Update app.py

app.py

@@ -6,7 +6,6 @@ import uvicorn
 from dotenv import load_dotenv
 from difflib import SequenceMatcher
 from tqdm import tqdm
-import multiprocessing
 
 load_dotenv()
 
@@ -47,6 +46,7 @@ class ChatRequest(BaseModel):
     top_p: float = 0.95
     temperature: float = 0.7
 
+# Función global para generar respuestas de chat
 def generate_chat_response(request, llm):
     try:
         user_input = normalize_input(request.message)
@@ -102,6 +102,9 @@ def filter_by_similarity(responses):
             break
     return best_response
 
+def worker_function(llm, request):
+    return generate_chat_response(request, llm)
+
 @app.post("/generate_chat")
 async def generate_chat(request: ChatRequest):
     if not request.message.strip():
@@ -109,12 +112,8 @@ async def generate_chat(request: ChatRequest):
     
     print(f"Procesando solicitud: {request.message}")
 
-    # Utilizar un ProcessPoolExecutor para procesar los modelos en paralelo
-    def worker_function(llm):
-        return generate_chat_response(request, llm)
-    
     with ProcessPoolExecutor() as executor:
-        futures = [executor.submit(worker_function, llm) for llm in llms.values()]
+        futures = [executor.submit(worker_function, llm, request) for llm in llms.values()]
         responses = []
 
         for future in tqdm(as_completed(futures), total=len(futures), desc="Generando respuestas"):