Update app.py

app.py

@@ -5,6 +5,9 @@ import json
 # Default model
 MODEL = "ZeppelinCorp/Charm_15"
 
+# Maximum number of messages to keep in chat history
+MAX_HISTORY_MESSAGES = 5
+
 # Generate response using Hugging Face Inference API
 def generate_response(prompt, model_name, temperature, top_k, top_p, chat_history):
     global MODEL
@@ -12,6 +15,10 @@ def generate_response(prompt, model_name, temperature, top_k, top_p, chat_histor
     if model_name != MODEL:
         MODEL = model_name
     
+    # Truncate chat history to the last MAX_HISTORY_MESSAGES messages
+    if len(chat_history) > MAX_HISTORY_MESSAGES:
+        chat_history = chat_history[-MAX_HISTORY_MESSAGES:]
+    
     # Combine chat history into a single prompt
     full_prompt = "\n".join([f"User: {msg['content']}" if msg['role'] == 'user' else f"Bot: {msg['content']}" for msg in chat_history])
     full_prompt += f"\nUser: {prompt}\nBot:"
@@ -32,6 +39,7 @@ def generate_response(prompt, model_name, temperature, top_k, top_p, chat_histor
         response = requests.post(API_URL, json=payload)
         if response.status_code == 200:
             generated_text = response.json()[0]['generated_text']
+            chat_history.append({"role": "user", "content": prompt})
             chat_history.append({"role": "assistant", "content": generated_text})
         else:
             chat_history.append({"role": "assistant", "content": f"Error: {response.status_code} - {response.text}"})