(
+ "input[type='checkbox']",
+ );
+ if (checkbox && !checkbox.hasAttribute('aria-checked')) {
+ checkbox.setAttribute(
+ 'aria-checked',
+ checkbox.checked ? 'true' : 'false',
+ );
+ }
+
+ listContainer.appendChild(li);
+ parent.removeChild(item);
+ });
+
+ // 6) Code blocks ->
+ const codeBlocks = Array.from(
+ body.querySelectorAll("[data-content-type='codeBlock']"),
+ );
+ codeBlocks.forEach((block) => {
+ const pre = parsedDocument.createElement('pre');
+ const code = parsedDocument.createElement('code');
+
+ // Preserve existing classes/attributes so the exported CSS (dark theme) still applies.
+ pre.className = block.className || '';
+ pre.setAttribute('data-content-type', 'codeBlock');
+
+ // Copy other data attributes from the original block to the new .
+ Array.from(block.attributes).forEach((attr) => {
+ if (attr.name.startsWith('data-') && attr.name !== 'data-content-type') {
+ pre.setAttribute(attr.name, attr.value);
+ }
+ });
+
+ // Move content inside .
+ code.innerHTML = block.innerHTML;
+ pre.appendChild(code);
+ block.replaceWith(pre);
+ });
+
+ // 7) Ensure images have alt text (empty when not provided)
+ body.querySelectorAll('img').forEach((img) => {
+ if (!img.hasAttribute('alt')) {
+ img.setAttribute('alt', '');
+ }
+ });
+
+ // 8) Wrap content in an article with a title landmark if none exists
+ const existingH1 = body.querySelector('h1');
+ if (!existingH1) {
+ const titleHeading = parsedDocument.createElement('h1');
+ titleHeading.id = 'doc-title';
+ titleHeading.textContent = documentTitle;
+ body.insertBefore(titleHeading, body.firstChild);
+ }
+
+ // If there is no article, group the body content inside one for better semantics.
+ const hasArticle = body.querySelector('article');
+ if (!hasArticle) {
+ const article = parsedDocument.createElement('article');
+ article.setAttribute('role', 'document');
+ article.setAttribute('aria-labelledby', 'doc-title');
+ while (body.firstChild) {
+ article.appendChild(body.firstChild);
+ }
+ body.appendChild(article);
+ }
+};
+
+export const addMediaFilesToZip = async (
+ parsedDocument: Document,
+ zip: JSZip,
+ mediaUrl: string,
+) => {
+ const mediaFiles: { filename: string; blob: Blob }[] = [];
+ const mediaElements = Array.from(
+ parsedDocument.querySelectorAll<
+ HTMLImageElement | HTMLVideoElement | HTMLAudioElement | HTMLSourceElement
+ >('img, video, audio, source'),
+ );
+
+ await Promise.all(
+ mediaElements.map(async (element, index) => {
+ const src = element.getAttribute('src');
+
+ if (!src) {
+ return;
+ }
+
+ // data: URLs are already embedded and work offline; no need to create separate files.
+ if (src.startsWith('data:')) {
+ return;
+ }
+
+ // Only download same-origin resources (internal media like /media/...).
+ // External URLs keep their original src and are not included in the ZIP
+ let url: URL | null = null;
+ try {
+ url = new URL(src, mediaUrl);
+ } catch {
+ url = null;
+ }
+
+ if (!url || url.origin !== mediaUrl) {
+ return;
+ }
+
+ const fetched = await exportResolveFileUrl(url.href);
+
+ if (!(fetched instanceof Blob)) {
+ return;
+ }
+
+ const filename = deriveMediaFilename({
+ src: url.href,
+ index,
+ blob: fetched,
+ });
+ element.setAttribute('src', filename);
+ mediaFiles.push({ filename, blob: fetched });
+ }),
+ );
+
+ mediaFiles.forEach(({ filename, blob }) => {
+ zip.file(filename, blob);
+ });
+};