更新

backend/app/services/whisper_service.py

@@ -212,6 +212,43 @@ class WhisperService:
 
         return result
 
+    async def transcribe(self, audio_path: str) -> str:
+        """
+        仅转录文本（用于提取文案）
+
+        Args:
+            audio_path: 音频/视频文件路径
+
+        Returns:
+            纯文本内容
+        """
+        import asyncio
+
+        def _do_transcribe_text():
+            model = self._load_model()
+            logger.info(f"Extracting script from: {audio_path}")
+
+            # 转录 (无需字级时间戳)
+            segments_iter, _ = model.transcribe(
+                audio_path,
+                language="zh",
+                word_timestamps=False,
+                vad_filter=True,
+            )
+
+            text_parts = []
+            for segment in segments_iter:
+                text_parts.append(segment.text.strip())
+
+            full_text = " ".join(text_parts)
+            logger.info(f"Extracted text length: {len(full_text)}")
+            return full_text
+
+        # 在线程池中执行
+        loop = asyncio.get_event_loop()
+        result = await loop.run_in_executor(None, _do_transcribe_text)
+        return result
+
     async def check_health(self) -> dict:
         """检查服务健康状态"""
         try: