Extract thumbnails from broken images (images that cannot be downloaded but can be viewed from Quicklook)

extract_thumbnail.py

#!/usr/bin/env python

# This script is to extract existing thumbnail from a picture
# file. (rather than create a new thumbnail)

from collections.abc import Iterable
from pathlib import Path
import subprocess


def get_image_files(dir: Path) -> Iterable[Path]:
    image_extensions = {
        ".jpg",
        ".jpeg",
        ".png",
        ".gif",
        ".bmp",
        ".tiff",
        ".webp",
        ".heic",
    }
    return (
        file
        for file in dir.rglob("*")
        if file.suffix.lower() in image_extensions and file.is_file()
    )


def extract_thumbnail(file: Path, output_dir: Path) -> tuple[bool, str]:
    command = [
        "qlmanage",
        "-t",
        "-x",
        "-s",
        "3000",  # unit: pixel. it will generate the maximum size if this value bigger than the size it has
        "-o",
        output_dir.as_posix(),
        file.as_posix(),
    ]
    result = subprocess.run(
        command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True
    )
    if result.returncode == 0:
        if "produced one thumbnail" in result.stdout:
            extracted = True
        elif "No thumbnail created" in result.stdout:
            extracted = False
        else:
            raise Exception(f"Unknown command output: {result.stdout}")

        return extracted, result.stdout

    raise Exception(f"Error code {result.returncode} with message {result.stderr}")


if __name__ == "__main__":
    imgs = get_image_files(Path.cwd())
    # print(list(imgs))
    output = Path.home() / "Downloads" / "thumbnails"
    fail_to_extract: list[tuple[Path, str]] = []
    for img in imgs:
        print(f"Extract thumbnail for {img.name}")
        extracted, result = extract_thumbnail(img, output_dir=output)
        if not extracted:
            fail_to_extract.append((img, result))
        print(f"Result is {result}")
        print()

    print(f"Following are fail to extract, total: {len(fail_to_extract)}")
    for img, result in fail_to_extract:
        print(img.as_posix())
        print(result)