Server Py update

Backend/server.py

@@ -52,12 +52,15 @@ manager = ConnectionManager()
 # Helper function to convert audio data
 async def decode_audio_data(audio_data: str) -> torch.Tensor:
     """Decode base64 audio data to a torch tensor"""
+    try:
         # Decode base64 audio data
         binary_data = base64.b64decode(audio_data.split(',')[1] if ',' in audio_data else audio_data)
         
-    # Load audio from binary data
-    buf = BytesIO(binary_data)
-    audio_tensor, sample_rate = torchaudio.load(buf)
+        # Save to a temporary WAV file first
+        temp_file = BytesIO(binary_data)
+        
+        # Load audio from binary data, explicitly specifying the format
+        audio_tensor, sample_rate = torchaudio.load(temp_file, format="wav")
         
         # Resample if needed
         if sample_rate != generator.sample_rate:
@@ -70,6 +73,10 @@ async def decode_audio_data(audio_data: str) -> torch.Tensor:
             audio_tensor = audio_tensor.squeeze(0)
             
         return audio_tensor
+    except Exception as e:
+        print(f"Error decoding audio: {str(e)}")
+        # Return a small silent audio segment as fallback
+        return torch.zeros(generator.sample_rate // 2)  # 0.5 seconds of silence
 
 
 async def encode_audio_data(audio_tensor: torch.Tensor) -> str:
@@ -95,6 +102,7 @@ async def websocket_endpoint(websocket: WebSocket):
             action = request.get("action")
             
             if action == "generate":
+                try:
                     text = request.get("text", "")
                     speaker_id = request.get("speaker", 0)
                     
@@ -116,8 +124,15 @@ async def websocket_endpoint(websocket: WebSocket):
                         "type": "audio_response",
                         "audio": audio_base64
                     })
+                except Exception as e:
+                    print(f"Error generating audio: {str(e)}")
+                    await websocket.send_json({
+                        "type": "error",
+                        "message": f"Error generating audio: {str(e)}"
+                    })
                 
             elif action == "add_to_context":
+                try:
                     text = request.get("text", "")
                     speaker_id = request.get("speaker", 0)
                     audio_data = request.get("audio", "")
@@ -132,6 +147,12 @@ async def websocket_endpoint(websocket: WebSocket):
                         "type": "context_updated",
                         "message": "Audio added to context"
                     })
+                except Exception as e:
+                    print(f"Error adding to context: {str(e)}")
+                    await websocket.send_json({
+                        "type": "error",
+                        "message": f"Error processing audio: {str(e)}"
+                    })
                 
             elif action == "clear_context":
                 context_segments = []