Update app.py

app.py

@@ -1,7 +1,8 @@
 import gradio as gr
 import os
 import sys
-import json 
+import json
+import copy
 import requests
 import random
 import timeout_decorator
@@ -30,6 +31,8 @@ sys.excepthook = exception_handler
 sys.tracebacklimit = 0
     
 def predict(inputs, top_p, temperature, chat_counter, chatbot, history, request:gr.Request):
+    orig_history = copy.deepcopy(history)
+    orig_chat_counter = chat_counter
     payload = {
         "model": MODEL,
         "messages": [{"role": "user", "content": f"{inputs}"}],
@@ -114,6 +117,9 @@ def predict(inputs, top_p, temperature, chat_counter, chatbot, history, request:
                     yield [(history[i], history[i + 1]) for i in range(0, len(history) - 1, 2) ], history, chat_counter, response, gr.update(interactive=False), gr.update(interactive=False)  # resembles {chatbot: chat, state: history}  
     except Exception as e:
         print (f'error found: {e}')
+        return [(orig_history[i], orig_history[i + 1]) for i in range(0, len(orig_history) - 1, 2) ], orig_history, orig_chat_counter, 'Error! Please try again', gr.update(interactive=True), gr.update(interactive=True)
+        
+
     yield [(history[i], history[i + 1]) for i in range(0, len(history) - 1, 2) ], history, chat_counter, response, gr.update(interactive=True), gr.update(interactive=True)
     print(json.dumps({"chat_counter": chat_counter, "payload": payload, "partial_words": partial_words, "token_counter": token_counter, "counter": counter}))