backup_unref_images.sh

#!/bin/bash
SCRIPT_DIR="$(cd $(dirname "$0") && pwd -P)"
LOGSEQ_DIR="$(dirname "$SCRIPT_DIR")"
PY_SCRIPT="clear_unref_images.py"

"$SCRIPT_DIR/$PY_SCRIPT" "$LOGSEQ_DIR" -b

echo 'DONE'
read

clear_unref_images.py

#!/usr/bin/env python3
import argparse
import os
import re
from pathlib import Path

md_img_pattern = re.compile('!\[[^\]]*\]\((.+\.(?:jpg|jpeg|png|gif))\)')


def search_images_in_assets(content):
    def path_relative_to_assets():
        for link in re.findall(md_img_pattern, content):
            if link.find('assets/') >= 0:
                yield link[link.find('assets/') + len('assets/'):]

    return [path for path in path_relative_to_assets()]


parser = argparse.ArgumentParser(description='找出 logseq 中未引用的图片')
parser.add_argument('-b', '--bak', action='store_true', help='备份图片到 assets/bak')
parser.add_argument('-d', '--delete', action='store_true', help='删除图片，有备份选项时忽略')
parser.add_argument('workdir', type=str, action='store', nargs=1, help='logseq 根目录')

if __name__ == '__main__':
    args = parser.parse_args()
    # print(args)
    workdir = Path(args.workdir[0])
    assets_dir = workdir / 'assets'
    bak_dir = assets_dir / 'bak'
    search_dirs = [workdir / 'pages', workdir / 'journals']

    images_in_notes = set()
    for search_dir in search_dirs:
        for root, dirs, files in os.walk(search_dir):
            for file in files:
                if file.endswith('.md'):
                    with open(Path(root, file)) as f:
                        images_in_notes.update(search_images_in_assets(f.read()))

    # print(images_in_notes)

    none_refs_images = set()
    for root, dirs, files in os.walk(assets_dir):
        for file in files:
            if file.endswith(('jpg', 'jpeg', 'png', 'gif')):
                image_relative_path = str(Path(root, file).relative_to(assets_dir))
                if image_relative_path not in images_in_notes:
                    none_refs_images.add(image_relative_path)

    if not (args.bak or args.delete):
        print('应该被删除的图片')
        print(none_refs_images)

    if args.bak:
        if not bak_dir.exists():
            bak_dir.mkdir()

    for images in none_refs_images:
        img_file = Path(assets_dir, images)
        if args.bak:
            print('备份:\t', img_file)
            img_file.rename(bak_dir / img_file.name)
        elif args.delete:
            print('删除:\t', img_file)
            os.remove(img_file)

delete_unref_images.sh

#!/bin/bash
SCRIPT_DIR="$(cd $(dirname "$0") && pwd -P)"
LOGSEQ_DIR="$(dirname "$SCRIPT_DIR")"
PY_SCRIPT="clear_unref_images.py"

"$SCRIPT_DIR/$PY_SCRIPT" "$LOGSEQ_DIR" -d

echo 'DONE'
read