Creating an ebook with pandoc
Starting with version 1.6, pandoc can produce output in the EPUB electronic book format. EPUB books can be viewed on iPads, Nooks, and other electronic book readers, including many smart phones. (They can also be converted to Kindle books using KindleGen.)
This means that it’s now very easy to produce an electronic book! Let’s try it.
A toy example
Use your text editor to create a file
mybook.txt, with the following
% My Book % Sam Smith This is my book! # Chapter One Chapter one is over. # Chapter Two Chapter two has just begun.
To make this into an ebook takes only one command:
pandoc mybook.txt -o mybook.epub
You can upload
to your ebook reader and try it out.
Note that if your markdown file contains links to local images, for example
pandoc will automatically include the images in the generated epub.
A real book
To see what this would look like for a real book, let’s convert Scott Chacon’s book Pro Git, which he wrote using pandoc’s markdown variant and released under a Creative Commons license. (If you use the book, please consider buying a copy to help support his excellent work.)
git clone https://github.com/progit/progit.git
This command will create a working directory called
progit on your machine. The actual
markdown sources for the English version of the book are in the
en subdirectory, so start by changing
to that directory:
As you can see, each chapter is a single text file in its own directory. Chacon does some postprocessing on these files, for example, to insert images. This is a placeholder for Figure 1-1, for example:
Insert 18333fig0101.png Figure 1-1. Local version control diagram.
The actual image file is called
18333fig0101-tn.png and lives in the
figures subdirectory of the
repository, as you can verify.
For demonstration purposes, we want pure markdown files, so let’s change this placeholder into a markdown image link. Pandoc will treat a paragraph containing a single image as a figure with a caption, which is what we want:
![Figure 1-1. Local version control diagram.](../figures/18333fig0101-tn.png)
We can make this change in all the files with a perl one-liner:
perl -i -0pe \ 's/^Insert\s*(.*)\.png\s*\n([^\n]*)$/!\[\2](..\/figures\/\1-tn.png)/mg' \ */*.markdown
This will modify the files in place. (We won’t worry about backing
them up; if we mess up, we can get the original files back with
git reset --hard.)
OK! Now we’re almost ready to make an ebook. We have the chapters,
each in its own file, but we still need a title. Create a file,
title.txt, with a pandoc YAML
--- title: Pro Git author: Scott Chacon rights: Creative Commons Non-Commercial Share Alike 3.0 language: en-US ...
See the User’s Guide for more information above these fields.
Now run pandoc to make the ebook, using our title page and modified chapter files as sources:
pandoc -o progit.epub title.txt \ 01-introduction/01-chapter1.markdown \ 02-git-basics/01-chapter2.markdown \ 03-git-branching/01-chapter3.markdown \ 04-git-server/01-chapter4.markdown \ 05-distributed-git/01-chapter5.markdown \ 06-git-tools/01-chapter6.markdown \ 07-customizing-git/01-chapter7.markdown \ 08-git-and-other-scms/01-chapter8.markdown \ 09-git-internals/01-chapter9.markdown
That’s it! The ebook,
progit.epub, is ready to be uploaded
to your reader.
Changing the format
You can use the
to specify a CSS file for the book. The default CSS is minimal and can
be found on
GitHub or in the
file in your data directory (see
--data-dir in the User’s Guide).
You can even embed fonts in the EPUB if you want; see the User’s Guide under
Pandoc has an EPUB3 writer. It renders LaTeX math into MathML, which EPUB3 readers are supposed to support (but unfortunately few do).
Of course, this isn’t much help if you want EPUB2 output (
pandoc -t epub2) or target readers
that don’t support MathML. Then you have two options:
- Use the
--webtexoption, which will use a web service to convert the TeX to an image.
- Use the
--gladtexoption to convert maths into SVG images on your local machine.
Both GladTeX and WebTeX add the LaTeX source of the formula as alternative text of the image, increasing accessibility for blind users.