Update app.py

app.py

@@ -37,6 +37,11 @@ class ConversationManager:
     def add_exchange(self, user_message, assistant_response, formatted_response):
         self.model_history.append((user_message, assistant_response))
         self.user_history.append((user_message, formatted_response))
+        # Log the exchange
+        print(f"\nModel History Exchange:")
+        print(f"User: {user_message}")
+        print(f"Assistant (Original): {assistant_response}")
+        print(f"Assistant (Formatted): {formatted_response}")
     
     def get_model_history(self):
         return self.model_history
@@ -76,7 +81,14 @@ def stream_chat(
     top_k: int = 1,
     penalty: float = 1.1,
 ):
+    print(f'\nNew Chat Request:')
+    print(f'Message: {message}')
+    print(f'History from UI: {history}')
+    print(f'System Prompt: {system_prompt}')
+    print(f'Parameters: temp={temperature}, max_tokens={max_new_tokens}, top_p={top_p}, top_k={top_k}, penalty={penalty}')
+    
     model_history = conversation_manager.get_model_history()
+    print(f'Model History: {model_history}')
     
     conversation = []
     for prompt, answer in model_history:
@@ -86,6 +98,8 @@ def stream_chat(
         ])
     
     conversation.append({"role": "user", "content": message})
+    print(f'\nFormatted Conversation for Model:')
+    print(conversation)
     
     input_ids = tokenizer.apply_chat_template(
         conversation, 
@@ -126,6 +140,10 @@ def stream_chat(
             formatted_buffer = format_response(buffer)
             
             if thread.is_alive() is False:
+                print(f'\nGeneration Complete:')
+                print(f'Original Response: {original_response}')
+                print(f'Formatted Response: {formatted_buffer}')
+                
                 conversation_manager.add_exchange(
                     message,
                     original_response,  # Original for model