Update app.py

app.py

@@ -192,9 +192,117 @@ Important rules:
                 if current_scene:
                     english_scenes.append(current_scene.strip())
 
+            # 중국어 번역 생성
+            chinese_prompt = f"""
+将以下韩语语音/对话转录翻译成中文，保持完全相同的时间戳。
+只翻译实际的语音内容，不要描述：
+
+{chr(10).join(korean_scenes)}
+
+重要规则：
+- 保持格式完全相同：**[MM:SS-MM:SS]**: "语音的中文翻译"
+- 对于像(배경음악)这样的非语音部分，翻译为(背景音乐)
+- 对于(...)保持原样
+- 对于像[이도학 교수]这样的采访标签，翻译为[李道学教授]
+- 对于[자막]，翻译为[字幕]
+- 保留实际语音的引号
+- 不要添加任何场景描述或解释
+"""
+            chinese_response = self.gemini_model.generate_content(chinese_prompt)
+            
+            chinese_scenes = []
+            if chinese_response and chinese_response.text:
+                lines = chinese_response.text.split("\n")
+                current_scene = ""
+
+                for line in lines:
+                    line = line.strip()
+                    if line.startswith("**[") and "]**:" in line:
+                        if current_scene:
+                            chinese_scenes.append(current_scene.strip())
+                        current_scene = line
+                    elif current_scene:
+                        current_scene += "\n" + line
+
+                if current_scene:
+                    chinese_scenes.append(current_scene.strip())
+
+            # 태국어 번역 생성
+            thai_prompt = f"""
+แปลคำบรรยายเสียง/บทสนทนาภาษาเกาหลีต่อไปนี้เป็นภาษาไทย โดยคงรูปแบบเวลาเดิมไว้
+แปลเฉพาะเนื้อหาเสียงจริงเท่านั้น ไม่ต้องบรรยาย:
+
+{chr(10).join(korean_scenes)}
+
+กฎสำคัญ:
+- คงรูปแบบเดิมไว้: **[MM:SS-MM:SS]**: "คำแปลภาษาไทยของเสียงพูด"
+- สำหรับส่วนที่ไม่ใช่เสียงพูด เช่น (배경음악) แปลเป็น (เพลงประกอบ)
+- สำหรับ (...) ให้คงเดิม
+- สำหรับป้ายสัมภาษณ์ เช่น [이도학 교수] แปลเป็น [ศาสตราจารย์ อี โด-ฮัก]
+- สำหรับ [자막] แปลเป็น [คำบรรยาย]
+- คงเครื่องหมายคำพูดสำหรับเสียงพูดจริง
+- ไม่ต้องเพิ่มคำบรรยายฉากหรือคำอธิบายใดๆ
+"""
+            thai_response = self.gemini_model.generate_content(thai_prompt)
+            
+            thai_scenes = []
+            if thai_response and thai_response.text:
+                lines = thai_response.text.split("\n")
+                current_scene = ""
+
+                for line in lines:
+                    line = line.strip()
+                    if line.startswith("**[") and "]**:" in line:
+                        if current_scene:
+                            thai_scenes.append(current_scene.strip())
+                        current_scene = line
+                    elif current_scene:
+                        current_scene += "\n" + line
+
+                if current_scene:
+                    thai_scenes.append(current_scene.strip())
+
+            # 러시아어 번역 생성
+            russian_prompt = f"""
+Переведите следующую корейскую транскрипцию речи/диалога на русский язык, сохраняя точно такие же временные метки.
+Переводите ТОЛЬКО фактическое содержание речи, а не описания:
+
+{chr(10).join(korean_scenes)}
+
+Важные правила:
+- Сохраняйте формат точно таким же: **[MM:SS-MM:SS]**: "Русский перевод речи"
+- Для частей без речи, таких как (배경음악), переведите как (фоновая музыка)
+- Для (...) оставьте как есть
+- Для тегов интервью, таких как [이도학 교수], переведите как [Профессор Ли До Хак]
+- Для [자막] переведите как [Субтитры]
+- Сохраняйте кавычки для фактической речи
+- НЕ добавляйте никаких описаний сцен или объяснений
+"""
+            russian_response = self.gemini_model.generate_content(russian_prompt)
+            
+            russian_scenes = []
+            if russian_response and russian_response.text:
+                lines = russian_response.text.split("\n")
+                current_scene = ""
+
+                for line in lines:
+                    line = line.strip()
+                    if line.startswith("**[") and "]**:" in line:
+                        if current_scene:
+                            russian_scenes.append(current_scene.strip())
+                        current_scene = line
+                    elif current_scene:
+                        current_scene += "\n" + line
+
+                if current_scene:
+                    russian_scenes.append(current_scene.strip())
+
             return {
                 "korean": korean_scenes if korean_scenes else self.generate_scene_breakdown_fallback(video_info)["korean"],
-                "english": english_scenes if english_scenes else self.generate_scene_breakdown_fallback(video_info)["english"]
+                "english": english_scenes if english_scenes else self.generate_scene_breakdown_fallback(video_info)["english"],
+                "chinese": chinese_scenes if chinese_scenes else [],
+                "thai": thai_scenes if thai_scenes else [],
+                "russian": russian_scenes if russian_scenes else []
             }
 
         except Exception as e:
@@ -202,7 +310,7 @@ Important rules:
             return self.generate_scene_breakdown_fallback(video_info)
 
