check.py

from pathlib import Path
from PIL import Image

BASE_DIR = Path("data")
TRAIN_DIR = BASE_DIR / "train"
VAL_DIR = BASE_DIR / "val"
BAD_DIR = BASE_DIR / "bad"

def check_dir(dir_path: Path):
    print(f"Checking {dir_path}...")
    count = 0
    bad = 0

    for img_path in dir_path.rglob("*"):
        if not img_path.is_file():
            continue

        # Optional: restrict to known extensions
        if img_path.suffix.lower() not in {".jpg", ".jpeg", ".png"}:
            continue

        count += 1
        try:
            with Image.open(img_path) as img:
                img.verify()  # check integrity
        except Exception as e:
            bad += 1
            print(f"  BAD: {img_path} ({e})")

            # Move to bad/ preserving subdirs
            rel = img_path.relative_to(BASE_DIR)
            target = BAD_DIR / rel
            target.parent.mkdir(parents=True, exist_ok=True)
            img_path.rename(target)

    print(f"Checked {count} images in {dir_path}, found {bad} bad.")

def main():
    BAD_DIR.mkdir(parents=True, exist_ok=True)

    if TRAIN_DIR.exists():
        check_dir(TRAIN_DIR)
    if VAL_DIR.exists():
        check_dir(VAL_DIR)

    print("Done.")

if __name__ == "__main__":
    main()