changed websocket binary type to base64

app.py

@@ -2,6 +2,7 @@ import gradio as gr
 import websockets
 import asyncio
 import json
+import base64
 
 async def process_audio_stream(audio_path, max_tokens):
     """
@@ -12,14 +13,17 @@ async def process_audio_stream(audio_path, max_tokens):
         return
     
     try:
-        # Read audio file
+        # Read audio file and convert to base64
         with open(audio_path, 'rb') as f:
-            audio_data = f.read()
-        
+            audio_bytes = f.read()
+            base64_bytes = base64.b64encode(audio_bytes)
+            # Convert to binary for websocket
+            binary_data = base64_bytes
+            
         # Connect to WebSocket
         async with websockets.connect('ws://localhost:8330/ws/process-audio/') as websocket:
-            # Send audio data
-            await websocket.send(audio_data)
+            # Send base64 encoded audio data
+            await websocket.send(binary_data)
             
             # Send parameters
             await websocket.send(json.dumps({
@@ -30,15 +34,23 @@ async def process_audio_stream(audio_path, max_tokens):
             # Initialize response
             response = ""
             
-            # Receive streaming tokens
+            # Receive streaming response
             async for message in websocket:
-                if message == "[DONE]":
-                    break
-                response += message
-                yield response
+                try:
+                    data = json.loads(message)
+                    if data["status"] == "generating":
+                        response += data["token"]
+                        yield response
+                    elif data["status"] == "complete":
+                        break
+                    elif data["status"] == "error":
+                        yield f"Error: {data['error']}"
+                        break
+                except json.JSONDecodeError:
+                    continue
                 
     except Exception as e:
-        yield f"Error processing audio: {str(e)}"
+        yield f"Error connecting to server: {str(e)}"
 
 # Create Gradio interface
 demo = gr.Interface(
@@ -65,9 +77,5 @@ demo = gr.Interface(
     ]
 )
 
-def clear_output(audio, max_tokens):
-    return ""
-demo.load_examples = clear_output
-
 if __name__ == "__main__":
-    demo.queue().launch()
+    demo.queue().launch(server_name="0.0.0.0", server_port=7860)