from PIL import Image
from io import BytesIO
import base64

from diffusers import StableDiffusionControlNetPipeline, ControlNetModel, UniPCMultistepScheduler
from diffusers.utils import load_image
import torch

import gradio as gr

controlnet = ControlNetModel.from_pretrained("rgres/sd-controlnet-aerialdreams", torch_dtype=torch.float16)
pipe = StableDiffusionControlNetPipeline.from_pretrained(
    "stabilityai/stable-diffusion-2-1-base", controlnet=controlnet, torch_dtype=torch.float16
)

pipe = pipe.to("cuda")

# CPU offloading for faster inference times
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)
pipe.enable_model_cpu_offload()

def generate_map(image, prompt, steps, seed):
    #image = Image.open(BytesIO(base64.b64decode(image_base64)))
    generator = torch.manual_seed(seed)

    image = Image.fromarray(image)

    image = pipe(
        prompt=prompt,
        num_inference_steps=steps,
        image=image
    ).images[0]

    return image

demo = gr.Interface(
  fn=generate_map,
  server_port=os.getenv('GRADIO_PORT', '7860'),
  inputs=["image", "text", gr.Slider(0,100), "number"],
  outputs="image")

demo.launch()