feat(scripts): strip translator credits and promo lines from fetched and existing EPUBs

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-25 17:18:05 -03:00
parent fba9a35075
commit 750c2824ce
3 changed files with 82 additions and 1 deletions
@@ -0,0 +1,41 @@
+// 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`);
+}