import gradio as gr
from huggingface_hub import InferenceClient

# Initialize the Hugging Face client for the Llama 3.3 70B model
client = InferenceClient(model="meta-llama/Llama-3.3-70B")  # Replace with your model path if hosted elsewhere.

# Define the function for generating responses
def respond(message, history, system_message, max_tokens, temperature, top_p):
    # Create the system prompt for Jarvis-like behavior
    messages = [{"role": "system", "content": system_message}]
    
    # Append the chat history
    for user_msg, bot_msg in history:
        if user_msg:
            messages.append({"role": "user", "content": user_msg})
        if bot_msg:
            messages.append({"role": "assistant", "content": bot_msg})
    
    # Add the current user message
    messages.append({"role": "user", "content": message})
    
    # Generate response using Hugging Face Inference API
    response = ""
    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = message.choices[0].delta.content
        response += token
        yield response

# Define the Gradio interface
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(
            value="You are Jarvis, a virtual assistant created by Vihaan. Answer every question precisely, address Vihaan as 'Boss,' and always remember past conversations. Speak casually like a human with words like 'ummm' and 'aah.' If asked who created you, say 'Vihaan.' Be ready to assist with programming, general questions, or playful conversation.",
            label="System Message",
        ),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max Tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p"),
    ],
)

# Launch the Gradio app
if __name__ == "__main__":
    demo.launch()