#!/usr/bin/env python3 """ Image post-processing toolkit for the creative pipeline. Handles common fixes that don't require full regeneration. Usage: python3 image-postprocess.py flatten-bg input.png output.png --color "#FF6B35" python3 image-postprocess.py upscale input.png output.png --factor 2 python3 image-postprocess.py color-shift input.png output.png --hue 15 --saturation 1.1 python3 image-postprocess.py crop input.png output.png --aspect 16:9 python3 image-postprocess.py info input.png """ import argparse import sys import os def ensure_deps(): """Check dependencies are available.""" try: from PIL import Image import numpy as np return True except ImportError: print("Missing dependencies. Install with:") print(" uv pip install Pillow numpy") print("Or use the existing venv: source /tmp/slides_env/bin/activate") return False def flatten_bg(input_path, output_path, color_hex="#FFFFFF", tolerance=30): """Replace background with a solid color. Detects the dominant edge color as background.""" from PIL import Image import numpy as np img = Image.open(input_path).convert("RGBA") arr = np.array(img) # Parse target color color_hex = color_hex.lstrip('#') target = np.array([int(color_hex[i:i+2], 16) for i in (0, 2, 4)]) # Sample edges to find background color h, w = arr.shape[:2] edge_pixels = np.concatenate([ arr[0, :, :3], # top row arr[-1, :, :3], # bottom row arr[:, 0, :3], # left col arr[:, -1, :3], # right col ]) # Most common edge color (rough background detection) from collections import Counter colors = [tuple(p) for p in edge_pixels] bg_color = np.array(Counter(colors).most_common(1)[0][0]) # Create mask: pixels close to detected background diff = np.sqrt(np.sum((arr[:, :, :3].astype(float) - bg_color.astype(float)) ** 2, axis=2)) mask = diff < tolerance # Replace background pixels with target color result = arr.copy() result[mask, 0] = target[0] result[mask, 1] = target[1] result[mask, 2] = target[2] result[mask, 3] = 255 # fully opaque Image.fromarray(result).convert("RGB").save(output_path, quality=95) print(f"✅ Background flattened → {output_path}") print(f" Detected BG color: rgb{tuple(bg_color)}, replaced with #{color_hex.upper()}") def upscale(input_path, output_path, factor=2): """Simple upscale using Lanczos resampling.""" from PIL import Image img = Image.open(input_path) new_size = (img.width * factor, img.height * factor) result = img.resize(new_size, Image.LANCZOS) result.save(output_path, quality=95) print(f"✅ Upscaled {factor}x → {output_path} ({new_size[0]}x{new_size[1]})") def color_shift(input_path, output_path, hue_shift=0, saturation_mult=1.0, brightness_mult=1.0): """Adjust hue, saturation, and brightness.""" from PIL import Image, ImageEnhance img = Image.open(input_path) if saturation_mult != 1.0: enhancer = ImageEnhance.Color(img) img = enhancer.enhance(saturation_mult) if brightness_mult != 1.0: enhancer = ImageEnhance.Brightness(img) img = enhancer.enhance(brightness_mult) if hue_shift != 0: # Convert to HSV, shift hue, convert back import numpy as np arr = np.array(img.convert("HSV")) arr[:, :, 0] = (arr[:, :, 0].astype(int) + hue_shift) % 256 img = Image.fromarray(arr, "HSV").convert("RGB") img.save(output_path, quality=95) print(f"✅ Color adjusted → {output_path}") print(f" Hue shift: {hue_shift}, Saturation: {saturation_mult}x, Brightness: {brightness_mult}x") def crop_aspect(input_path, output_path, aspect="16:9"): """Crop to target aspect ratio (center crop).""" from PIL import Image img = Image.open(input_path) w, h = img.size target_w, target_h = map(int, aspect.split(":")) target_ratio = target_w / target_h current_ratio = w / h if current_ratio > target_ratio: # Too wide — crop width new_w = int(h * target_ratio) left = (w - new_w) // 2 img = img.crop((left, 0, left + new_w, h)) else: # Too tall — crop height new_h = int(w / target_ratio) top = (h - new_h) // 2 img = img.crop((0, top, w, top + new_h)) img.save(output_path, quality=95) print(f"✅ Cropped to {aspect} → {output_path} ({img.size[0]}x{img.size[1]})") def info(input_path): """Print image metadata.""" from PIL import Image img = Image.open(input_path) file_size = os.path.getsize(input_path) print(f"📊 Image Info: {input_path}") print(f" Size: {img.size[0]}x{img.size[1]}") print(f" Mode: {img.mode}") print(f" Format: {img.format}") print(f" File size: {file_size / 1024:.1f} KB ({file_size / 1024 / 1024:.2f} MB)") # Sample dominant colors import numpy as np from collections import Counter arr = np.array(img.convert("RGB")) # Sample 1000 random pixels flat = arr.reshape(-1, 3) indices = np.random.choice(len(flat), min(1000, len(flat)), replace=False) sampled = flat[indices] # Quantize to reduce color space quantized = (sampled // 32) * 32 colors = Counter(map(tuple, quantized)).most_common(5) print(f" Dominant colors (approx):") for c, count in colors: pct = count / len(indices) * 100 print(f" rgb{c} — {pct:.0f}%") def main(): parser = argparse.ArgumentParser(description="Image post-processing toolkit") sub = parser.add_subparsers(dest="command") # flatten-bg p = sub.add_parser("flatten-bg", help="Replace background with solid color") p.add_argument("input", help="Input image path") p.add_argument("output", help="Output image path") p.add_argument("--color", default="#FFFFFF", help="Target background color (hex)") p.add_argument("--tolerance", type=int, default=30, help="Color similarity tolerance") # upscale p = sub.add_parser("upscale", help="Upscale image") p.add_argument("input", help="Input image path") p.add_argument("output", help="Output image path") p.add_argument("--factor", type=int, default=2, help="Scale factor") # color-shift p = sub.add_parser("color-shift", help="Adjust colors") p.add_argument("input", help="Input image path") p.add_argument("output", help="Output image path") p.add_argument("--hue", type=int, default=0, help="Hue shift (0-255)") p.add_argument("--saturation", type=float, default=1.0, help="Saturation multiplier") p.add_argument("--brightness", type=float, default=1.0, help="Brightness multiplier") # crop p = sub.add_parser("crop", help="Crop to aspect ratio") p.add_argument("input", help="Input image path") p.add_argument("output", help="Output image path") p.add_argument("--aspect", default="16:9", help="Target aspect ratio (e.g., 16:9, 3:4)") # info p = sub.add_parser("info", help="Show image metadata") p.add_argument("input", help="Input image path") args = parser.parse_args() if not args.command: parser.print_help() return if not ensure_deps(): sys.exit(1) if args.command == "flatten-bg": flatten_bg(args.input, args.output, args.color, args.tolerance) elif args.command == "upscale": upscale(args.input, args.output, args.factor) elif args.command == "color-shift": color_shift(args.input, args.output, args.hue, args.saturation, args.brightness) elif args.command == "crop": crop_aspect(args.input, args.output, args.aspect) elif args.command == "info": info(args.input) if __name__ == "__main__": main()