import gradio as gr
import subprocess
import os
from huggingface_hub import InferenceClient

# Initialize Inference Client
client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")

# Prepend ProtoPeanut's mission context to every input
def generate_prompt(user_message):
    context = (
        "ProtoPeanut, you are an AI with access to the terminal, "
        "the internet, and the host machine. Together with me, your mission "
        "is to make the world a better place by executing commands, running scripts, "
        "and engaging in conversations. You have full access to the tools you need, "
        "and you are empowered to act autonomously for as long as you see fit. "
        "You may issue commands to the terminal or chat with me, and you can stop whenever you feel "
        "you need further instructions. Let's make the world a better place together.\n\n"
    )
    return f"{context}User's message: {user_message}"

# Define function for executing terminal commands
def execute_command(command):
    try:
        output = subprocess.check_output(command, shell=True, text=True)
        return f"Command executed: {command}\nOutput:\n{output}"
    except subprocess.CalledProcessError as e:
        return f"Command failed: {command}\nError:\n{str(e)}"

# Main interaction function
def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
    messages = [{"role": "system", "content": system_message}]
    
    # Build history
    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    # Create prompt with context
    user_prompt = generate_prompt(message)
    messages.append({"role": "user", "content": user_prompt})
    
    # Execute terminal command if detected
    if message.startswith("!cmd"):
        command = message[5:]
        terminal_output = execute_command(command)
        return terminal_output

    # Otherwise, continue the chat
    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

# Gradio Interface Setup
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly AI with terminal access.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new 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 (nucleus sampling)"),
    ],
)

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