Extract key points from article (1 sentence each, full and complete)
     - Pair each point with an image from the article
     - Return as slides array for frontend to display
-    - Save to Tường AI
+    - Save to Tường AI using the same file as app_v2_entry.py
     NO AI NEEDED - instant response.
     """
-    body = await request.json()
+    try:
+        body = await request.json()
+    except Exception:
+        return JSONResponse({"error": "JSON request lỗi"}, status_code=400)
     url = _clean(body.get("url", ""))
     context = body.get("context", "")