Upload audio_routes.py

routes/audio_routes.py

@@ -15,6 +15,7 @@ from utils.logger import log_info, log_error, log_warning, log_debug
 from tts.tts_factory import TTSFactory
 from tts.tts_preprocessor import TTSPreprocessor
 from config.config_provider import ConfigProvider
+from chat_session.event_bus import Event, EventType
 
 router = APIRouter(tags=["audio"])
 
@@ -43,10 +44,10 @@ async def generate_tts(request: TTSRequest, req: Request):
         # Check if we should use event-driven mode
         if request.session_id and hasattr(req.app.state, 'event_bus'):
             # Event-driven mode for realtime sessions
-            from event_bus import Event, EventType
-            
+            from chat_session.event_bus import Event, EventType
+
             log_info(f"🎤 TTS request via event bus for session: {request.session_id}")
-            
+
             # Publish TTS event
             await req.app.state.event_bus.publish(Event(
                 type=EventType.TTS_STARTED,
@@ -58,14 +59,14 @@ async def generate_tts(request: TTSRequest, req: Request):
                     "is_api_call": True  # Flag to indicate this is from REST API
                 }
             ))
-            
+
             # Return a response indicating audio will be streamed via WebSocket
             return {
                 "status": "processing",
                 "message": "TTS audio will be streamed via WebSocket connection",
                 "session_id": request.session_id
             }
-        
+
         else:
             # Direct TTS generation (legacy mode)
             tts_provider = TTSFactory.create_provider()
@@ -179,10 +180,8 @@ async def transcribe_audio(request: STTRequest, req: Request):
         # Check if we should use event-driven mode
         if request.session_id and hasattr(req.app.state, 'event_bus'):
             # Event-driven mode for realtime sessions
-            from event_bus import Event, EventType
-            
             log_info(f"🎧 STT request via event bus for session: {request.session_id}")
-            
+
             # Publish audio chunk event
             await req.app.state.event_bus.publish(Event(
                 type=EventType.AUDIO_CHUNK_RECEIVED,
@@ -194,14 +193,14 @@ async def transcribe_audio(request: STTRequest, req: Request):
                     "is_api_call": True
                 }
             ))
-            
+
             # Return a response indicating transcription will be available via WebSocket
             return {
                 "status": "processing",
                 "message": "Transcription will be available via WebSocket connection",
                 "session_id": request.session_id
             }
-        
+
         else:
             # Direct STT transcription (legacy mode)
             from stt.stt_factory import STTFactory
@@ -327,11 +326,11 @@ async def audio_websocket(websocket: WebSocket, session_id: str, request: Reques
     This is a dedicated audio stream separate from the main conversation WebSocket
     """
     from fastapi import WebSocketDisconnect
-    
+
     try:
         await websocket.accept()
         log_info(f"🎵 Audio WebSocket connected for session: {session_id}")
-        
+
         if not hasattr(request.app.state, 'event_bus'):
             await websocket.send_json({
                 "type": "error",
@@ -339,16 +338,14 @@ async def audio_websocket(websocket: WebSocket, session_id: str, request: Reques
             })
             await websocket.close()
             return
-        
+
         while True:
             try:
                 # Receive audio data
                 data = await websocket.receive_json()
-                
+
                 if data.get("type") == "audio_chunk":
                     # Forward to event bus
-                    from event_bus import Event, EventType
-                    
                     await request.app.state.event_bus.publish(Event(
                         type=EventType.AUDIO_CHUNK_RECEIVED,
                         session_id=session_id,
@@ -358,13 +355,11 @@ async def audio_websocket(websocket: WebSocket, session_id: str, request: Reques
                             "chunk_index": data.get("chunk_index", 0)
                         }
                     ))
-                    
+
                 elif data.get("type") == "control":
                     action = data.get("action")
-                    
+
                     if action == "start_recording":
-                        from event_bus import Event, EventType
-                        
                         await request.app.state.event_bus.publish(Event(
                             type=EventType.STT_STARTED,
                             session_id=session_id,
@@ -373,16 +368,14 @@ async def audio_websocket(websocket: WebSocket, session_id: str, request: Reques
                                 "format": data.get("format", "webm")
                             }
                         ))
-                        
+
                     elif action == "stop_recording":
-                        from event_bus import Event, EventType
-                        
                         await request.app.state.event_bus.publish(Event(
                             type=EventType.STT_STOPPED,
                             session_id=session_id,
                             data={"reason": "user_request"}
                         ))
-                        
+
             except WebSocketDisconnect:
                 break
             except Exception as e:
@@ -391,7 +384,7 @@ async def audio_websocket(websocket: WebSocket, session_id: str, request: Reques
                     "type": "error",
                     "message": str(e)
                 })
-                
+
     except Exception as e:
         log_error(f"Audio WebSocket error", error=str(e))
     finally: