Update app.py

app.py

@@ -8,8 +8,8 @@
 import uvicorn
 import logging
 import asyncio
-from dashscope.audio.omni import OmniRealtimeConversation, OmniRealtimeConfig, OmniRealtimeParam, OmniRealtimeCallback
-from dashscope.audio.omni import OmniRealtimeAudioFormat, OmniRealtimeModality, OmniRealtimeTranslationParam
+from dashscope.audio.qwen_omni import OmniRealtimeConversation, OmniRealtimeCallback
+from dashscope.audio.qwen_omni.omni_realtime import TranslationParams, MultiModality, AudioFormat
 
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
@@ -48,8 +48,8 @@ def on_event(self, message):
         except Exception as e:
             logger.error(f"Error processing message: {e}")
 
-    def on_close(self, code, reason):
-        logger.info(f"Connection closed: {code} - {reason}")
+    def on_close(self):
+        logger.info("Connection closed.")
 
 @app.get("/", response_class=HTMLResponse)
 async def get_index(request: Request):
@@ -64,47 +64,59 @@ async def websocket_endpoint(websocket: WebSocket):
         try:
             msg = json.loads(data)
             api_key = msg.get("api_key")
-            workspace_id = msg.get("workspace_id", "your_workspace_id")
+            workspace_id = msg.get("workspace_id")
             region = msg.get("region", "cn-beijing")
         except:
-            await websocket.send_text(json.dumps({"type": "error", "message": "Invalid API key format"}))
+            await websocket.send_text(json.dumps({"type": "error", "message": "Invalid config format"}))
             return
 
+        # 优先使用前端传入的 API Key，如果没有则从环境变量读取
         if not api_key:
             api_key = os.getenv("DASHSCOPE_API_KEY")
             if not api_key:
                 await websocket.send_text(json.dumps({"type": "error", "message": "API key is required"}))
                 return
 
-        param = OmniRealtimeParam.builder() \
-            .model("qwen3.5-livetranslate-flash-realtime") \
-            .url(f"wss://{workspace_id}.{region}.maas.aliyuncs.com/api-ws/v1/realtime") \
-            .build()
-
-        config = OmniRealtimeConfig.builder() \
-            .modalities([OmniRealtimeModality.AUDIO, OmniRealtimeModality.TEXT]) \
-            .translationConfig(
-                OmniRealtimeTranslationParam.builder()
-                    .language("auto")
-                    .build()
-            ) \
-            .inputAudioFormat(OmniRealtimeAudioFormat.PCM_16000HZ_MONO_16BIT) \
-            .outputAudioFormat(OmniRealtimeAudioFormat.PCM_24000HZ_MONO_16BIT) \
-            .enableTurnDetection(True) \
-            .build()
+        if not workspace_id:
+            await websocket.send_text(json.dumps({"type": "error", "message": "Workspace ID is required"}))
+            return
+
+        # 构建 WebSocket URL
+        ws_url = f"wss://{workspace_id}.{region}.maas.aliyuncs.com/api-ws/v1/realtime"
 
+        # 创建回调实例
         callback = TranslateCallback(websocket)
-        conversation = OmniRealtimeConversation(param, callback)
+
+        # 建立连接
+        conversation = OmniRealtimeConversation(
+            model='qwen3.5-livetranslate-flash-realtime',
+            url=ws_url,
+            api_key=api_key,
+            callback=callback
+        )
         conversation.connect()
-        conversation.updateSession(config)
+
+        # 配置会话参数
+        translation_params = TranslationParams(
+            language='auto'  # auto 表示自动检测源语言
+        )
+        conversation.update_session(
+            output_modalities=[MultiModality.TEXT, MultiModality.AUDIO],
+            voice='Tina',
+            translation_params=translation_params,
+            input_audio_format=AudioFormat.PCM_16000HZ_MONO_16BIT,
+            output_audio_format=AudioFormat.PCM_24000HZ_MONO_16BIT,
+            enable_turn_detection=True
+        )
 
         await websocket.send_text(json.dumps({"type": "ready", "message": "Connected to translation service"}))
 
+        # 循环接收音频数据
         while True:
             audio_data = await websocket.receive_bytes()
             if audio_data:
                 audio_base64 = base64.b64encode(audio_data).decode('utf-8')
-                conversation.appendAudio(audio_base64)
+                conversation.append_audio(audio_base64)
 
     except WebSocketDisconnect:
         logger.info("WebSocket disconnected")
@@ -117,8 +129,7 @@ async def websocket_endpoint(websocket: WebSocket):
     finally:
         if conversation:
             try:
-                conversation.endSession()
-                conversation.close(1000, "Session ended")
+                conversation.close()
             except:
                 pass