import gradio as gr
import torch
from diffusers import StableDiffusionPipeline
from PIL import Image
import traceback
from typing import Optional

model_id: str = "runwayml/stable-diffusion-v1-5"
device: str = "cpu"  # force CPU usage for compatibility

image_generator_pipe: Optional[StableDiffusionPipeline] = None
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float32)
image_generator_pipe = pipe.to(device)

def generate_image_sd(prompt: str, negative_prompt: str, guidance_scale: float, num_inference_steps: int) -> Image.Image:
    with torch.no_grad():
        output = image_generator_pipe(
            prompt=prompt,
            negative_prompt=negative_prompt,
            guidance_scale=guidance_scale,
            num_inference_steps=num_inference_steps
        )
        image = output.images[0] if output.images else None

    if not image:
        raise RuntimeError("No image was returned from the generation pipeline.")
    return image


with gr.Blocks(theme=gr.themes.Soft()) as demo:
    with gr.Row():
        with gr.Column(scale=1):
            prompt = gr.Textbox(label="Prompt", placeholder="A beautiful futuristic city skyline at night")
            neg_prompt = gr.Textbox(label="Negative Prompt", placeholder="blurry, distorted, watermark")
            guidance = gr.Slider(1.0, 15.0, value=7.5, step=0.5, label="Guidance Scale")
            steps = gr.Slider(10, 50, value=25, step=1, label="Inference Steps")
            generate_btn = gr.Button("Generate Image")

        with gr.Column(scale=1):
            output_image = gr.Image(label="Generated Image", type="pil")

    generate_btn.click(
        fn=generate_image_sd,
        inputs=[prompt, neg_prompt, guidance, steps],
        outputs=output_image
    )

if __name__ == "__main__":
    if not image_generator_pipe:
        print("WARNING: Image generator pipeline is not available. UI will launch, but generation will fail.")
    demo.launch(server_name="0.0.0.0", server_port=7860)