import gradio as gr
from huggingface_hub import InferenceClient

# Initialize the InferenceClient with your chat model.
client = InferenceClient("amusktweewt/tiny-model-500M-chat-v2")

def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
    """
    Builds a chat prompt using a simple template:
      - Optionally includes a system message.
      - Iterates over conversation history (each exchange as a tuple of (user, assistant)).
      - Adds the new user message and appends an empty assistant turn.
    Then it streams the response from the model.
    """
    messages = []
    
    # Include the system prompt if provided.
    if system_message:
        messages.append({"role": "system", "content": system_message})
    
    # Append conversation history.
    if history:
        for user_msg, bot_msg in history:
            messages.append({"role": "user", "content": user_msg})
            messages.append({"role": "assistant", "content": bot_msg})
    
    # Add the new user message and an empty assistant response
    # (this mimics your template where the assistant turn is empty to be filled).
    messages.append({"role": "user", "content": message})
    messages.append({"role": "assistant", "content": ""})
    
    response_text = ""
    # Stream the response token-by-token.
    for resp in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = resp.choices[0].delta.content
        response_text += token
        yield response_text

# Create a Gradio ChatInterface.
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", 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)"),
    ],
)

if __name__ == "__main__":
    demo.launch()