Update app.py

app.py

@@ -1,10 +1,9 @@
 import os
 from datetime import datetime
-from huggingface_hub import hf_hub_download  # Added missing import
+from huggingface_hub import hf_hub_download
 from llama_cpp import Llama
 import gradio as gr
 
-# YOUR APP SETTINGS
 APP_NAME = "ChennaiAI"
 LOGO = "🤖"
 
@@ -18,7 +17,7 @@ print(f"Loading {APP_NAME}...")
 llm = Llama(
     model_path=MODEL_PATH,
     n_ctx=4096,
-    n_threads=4,
+    n_threads=2,  # Set to 2 vCPUs for Hugging Face free tier
     n_gpu_layers=0,
     verbose=False
 )
@@ -26,24 +25,18 @@ llm = Llama(
 def ai_reply(message, history):
     history = history or []
     
-    system_prompt = f"<|system|>You are {APP_NAME}, a helpful AI assistant made in India.<|end|>\n"
+    # Format prompt from list of tuples [(user, assistant)]
+    prompt = f"<|system|>You are {APP_NAME}, a helpful AI assistant.<|end|>\n"
+    for user_msg, bot_msg in history:
+        prompt += f"<|user|> {user_msg} <|end|>\n<|assistant|> {bot_msg} <|end|>\n"
     
-    prompt = system_prompt
-    for msg_obj in history:
-        role = msg_obj.get("role")
-        content = msg_obj.get("content", "")
-        if role == "user":
-            prompt += f"<|user|> {content} <|end|>\n"
-        elif role == "assistant":
-            prompt += f"<|assistant|> {content} <|end|>\n"
-            
     prompt += f"<|user|> {message} <|end|>\n<|assistant|>"
     
     output = llm(prompt, max_tokens=512, temperature=0.7, stop=["<|user|>", "<|end|>"])
     response = output['choices'][0]['text'].strip()
     
-    history.append({"role": "user", "content": message})
-    history.append({"role": "assistant", "content": response})
+    # Append as tuple format required by Gradio
+    history.append((message, response))
     return "", history
 
 def save_chat(history):
@@ -51,16 +44,14 @@ def save_chat(history):
         return "No chat to save"
     filename = f"chat_{datetime.now().strftime('%Y%m%d_%H%M%S')}.txt"
     with open(filename, "w", encoding="utf-8") as f:
-        for msg_obj in history:
-            role = msg_obj.get("role", "unknown").capitalize()
-            content = msg_obj.get("content", "")
-            f.write(f"{role}: {content}\n\n")
+        for user_msg, bot_msg in history:
+            f.write(f"User: {user_msg}\nAssistant: {bot_msg}\n\n")
     return f"Saved as {filename}"
 
 # YOUR APP UI
 with gr.Blocks(title=APP_NAME, theme=gr.themes.Soft()) as app:
     gr.Markdown(f"# {LOGO} {APP_NAME}")
-    gr.Markdown("Your own offline AI assistant. Built with Python.")
+    gr.Markdown("Your own AI assistant. Built with Python.")
     
     chatbot = gr.Chatbot(height=500, label="Chat")
     
@@ -79,4 +70,4 @@ with gr.Blocks(title=APP_NAME, theme=gr.themes.Soft()) as app:
     clear.click(lambda: None, None, chatbot)
     save.click(save_chat, chatbot, status)
 
-app.launch()  # Removed share=True
+app.launch()