Update websocket_handler.py

websocket_handler.py

@@ -12,6 +12,7 @@ from enum import Enum
 import numpy as np
 import traceback
 
+from realtime_session_manager import send_tts_welcome_message
 from session import Session, session_store
 from config_provider import ConfigProvider
 from chat_handler import handle_new_message, handle_parameter_followup
@@ -272,55 +273,29 @@ async def websocket_endpoint(websocket: WebSocket, session_id: str):
     if not stt_initialized:
         await websocket.send_json({
             "type": "error",
-            "message": "Speech-to-Text service initialization failed. Please check your configuration.",
-            "error_type": "stt_init_failed"
+            "message": "STT initialization failed"
         })
-        await realtime_session.cleanup()
-        await websocket.close()
-        return
     
-    # Send welcome message if exists
-    if session.chat_history and len(session.chat_history) > 0:
-        first_message = session.chat_history[0]
-        if first_message.get('role') == 'assistant':
-            # Send welcome message
-            await websocket.send_json({
-                "type": "assistant_response",
-                "text": first_message.get('content', ''),
-                "is_welcome": True
-            })
-            
-            # Generate TTS for welcome message
-            tts_provider = TTSFactory.create_provider()
-            if tts_provider:
-                try:
-                    welcome_text = first_message.get('content', '')
-                    audio_data = await tts_provider.synthesize(welcome_text)
-                    
-                    # Send audio in chunks
-                    chunk_size = realtime_session.audio_chunk_size
-                    total_chunks = (len(audio_data) + chunk_size - 1) // chunk_size
-                    
-                    for i in range(0, len(audio_data), chunk_size):
-                        chunk = audio_data[i:i + chunk_size]
-                        chunk_index = i // chunk_size
-                        
-                        await websocket.send_json({
-                            "type": "tts_audio",
-                            "data": base64.b64encode(chunk).decode('utf-8'),
-                            "chunk_index": chunk_index,
-                            "total_chunks": total_chunks,
-                            "is_last": chunk_index == total_chunks - 1,
-                            "is_welcome": True
-                        })
-                        
-                        await asyncio.sleep(0.01)
-                    
-                    log_info(f"✅ Welcome message TTS sent", session_id=session_id)
-                except Exception as e:
-                    log_error(f"Failed to generate welcome TTS", error=str(e))
-
-    # STT başarılı, devam et
+    # Generate and send welcome message TTS
+    tts_provider = TTSFactory.create_provider()
+    if tts_provider:
+        # Get welcome text from session history
+        welcome_text = None
+        if session.chat_history and len(session.chat_history) > 0:
+            # Find the first assistant message (welcome message)
+            for msg in session.chat_history:
+                if msg.get('role') == 'assistant':
+                    welcome_text = msg.get('content', '')
+                    break
+        
+        if welcome_text:
+            await send_tts_welcome_message(
+                websocket, 
+                session_id, 
+                tts_provider, 
+                welcome_text
+            )
+    
     try:
         while True:
             # Receive message