Update app.py

app.py

@@ -1,10 +1,8 @@
 import argparse
 import os
-
 import gradio as gr
 from loguru import logger
 from similarities import BertSimilarity
-
 from chatpdf import ChatPDF
 
 pwd_path = os.path.abspath(os.path.dirname(__file__))
@@ -25,9 +23,9 @@ if __name__ == '__main__':
     parser.add_argument("--server_port", type=int, default=8082)
     parser.add_argument("--share", action='store_true', help="share model")
     args = parser.parse_args()
-
     logger.info(args)
-
+    
+    # Inicializar el modelo
     sim_model = BertSimilarity(model_name_or_path=args.sim_model_name, device=args.device)
     model = ChatPDF(
         similarity_model=sim_model,
@@ -43,21 +41,17 @@ if __name__ == '__main__':
     )
     logger.info(f"chatpdf model: {model}")
 
-    def predict_stream(message, history):
+    # Función para generar respuesta (sin yield)
+    def predict(message, history):
         history_format = []
         for human, assistant in history:
             history_format.append([human, assistant])
         model.history = history_format
-        for chunk in model.predict_stream(message):
-            yield chunk
-
-    def predict(message, history):
-        logger.debug(message)
         response, reference_results = model.predict(message)
         r = response + "\n\n" + '\n'.join(reference_results)
-        logger.debug(r)
-        return r
+        return r, history_format
 
+    # Crear el chatbot
     chatbot_stream = gr.Chatbot(
         height=600,
         avatar_images=(
@@ -66,24 +60,28 @@ if __name__ == '__main__':
         ),
         bubble_full_width=False
     )
-
-    title = "🤖ChatPDF Zonia🤖"
+    
+    title = " 🤖ChatPDF Zonia🤖 "
     css = """.toast-wrap { display: none !important } """
     examples = ['Puede hablarme del PNL?', 'Introducción a la PNL']
 
-    # Configura la interfaz de chat
-    chat_interface_stream = gr.ChatInterface(
-        predict_stream,
-        textbox=gr.Textbox(lines=4, placeholder="Hazme una pregunta", scale=7),
+    # Crear la interfaz sin utilizar la función que usa yield
+    chat_interface = gr.Interface(
+        fn=predict,
+        inputs=[gr.Textbox(lines=4, placeholder="Ask me question", scale=7), gr.State()],
+        outputs=[chatbot_stream, gr.State()],
         title=title,
-        chatbot=chatbot_stream,
         css=css,
         examples=examples,
         theme='soft',
     ).queue()  # Asegúrate de habilitar la cola aquí
 
     with gr.Blocks() as demo:
-        chat_interface_stream.render()
+        chat_interface.render()
     
-    # Lanza la aplicación
-    demo.queue().launch(server_name=args.server_name, server_port=args.server_port, share=args.share)
+    # Lanzar la aplicación con `.queue()`
+    demo.queue().launch(
+        server_name=args.server_name,
+        server_port=args.server_port,
+        share=args.share
+    )