azw3-or-mobi-to-cbz

A converter from AZW/AZW3 and MOBI fixed-layout comic book e-books to CBZ.

Motivation

I got my own library of DRM-free *.azw3 files that I got from my Amazon Kindle and ComiXology account. (Getting such files and removing the DRM from them is outside the scope of this repository.) Since the AZW3 format is very similar to the Mobipocket format, many e-book readers can open them.

Unfortunately, although most e-book readers can open such files, very few of them support fixed-layout e-books. This means any comic book becomes basically unreadable.

My first idea was to use Calibre to convert those formats to some other format, such as epub or PDF. Unfortunately, Calibre also doesn't support fixed-layout e-books, and the conversion leads to files with a completely broken layout. It seems like supporting fixed-layout e-books isn't in the current plans for the Calibre project. (But if you now how to code and want to implement it yourself, I believe the project will accept your contributions.)

People on reddit (1) (2) suggested a workaround: use Calibre to convert to ZIP, and then rename the ZIP to CBZ. Although it can work, we also lose all the metadata. And we first have to figure out which of the files are fixed-layout, and only convert those. While this manual workaround could work on a couple of files, it requires too many manual steps to be doable for a library of thousand(s) of files. I needed a better tool. So I wrote it myself.

How it works (and limitations)

The most accurate way to do a conversion from a fixed-layout e-book to something else is by actually rendering the e-book, or by at least having a full HTML+CSS+SVG rendering engine. Having such engine would allow for properly rendering absolutely-positioned text on top of a background image.

This project doesn't do that.

Having a full rendering engine is a lot of work and comes with a lot of extra challenges as well. Instead, this project takes shortcuts. Many shortcuts.

Instead of having a proper interpreter for HTML+CSS+SVG semantics, this project just looks at the linked images from the e-book source code. This project assumes that most comic books will have one single image per page, and thus this project looks over each page to find the name of such image, blindly ignoring all the other markup and styles.

This works surprisingly well. Most comic books can be correctly converted using this naïve approach.

Some comic books have two images per page, because they include the left and right pages together in a single page. This project also handles that.

Then, after collecting the list of the images (per page), this project creates a zip file with the CBZ extension adding all those images, in the order they were found, while also adding a simple comicinfo.xml generated from the book metadata. That's all. You now understand what this project does, what it can do, and what it can't do.

In summary:

• Reflowable e-books: not supported.
• Fixed-layout e-books:
  • With a single image per page: supported.
  • With two images per page: supported.
  • With multiple images per page: converted, but with warnings. The result is probably wrong.
  • With text overlaid on top of images: not supported. They still get converted, but the result will lack any text.

After converting the e-books, you may want to manually check those that had any kind of warning. Those may have been converted incorrectly, and they may need manual fixing, or they may be impossible to losslessly convert to CBZ.

Future of this project

This is a side-project I wrote myself to scratch my own itch. I wrote it during the rare spare time I had. I've used it once, and I don't plan on using it again.

I have no plans on supporting this project. Feel free to use it for whatever purposes you want, but I won't provide any support. It's unlikely I'll ever touch this code again. In fact, I may archive this repository to reflect this status.

That said, I used it once and it indeed worked. It successfully converted over a thousand comic book e-books I had. It may work for you. It may also require some tweaks before it works for you. In any case, it can be helpful to someone else, and that's why I'm sharing it.

Prerequisites

• You are comfortable with a little bit of coding.
• You have a recent version of Python on your computer. If you are running Linux or Mac OS X, you probably already have.
• You can install JupyterLab or Jupyter Notebook and you are comfortable using it.
• If you don't know virtual environments, it may be worth learning how to use them before installing Jupyter or any other Python dependency.
• You have a bunch of *.azw3 files (or *.mobi) without any DRM. Removing DRM from your own e-book library is left as an exercise to the reader.

Getting started

Assuming you have both Python and Jupyter already installed, possibly inside a virtual environment, please also install the dependencies of this project:

pip install -r requirements.txt

Now open the Jupyter notebook (the file with .ipynb extension) and follow the instructions over there.

Related projects

• mobi (at PyPI: the library I used to extract the files from the e-books. Please note it will extract the files to a temporary directory, and that directory is deleted after the conversion is finished.
• CbzMage: Claims to convert azw and pdf comic books to cbz files. I have not tested this project, I cannot tell you anything else about it.