app.py

# app.py

import gradio as gr
from PIL import Image
import os
import torch
from diffusers import StableDiffusionInpaintPipeline
from utils.planner import extract_scene_plan  # 🧠 Brain Layer

# ----------------------------
# 🔧 Device Setup
# ----------------------------
device = "cuda" if torch.cuda.is_available() else "cpu"
dtype = torch.float16 if device == "cuda" else torch.float32

# ----------------------------
# 📦 Load Inpainting Model
# ----------------------------


pipe = StableDiffusionInpaintPipeline.from_pretrained(
    "runwayml/stable-diffusion-inpainting",
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32
).to("cuda" if torch.cuda.is_available() else "cpu")

pipe.enable_attention_slicing()
pipe.enable_model_cpu_offload()

# ----------------------------
# 🎨 Image Generation Function
# ----------------------------
def process_image(prompt, image, mask, num_variations):
    try:
        print("🧠 Prompt received:", prompt)

        # 🧠 Step 1: Brain Layer
        reasoning_json = extract_scene_plan(prompt)
        print("🧠 Scene plan extracted:", reasoning_json)

        # Resize inputs to 1024x1024 (required for SDXL)
        image = image.resize((1024, 1024)).convert("RGB")
        mask = mask.resize((1024, 1024)).convert("L")

        results = []
        for i in range(num_variations):
            print(f"🎨 Generating variation {i + 1}...")
            output = pipe(
                prompt=prompt,
                image=image,
                mask_image=mask,
                strength=0.98,
                guidance_scale=7.5,
                num_inference_steps=40
            ).images[0]
            results.append(output)

        return results, reasoning_json

    except Exception as e:
        print("❌ Error during generation:", e)
        return ["❌ Generation failed"], {"error": str(e)}

# ----------------------------
# 🖼️ Gradio UI
# ----------------------------
with gr.Blocks() as demo:
    gr.Markdown("## 🧠 NewCrux Inpainting Demo (SDXL)\nUpload a product image, a mask, and a prompt to generate realistic content.")

    with gr.Row():
        with gr.Column():
            prompt_input = gr.Textbox(label="Prompt")
            image_input = gr.Image(type="pil", label="Upload Product Image")
            mask_input = gr.Image(type="pil", label="Upload Mask (white = keep, black = replace)")
            variation_slider = gr.Slider(1, 4, step=1, value=1, label="Number of Variations")
            generate_btn = gr.Button("Generate")

        with gr.Column():
            output_gallery = gr.Gallery(
                label="Generated Variations",
                columns=2,
                rows=2,
                height="auto"
            )
            json_output = gr.JSON(label="🧠 Brain Layer Reasoning")

    generate_btn.click(
        fn=process_image,
        inputs=[prompt_input, image_input, mask_input, variation_slider],
        outputs=[output_gallery, json_output]
    )

demo.launch()