     # ---------------------------------------------------------
-    # Fallback 음성/대사 추출 (한글/영어)
+    # Fallback 음성/대사 추출 (5개 언어)
     # ---------------------------------------------------------
     def generate_scene_breakdown_fallback(self, video_info):
         duration = video_info.get("duration", 0)
@@ -213,7 +321,10 @@ Important rules:
         if not duration:
             return {
                 "korean": ["**[재생시간 알 수 없음]**: 타임스탬프를 생성할 수 없습니다"],
-                "english": ["**[Duration Unknown]**: Unable to generate timestamped breakdown"]
+                "english": ["**[Duration Unknown]**: Unable to generate timestamped breakdown"],
+                "chinese": ["**[持续时间未知]**: 无法生成带时间戳的分解"],
+                "thai": ["**[ไม่ทราบระยะเวลา]**: ไม่สามารถสร้างการแบ่งส่วนตามเวลาได้"],
+                "russian": ["**[Продолжительность неизвестна]**: Невозможно создать временную разбивку"]
             }
 
         # 비디오 타입에 따른 대사 템플릿
@@ -228,8 +339,11 @@ Important rules:
 
         korean_scenes = []
         english_scenes = []
+        chinese_scenes = []
+        thai_scenes = []
+        russian_scenes = []
+        
         num_segments = min(duration // segment_length + 1, 20)
-        video_type = self.detect_video_type_detailed(title, description)
 
         for i in range(num_segments):
             start_time = i * segment_length
@@ -240,19 +354,31 @@ Important rules:
             
             # 음성 추출이 불가능한 경우의 기본 템플릿
             if i == 0:
-                korean_desc = f"(음성 추출을 위해 Gemini API가 필요합니다. 인트로 부분...)"
-                english_desc = f"(Gemini API required for speech extraction. Intro section...)"
+                korean_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (음성 추출을 위해 Gemini API가 필요합니다. 인트로 부분...)")
+                english_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (Gemini API required for speech extraction. Intro section...)")
+                chinese_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (需要Gemini API进行语音提取。介绍部分...)")
+                thai_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (ต้องใช้ Gemini API สำหรับการดึงเสียง ส่วนเปิดตัว...)")
+                russian_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (Требуется Gemini API для извлечения речи. Вступительная часть...)")
             elif i == num_segments - 1:
-                korean_desc = f"(아웃트로 부분...)"
-                english_desc = f"(Outro section...)"
+                korean_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (아웃트로 부분...)")
+                english_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (Outro section...)")
+                chinese_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (结尾部分...)")
+                thai_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (ส่วนจบ...)")
+                russian_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (Заключительная часть...)")
             else:
-                korean_desc = f"(본문 내용...)"
-                english_desc = f"(Main content...)"
-            
-            korean_scenes.append(f"**[{start_fmt}-{end_fmt}]**: {korean_desc}")
-            english_scenes.append(f"**[{start_fmt}-{end_fmt}]**: {english_desc}")
-
-        return {"korean": korean_scenes, "english": english_scenes}
+                korean_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (본문 내용...)")
+                english_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (Main content...)")
+                chinese_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (主要内容...)")
+                thai_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (เนื้อหาหลัก...)")
+                russian_scenes.append(f"**[{start_fmt}-{end_fmt}]**: (Основное содержание...)")
+
+        return {
+            "korean": korean_scenes, 
+            "english": english_scenes,
+            "chinese": chinese_scenes,
+            "thai": thai_scenes,
+            "russian": russian_scenes
+        }
 
     # ---------------------------------------------------------
     # 비디오 유형 감지(상세)
@@ -463,7 +589,7 @@ Important rules:
         return str(num)
 
     # ---------------------------------------------------------
-    # 최종 리포트 생성 (한글/영어 분리)
+    # 최종 리포트 생성 (5개 언어)
     # ---------------------------------------------------------
     def format_video_info(self, video_info):
         if not video_info:
@@ -484,6 +610,9 @@ Important rules:
         scene_data = self.generate_scene_breakdown_gemini(video_info)
         korean_scenes = scene_data.get("korean", [])
         english_scenes = scene_data.get("english", [])
+        chinese_scenes = scene_data.get("chinese", [])
+        thai_scenes = scene_data.get("thai", [])
+        russian_scenes = scene_data.get("russian", [])
         
         vtype = self.detect_video_type(title, video_info.get("description", ""))
         bgm = self.detect_background_music(video_info)
@@ -523,10 +652,17 @@ Important rules:
 {'─'*30}
 {chr(10).join(english_scenes)}
 
-📝 설명 미리보기 / DESCRIPTION PREVIEW
-{'─'*25}
-{video_info.get('description', 'No description available')[:500]}
-{'...(생략/truncated)' if len(video_info.get('description', '')) > 500 else ''}
+🎙️ 音频/对话提取 (中文) / SPEECH/DIALOGUE EXTRACTION (CHINESE)
+{'─'*30}
+{chr(10).join(chinese_scenes) if chinese_scenes else "(중국어 번역 없음 / No Chinese translation available)"}
+
+🎙️ การดึงเสียง/บทสนทนา (ไทย) / SPEECH/DIALOGUE EXTRACTION (THAI)
+{'─'*30}
+{chr(10).join(thai_scenes) if thai_scenes else "(태국어 번역 없음 / No Thai translation available)"}
+
+🎙️ Извлечение речи/диалога (Русский) / SPEECH/DIALOGUE EXTRACTION (RUSSIAN)
+{'─'*30}
+{chr(10).join(russian_scenes) if russian_scenes else "(러시아어 번역 없음 / No Russian translation available)"}
 
 {'='*50}
 📊 **분석 완료/Analysis completed:** {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}