Update app.py

app.py

@@ -133,14 +133,13 @@ async def get_transcript(youtube_url: str):
         ydl_opts = {
             'skip_download': True,
             'writesubtitles': True,
-            'writeautomaticsub': True,  # 启用自动字幕作为备选
-            'subtitleslangs': ['original'],  # 优先原始语言
+            'writeautomaticsub': True,
+            'subtitleslangs': ['all'],  # 获取所有字幕
             'subtitlesformat': 'best',
             'outtmpl': '%(id)s.%(ext)s',
             'noplaylist': True,
             'cookiefile': "firefox-cookies.txt",
-            'compat_opts': {'no-sub-translate'},  # 保持原始语言
-            'retries': 3,
+            'ignoreerrors': True,
         }
 
         env_to_cookies_from_env("firefox-cookies.txt")
@@ -149,55 +148,43 @@ async def get_transcript(youtube_url: str):
             info = ydl.extract_info(youtube_url, download=True)
             video_id = info['id']
             
-            # 精确获取原始语言（修复YouTube API返回不一致问题）
-            original_lang = info.get('original_lang') 
-            if not original_lang:
-                original_lang = info.get('subtitles', {}).get('original', [{}])[0].get('language')
-            original_lang = original_lang or 'en'
-
-            # 扩展匹配模式（包含自动生成字幕）
-            subtitle_pattern = re.compile(
-                rf"^{re.escape(video_id)}\.(a\.)?{re.escape(original_lang)}\..+$"
-            )
+            # 合并手动和自动字幕列表
+            all_subs = list(info.get('subtitles', {}).keys()) + list(info.get('automatic_captions', {}).keys())
             
-            # 查找所有可能匹配的文件
-            subtitle_files = [
-                f for f in os.listdir('.') 
-                if subtitle_pattern.match(f)
-            ]
-
-            logger.info(f"匹配到的字幕文件: {subtitle_files}")
-
-            if subtitle_files:
-                # 优先手动上传的字幕（非自动生成）
-                preferred_files = [f for f in subtitle_files if 'a.' not in f]
-                selected_file = preferred_files[0] if preferred_files else subtitle_files[0]
-
-                with open(selected_file, 'r', encoding='utf-8') as f:
+            # 过滤有效语言代码（排除翻译字幕如en-zh-Hant）
+            valid_langs = [lang for lang in all_subs if '-' not in lang.split('.')[-1]]
+            
+            if not valid_langs:
+                return {"error": "无可用字幕"}
+            
+            # 直接取第一个有效语言
+            target_lang = valid_langs[0]
+            
+            # 构建字幕文件名模式
+            subtitle_file = f"{video_id}.{target_lang}.vtt"  # 优先vtt格式
+            if not os.path.exists(subtitle_file):
+                subtitle_file = f"{video_id}.{target_lang}.srt"
+            
+            if os.path.exists(subtitle_file):
+                with open(subtitle_file, 'r', encoding='utf-8') as f:
                     content = f.read()
-
-                # 清理字幕内容
+                
+                # 基础清理
                 clean_content = re.sub(
-                    r'^\d+\n\d{2}:\d{2}:\d{2}[\.,]\d{3} --> \d{2}:\d{2}:\d{2}[\.,]\d{3}\n',
+                    r'\d{2}:\d{2}:\d{2}[,.]\d{3}.*?\n', 
                     '', 
-                    content,
-                    flags=re.MULTILINE
+                    content
                 ).strip()
-
+                
                 return {
                     "transcript": clean_content,
-                    "language": original_lang,
-                    "sub_type": "manual" if 'a.' not in selected_file else "auto-generated"
+                    "detected_lang": target_lang
                 }
 
-            return {
-                "error": "未找到原始语言字幕",
-                "available_langs": list(info.get('subtitles', {}).keys()) 
-                                 + list(info.get('automatic_captions', {}).keys())
-            }
+            return {"error": "字幕文件未找到"}
 
     except Exception as e:
-        return {"error": f"处理失败: {str(e)}"}
+        return {"error": str(e)}