Update app.py

app.py

@@ -49,7 +49,7 @@ def normalize_input(input_text):
     return input_text.strip()
 
 def select_best_response(responses, request):
-    coherent_responses = filter_by_coherence([resp['response'] for resp in responses], request)
+    coherent_responses = filter_by_coherence(responses, request)
     best_response = filter_by_similarity(coherent_responses)
     return best_response
 
@@ -75,19 +75,20 @@ async def generate_chat(request: ChatRequest):
     print(f"Procesando solicitud: {request.message}")
 
     # Crear un ThreadPoolExecutor para ejecutar las tareas en paralelo
-    with ThreadPoolExecutor(max_workers=None) as executor:
+    with ThreadPoolExecutor() as executor:
         # Usar tqdm para mostrar la barra de progreso
         futures = [executor.submit(generate_chat_response, request, llm) for llm in llms]
         responses = []
-        
+
         for future in tqdm(as_completed(futures), total=len(futures), desc="Generando respuestas"):
             response = future.result()
             responses.append(response)
             print(f"Modelo procesado: {response['literal'][:30]}...")  # Muestra los primeros 30 caracteres de la respuesta
     
     # Verificar si hay errores en las respuestas
-    if any("Error" in response['response'] for response in responses):
-        error_response = next(response for response in responses if "Error" in response['response'])
+    error_responses = [resp for resp in responses if "Error" in resp['response']]
+    if error_responses:
+        error_response = error_responses[0]
         raise HTTPException(status_code=500, detail=error_response['response'])
     
     best_response = select_best_response([resp['response'] for resp in responses], request)