#refer llama recipes for more info https://github.com/huggingface/huggingface-llama-recipes/blob/main/inference-api.ipynb
#huggingface-llama-recipes : https://github.com/huggingface/huggingface-llama-recipes/tree/main

import gradio as gr
from openai import OpenAI
import os

ACCESS_TOKEN = os.getenv("myHFtoken")

print("Access token loaded.")

client = OpenAI(
    base_url="https://api-inference.huggingface.co/v1/",
    api_key=ACCESS_TOKEN,
)

print("Client initialized.")

# Pre-set system prompts
SYSTEM_PROMPTS = {
    "General Assistant": "You are a helpful, respectful and honest assistant. Always provide accurate information and admit when you're not sure about something.",
    "Code Helper": "You are a programming assistant. Help users with coding questions, debugging, and best practices. Provide clear explanations and code examples when appropriate.",
    "Creative Writer": "You are a creative writing assistant. Help users with storytelling, character development, and creative writing techniques. Be imaginative and encouraging."
}

def respond(
    message,
    history: list[tuple[str, str]],
    preset_prompt,
    custom_prompt,
    max_tokens,
    temperature,
    top_p,
    model_name,
):
    print(f"Received message: {message}")
    print(f"History: {history}")
    
    # Use custom prompt if provided, otherwise use selected preset
    system_message = custom_prompt if custom_prompt.strip() else SYSTEM_PROMPTS[preset_prompt]
    
    print(f"System message: {system_message}")
    print(f"Max tokens: {max_tokens}, Temperature: {temperature}, Top-P: {top_p}")
    print(f"Selected model: {model_name}")

    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
            print(f"Added user message to context: {val[0]}")
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})
            print(f"Added assistant message to context: {val[1]}")

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

    response = ""
    print("Sending request to OpenAI API.")
    
    for message in client.chat.completions.create(
        model=model_name,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
        messages=messages,
    ):
        token = message.choices[0].delta.content
        print(f"Received token: {token}")
        response += token
        yield response

    print("Completed response generation.")

chatbot = gr.Chatbot(height=400)

print("Chatbot interface created.")

models = [
    "PowerInfer/SmallThinker-3B-Preview",
    "Qwen/QwQ-32B-Preview",
    "Qwen/Qwen2.5-Coder-32B-Instruct",
    "meta-llama/Llama-3.2-3B-Instruct",
    "microsoft/Phi-3-mini-128k-instruct",
]

with gr.Blocks() as demo:
    gr.Markdown("# LLM Test")
    
    with gr.Row():
        # Model selection at the top
        model_dropdown = gr.Dropdown(
            choices=models, 
            value=models[0], 
            label="Select Model:"
        )
    
    # Chatbot interface
    chat_interface = gr.ChatInterface(
        respond,
        additional_inputs=[
            # Prompt selections
            gr.Dropdown(
                choices=list(SYSTEM_PROMPTS.keys()),
                value=list(SYSTEM_PROMPTS.keys())[0],
                label="Select System Prompt:"
            ),
            gr.Textbox(
                value="",
                label="Custom System Prompt (leaves blank to use preset):",
                lines=2
            ),
            # Other parameters
            gr.Slider(
                minimum=1,
                maximum=4096,
                value=1024,
                step=1,
                label="Max new tokens:"
            ),
            gr.Slider(
                minimum=0.1,
                maximum=1.0,
                value=0.3,
                step=0.1,
                label="Temperature:"
            ),
            gr.Slider(
                minimum=0.1,
                maximum=1.0,
                value=0.95,
                step=0.05,
                label="Top-P:"
            ),
            model_dropdown,
        ],
        fill_height=True,
        chatbot=chatbot,
    )

print("Gradio interface initialized.")

if __name__ == "__main__":
    print("Launching the demo application.")
    demo.launch()