import gradio as gr
from llava.model.builder import load_pretrained_model
from llava.mm_utils import process_images, tokenizer_image_token
from llava.constants import IMAGE_TOKEN_INDEX, DEFAULT_IMAGE_TOKEN
from llava.conversation import conv_templates
from PIL import Image
import copy
import torch
import warnings
import requests

warnings.filterwarnings("ignore")

pretrained = "AI-Safeguard/Ivy-VL-llava"
model_name = "llava_qwen"
device = "cuda"
device_map = "auto"

# Load model, tokenizer, and image processor
tokenizer, model, image_processor, max_length = load_pretrained_model(pretrained, None, model_name, device_map=device_map)
model.eval()

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
    image=None,
):
    messages = [{"role": "system", "content": system_message}]

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

    if image:
        # Process image if provided
        image_tensor = process_images([image], image_processor, model.config)
        image_tensor = [_image.to(dtype=torch.float16, device=device) for _image in image_tensor]

        conv_template = "qwen_1_5"
        question = DEFAULT_IMAGE_TOKEN + "\n" + message
        conv = copy.deepcopy(conv_templates[conv_template])
        conv.append_message(conv.roles[0], question)
        conv.append_message(conv.roles[1], None)
        prompt_question = conv.get_prompt()

        input_ids = tokenizer_image_token(prompt_question, tokenizer, IMAGE_TOKEN_INDEX, return_tensors="pt").unsqueeze(0).to(device)
        image_sizes = [image.size]

        cont = model.generate(
            input_ids,
            images=image_tensor,
            image_sizes=image_sizes,
            do_sample=False,
            temperature=temperature,
            max_new_tokens=max_tokens,
        )

        response = tokenizer.batch_decode(cont, skip_special_tokens=True)[0]
    else:
        messages.append({"role": "user", "content": message})

        payload = {
            "messages": messages,
            "max_tokens": max_tokens,
            "temperature": temperature,
            "top_p": top_p,
        }

        response = ""
        for message in client.chat_completion(
            payload,
            stream=True,
        ):
            token = message.choices[0].delta.content
            response += token

    yield response

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)",
        ),
        gr.Image(type="filepath", label="Input Image (optional)"),
    ],
)

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