convert CBZ comic ZIP archives to fixed-layout EPUB

[panaC]

Fixes #3816

• Accept .cbz files through the file picker, drag-and-drop, CLI import, and OS file associations.
• Convert each comic image into a fixed-layout XHTML page.
• Generate a valid EPUB container with an uncompressed, first-position mimetype entry.
• Naturally sort image filenames so pages such as 1, 2, and 10 appear in the expected order.
• Read available ComicInfo.xml metadata:
  • title, writer, summary, and year
  • cover designation and page bookmarks
  • right-to-left manga progression
• Generate EPUB navigation, page-list, cover, viewport, and page-spread metadata.
• Remove temporary conversion files after import, including when importing fails.
• Add automated coverage for archive structure and the existing EPUB parsing pipeline.

Tested with samples from https://www.contrapositivediary.com/?p=1197 and https://github.com/clach04/sample_reading_media

[panaC]

EPUB’s mimetype entry has special ZIP requirements:

• It must be the first archive entry.
• Its content must be exactly application/epub+zip.
• It must be stored without compression.

createZip() previously compressed every entry. With:

const doDeflate = (zipPath: string) => zipPath !== "mimetype";

Yazl uses ZIP method STORE (method === 0) for the root mimetype file while continuing to deflate every other resource. The accompanying buffersFirst option ensures it is also written first.

The condition matches only the exact root entry named mimetype, so existing WebPub and customization archives are otherwise unchanged. Without it, the converted file may render but is not a conforming EPUB and can fail validation or be rejected by stricter readers.

[panaC]

Official reference: W3C EPUB 3.3 — §4.3.3 OCF ZIP container media type identification.
It requires that the mimetype file:

• Is the first ZIP entry.
• Contains exactly application/epub+zip.
• Is not compressed or encrypted.
• Has no ZIP-header extra field.