import dotenv
import gradio as gr
from huggingface_hub import InferenceClient

def respond_chat_completion(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages_in = [
        {
            "role": "system",
            "content": system_message
        }
        ]

    for val in history:
        if val[0]:
            messages_in.append(
                {
                    "role": "user",
                    "content": val[0]
                }
                )
        if val[1]:
            messages_in.append(
                {
                    "role": "assistant",
                    "content": val[1]
                }
                )

    messages_in.append(
        {
            "role": "user",
            "content": message
        }
        )

    dotenv.load_dotenv(dotenv.find_dotenv())  # Load env. variable HF_TOKEN
    client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")

    messages_out = client.chat_completion(
        messages_in,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p
        )

    response = ""
    for message in messages_out:
        token = message.choices[0].delta.content
        response += token
        yield response


chatbot = gr.ChatInterface(
    respond_chat_completion,
    type='messages',
    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__":
    chatbot.launch()