#!/usr/bin/env python3

import numpy as np
import cv2
import argparse
import os

def compute_sdf(img, scale_factor):
    # Convert to binary image if not already
    _, binary = cv2.threshold(img, 127, 255, cv2.THRESH_BINARY)
    
    # Compute distance transform for both foreground and background
    dist_transform_fg = cv2.distanceTransform(binary, cv2.DIST_L2, 5)
    dist_transform_bg = cv2.distanceTransform(255 - binary, cv2.DIST_L2, 5)
    
    # Combine the distance fields and scale by factor
    sdf = (dist_transform_fg - dist_transform_bg) / scale_factor
    
    # Normalize values to [0, 255] range
    sdf_min = np.min(sdf)
    sdf_max = np.max(sdf)
    sdf = ((sdf - sdf_min) * 255 / (sdf_max - sdf_min))
    
    return sdf.astype(np.uint8)

def main():
    parser = argparse.ArgumentParser(description='Generate SDF from black and white image')
    parser.add_argument('input_images', nargs='+', help='Path to input image(s)')
    parser.add_argument('--scale', type=float, default=1.0, 
                        help='Scale factor for distance (in texels)')
    args = parser.parse_args()
    
    # Process each input image
    for input_path in args.input_images:
        # Get input and output paths
        filename, ext = os.path.splitext(input_path)
        output_path = f"{filename}-sdf{ext}"
        
        # Read input image
        img = cv2.imread(input_path, cv2.IMREAD_GRAYSCALE)
        if img is None:
            print(f"Error: Could not read image {input_path}")
            continue
        
        # Compute SDF with scale factor
        sdf = compute_sdf(img, args.scale)
        
        # Save result
        cv2.imwrite(output_path, sdf)
        print(f"SDF generated and saved to {output_path}")

if __name__ == "__main__":
    main()