import gradio as gr
from huggingface_hub import InferenceClient

"""
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
"""
client = InferenceClient("R3troR0b/What-If-Explorer")


def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    # Prepare the prompt based on the message and history
    prompt = system_message + "\n"
    for val in history:
        if val[0]:
            prompt += "User: " + val[0] + "\n"
        if val[1]:
            prompt += "Assistant: " + val[1] + "\n"

    prompt += "User: " + message + "\nAssistant:"

    response = ""

    # Prevent infinite loops by limiting history and avoiding repeated responses
    if len(history) > 5:  # Limit history to the last 5 exchanges
        history = history[-5:]

    # Detect if responses are getting repetitive and stop the loop
    if len(set([h[1] for h in history])) == 1:  # All assistant's responses are the same
        yield "It seems we're repeating ourselves. Let's move to a new topic."
        return

    # Use text-generation instead of chat-completion
    for message in client.text_generation(
        prompt=prompt,
        max_new_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
        stream=True,
    ):
        # Since the message is a string, no need for indexing
        token = message.replace(prompt, '')
        response += token
        yield response


"""
For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/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()