Bookmark

epub,mobi,azwのサムネイルをエクスプローラーで表示する方法【Windows10,11】 

EPUBファイルから画像を抽出する

from __future__ import print_function, unicode_literals

import os
import time
import sys
import subprocess
import shutil
from xml.etree import ElementTree

TEMP_DIR = '/tmp/epub-extract-{}'.format(int(time.time()))


def procedure(file_path):
    if not os.path.exists(file_path):
        print("{} is not exist.".format(file_path), file=sys.stderr)
        return

    output_dir, ext = os.path.splitext(file_path)

    if ext != '.epub':
        print("{} is not epub.".format(file_path), file=sys.stderr)
        return

    if os.path.exists(output_dir):
        print("{} is already exists.".format(output_dir), file=sys.stderr)
        return

    os.mkdir(TEMP_DIR)

    subprocess.Popen(
        ('unzip', file_path, "-d", TEMP_DIR),
        stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()

    os.mkdir(output_dir)

    container_xml_path = os.path.join(TEMP_DIR, 'META-INF', 'container.xml')
    etree = ElementTree.parse(container_xml_path)
    rootfile_node = etree.find(
        ".//{urn:oasis:names:tc:opendocument:xmlns:container}rootfile")
    content_opf_path = rootfile_node.attrib['full-path']

    content_xml_path = os.path.join(TEMP_DIR, content_opf_path)
    etree = ElementTree.parse(content_xml_path)
    manifest = etree.find('.//{http://www.idpf.org/2007/opf}manifest')
    items = manifest.findall('.//{http://www.idpf.org/2007/opf}item')

    image_paths = []
    for item in items:
        if item.attrib['media-type'] == 'image/jpeg':
            image_paths.append(item.attrib['href'])

    root_dir = os.path.dirname(content_xml_path)

    for i, image_path in enumerate(image_paths, start=1):
        destination_image_name = '{:03d}.jpg'.format(i)
        source_image_path = os.path.join(root_dir, image_path)
        destination_image_path = os.path.join(
            output_dir, destination_image_name)
        shutil.move(source_image_path, destination_image_path)
        print('{} -> {}'.format(image_path, destination_image_name))

    shutil.rmtree(TEMP_DIR)


def main():
    for arg in sys.argv[1:]:
        procedure(arg)


if __name__ == '__main__':
    main()

画像結合ツール