tingwu_new

backend/ai_services/services.py

@@ -103,6 +103,11 @@ class AliyunTingwuService:
                 "Transcoding": {
                     "TargetAudioFormat": "mp3"
                 },
+                "Transcription": {
+                    "DiarizationEnabled": True, 
+                    "ChannelId": 0
+                },
+                "TranscriptionEnabled": True,
                 "AutoChaptersEnabled": True,
                 "SummarizationEnabled": True,
                 "Summarization": {

backend/ai_services/views.py

@@ -146,8 +146,10 @@ class TranscriptionTaskViewSet(viewsets.ModelViewSet):
         """
         task = self.get_object()
         
-        # 如果任务已经完成或失败，直接返回当前状态
+        # 如果任务已经完成或失败，但逐字稿为空，允许重新刷新
-        if task.status in [TranscriptionTask.Status.SUCCEEDED, TranscriptionTask.Status.FAILED]:
+        if task.status == TranscriptionTask.Status.SUCCEEDED and not task.transcription:
+            pass # 继续执行刷新逻辑
+        elif task.status in [TranscriptionTask.Status.SUCCEEDED, TranscriptionTask.Status.FAILED]:
             serializer = self.get_serializer(task)
             return Response(serializer.data)
 
@@ -191,9 +193,11 @@ class TranscriptionTaskViewSet(viewsets.ModelViewSet):
                 
                 # 解析结果
                 task_result = data_obj.get('Result', {})
+                logger.info(f"Task result keys: {task_result.keys()}")
                 
                 # 提取逐字稿
                 transcription_data = task_result.get('Transcription', {})
+                logger.info(f"Raw transcription data type: {type(transcription_data)}")
                 
                 # 如果是 URL (字符串)，尝试下载内容
                 if isinstance(transcription_data, str) and transcription_data.startswith('http'):
@@ -203,6 +207,7 @@ class TranscriptionTaskViewSet(viewsets.ModelViewSet):
                         t_resp = requests.get(transcription_data)
                         if t_resp.status_code == 200:
                             transcription_data = t_resp.json()
+                            logger.info(f"Downloaded transcription keys: {transcription_data.keys() if isinstance(transcription_data, dict) else 'Not a dict'}")
                         else:
                             logger.warning(f"Failed to download transcription: {t_resp.status_code}")
                             transcription_data = {}
@@ -218,6 +223,7 @@ class TranscriptionTaskViewSet(viewsets.ModelViewSet):
                         t_resp = requests.get(url)
                         if t_resp.status_code == 200:
                             transcription_data = t_resp.json()
+                            logger.info(f"Downloaded transcription keys: {transcription_data.keys() if isinstance(transcription_data, dict) else 'Not a dict'}")
                      except Exception as e:
                          logger.error(f"Error downloading transcription nested url: {e}")
                 
@@ -225,12 +231,38 @@ class TranscriptionTaskViewSet(viewsets.ModelViewSet):
                     # 尝试多种可能的路径提取句子
                     # 1. 直接在根目录: {"Sentences": [...]}
                     # 2. 在 Transcription 字段下: {"Transcription": {"Sentences": [...]}}
+                    # 3. 可能是 Paragraphs 结构
                     sentences = transcription_data.get('Sentences', [])
                     if not sentences and 'Transcription' in transcription_data:
                          sentences = transcription_data['Transcription'].get('Sentences', [])
                     
-                    full_text = " ".join([s.get('Text', '') for s in sentences])
+                    if sentences:
-                    task.transcription = full_text
+                        full_text = " ".join([s.get('Text', '') for s in sentences])
+                        task.transcription = full_text
+                    elif 'Paragraphs' in transcription_data:
+                        # 尝试从段落中提取
+                        paragraphs = transcription_data.get('Paragraphs', {})
+                        if isinstance(paragraphs, dict):
+                             # Paragraphs 可能是一个字典，包含 list
+                             para_list = paragraphs.get('Paragraphs', []) # 有时候是嵌套的
+                             if not para_list and isinstance(paragraphs, list):
+                                 para_list = paragraphs
+                        elif isinstance(paragraphs, list):
+                             para_list = paragraphs
+                        else:
+                             para_list = []
+                             
+                        texts = []
+                        for p in para_list:
+                            if 'Text' in p:
+                                texts.append(p['Text'])
+                            elif 'Sentences' in p:
+                                for s in p['Sentences']:
+                                    if 'Text' in s:
+                                        texts.append(s['Text'])
+                        task.transcription = "\n".join(texts)
+                    else:
+                        logger.warning(f"Could not find Sentences or Paragraphs in transcription data: {transcription_data.keys()}")
                 
                 # 提取总结
                 # 总结结果结构可能因配置不同而异，这里尝试获取摘要