Update app.py

app.py

@@ -5,6 +5,7 @@ import re
 import httpx
 from spaces import GPU
 import asyncio
+import gradio as gr
 
 global_data = {
     'models': {},
@@ -90,39 +91,6 @@ def generate_model_response(model, inputs):
         print(f"Error generating model response: {e}")
         return ""
 
-async def handle_request(request):
-    if request.method == "POST" and request.url.path == "/generate":
-        try:
-            chat_request = ChatRequest(**request.json())
-            inputs = normalize_input(chat_request.message)
-            with ThreadPoolExecutor() as executor:
-                futures = [
-                    executor.submit(generate_model_response, model, inputs)
-                    for model in global_data['models'].values()
-                ]
-                responses = [{'model': model_name, 'response': future.result()} for model_name, future in zip(global_data['models'].keys(), as_completed(futures))]
-            unique_responses = remove_repetitive_responses(responses)
-            return httpx.Response(status_code=200, json=unique_responses)
-        except Exception as e:
-            print(f"Error handling request: {e}")
-            return httpx.Response(status_code=500, json={"error": f"Error handling request: {e}"})
-
-    else:
-        return httpx.Response(status_code=404, text="Not Found")
-
-
-async def run_server(port: int):
-    async def serve_request(request: httpx.Request) -> httpx.Response:
-        return await handle_request(request)
-
-    from uvicorn.config import Config
-    from uvicorn.main import Server
-
-    config = Config(app=serve_request, host="127.0.0.1", port=port, log_level="info")
-    server = Server(config=config)
-
-    await server.serve()
-
 def remove_repetitive_responses(responses):
     unique_responses = {}
     for response in responses:
@@ -130,6 +98,33 @@ def remove_repetitive_responses(responses):
             unique_responses[response['model']] = response['response']
     return unique_responses
 
-if __name__ == "__main__":
-    port = 7860
-    asyncio.run(run_server(port))
+async def process_message(message):
+    inputs = normalize_input(message)
+    with ThreadPoolExecutor() as executor:
+        futures = [
+            executor.submit(generate_model_response, model, inputs)
+            for model in global_data['models'].values()
+        ]
+        responses = [{'model': model_name, 'response': future.result()} for model_name, future in zip(global_data['models'].keys(), as_completed(futures))]
+    unique_responses = remove_repetitive_responses(responses)
+    formatted_response = ""
+    for model, response in unique_responses.items():
+        formatted_response += f"**{model}:**\n{response}\n\n"
+    
+    curl_command = f"""
+    curl -X POST -H "Content-Type: application/json" \\
+         -d '{{"message": "{message}"}}' \\
+         http://localhost:7860/generate
+    """
+    return formatted_response, curl_command
+
+
+iface = gr.Interface(
+    fn=process_message,
+    inputs=gr.Textbox(lines=2, placeholder="Enter your message here..."),
+    outputs=[gr.Markdown(), gr.Textbox(label="cURL command")],
+    title="Multi-Model LLM API",
+    description="Enter a message and get responses from multiple LLMs.",
+)
+
+iface.launch(server_port=7860)