Files

42 lines
1.3 KiB
JavaScript

// Strip translator credits and promo lines from an existing EPUB, in place.
// Usage: node scripts/clean-epub.js <file.epub>...
import JSZip from 'jszip';
import { load } from 'cheerio';
import fs from 'fs';
import { stripPromo } from './clean-chapter.js';
const files = process.argv.slice(2);
if (!files.length) {
console.error('Usage: node scripts/clean-epub.js <file.epub>...');
process.exit(2);
}
for (const file of files) {
const zip = await JSZip.loadAsync(fs.readFileSync(file));
let removed = 0;
for (const name of Object.keys(zip.files)) {
if (!/\.x?html?$/i.test(name) || /(^|\/)nav\.xhtml$/i.test(name)) continue;
const $ = load(await zip.file(name).async('string'), { xml: true });
const n = stripPromo($);
if (n) {
zip.file(name, $.xml());
removed += n;
}
}
if (removed) {
const tmp = `${file}.tmp`;
// EPUB requires mimetype stored uncompressed; loaded entries lose that option.
zip.file('mimetype', 'application/epub+zip', { compression: 'STORE' });
fs.writeFileSync(
tmp,
await zip.generateAsync({
type: 'nodebuffer',
compression: 'DEFLATE',
mimeType: 'application/epub+zip',
}),
);
fs.renameSync(tmp, file);
}
process.stdout.write(`${file}: removed ${removed} promo paragraphs\n`);
}