Lightweight Gitbook Generator

This is a simple script that can turn any directory in your repo into a base folder for your Gitbook.

Quick start:

gradlew run --args="Fullpath true"

What Does This Do?

Gitbook is very useful for writers, or programmers that sometimes write.

Getting it to run now

Organize your markdown files like this:

Suppose this is a book base folder anywhere in your repo, and this folder's relative path to where the Gen files are located is ./demo, then run

$ java Gen demo

or

$ java Gen ./demo

And add a README.md and a book.json:

{
    "root": "./demo"
}

to the root of your repo. The folder ./demo is ready to be linked to your gitbook. Just create a book in Gitbook and link to your repo.

List Of Features

This is a list of features introduced in detail below

• Recursively generating SUMMARY.md.
• File name splitting, which allows user-specified regex rules.
• Arbitrary target position for the book base folder.
• Full confirmation to official Gitbook documentation specification.
• Additional file content processing:
  • Newline handling to avoid github's append-two-spaces-at-end rule
  • Automatically handling of dropbox direct screenshot link. Additional width-attribute scaling enabled.

A Little More Explanation

This is a compact script. All you need is Java for it to work. Copy Gen.class (or Gen.java if you want to compile yourself) to a convenient folder, preferably the root of your repo.

The introduction above aims to get you going as quick as possible. Here are some more explanation. First, please refer to the wonderful documentation here for detailed information. Here I give as little information as possible.

First, I provided a sample source folder ./original_demo that you can experiment with. Copy everything in that folder to ./demo and run the script as shown in the previous section to see the result:

• SUMMARY.md is generated.
• Additional file content processing is provided, as discussed below. To enable this processing run $ java Gen demo true instead.

The book generated is provided here. There will be short gitbook linking instructions in the end of this article.

Organize your folder like this (placed anywhere in your repo, as long as you know the relative path to the root, which we denote as base_to_root):

.
├── README.md
├── SUMMARY.md
├── chapter-1/
|   ├── README.md
|   └── something.md
└── chapter-2/
    ├── README.md
    └── something.md

File	Description	Notes
book.json	Stores configuration data (optional)	See below.
README.md	Preface / Introduction for your book (required)	Attention: This is the cover of your book, and you have to supply this.
SUMMARY.md	Table of Contents (See Pages) (optional)	You don't have to worry about this. If you already have this file, it will be overwritten, so back it up if necessary.

For book.json, put it in the root of your repo, and the gitbook server will use it to find the folder you want to use as the base, an example:

{
    "root": "./articles"
}

This will use the folder ./articles as the base of the gitbook. I could have automated this part, but this reduced the flexibility of the script. It would be useful to seperate the root of the repo, base folder that store the book files and folder that has Gen script.

Suggestions on File Organization

These are not hard guidelines: as long as you confirm to the required structure shown above, which is what Gitbook itself asks, you are okay to go. But generally, in a structure like:

.
├── README.md
├── SUMMARY.md
├── chapter-1/
|   ├── README.md
|   └── something.md
└── chapter-2/
    ├── README.md
    └── something.md

The README.md in the root of your book base folder has been introduced above: it's the introduction cover of your book.

You usually want to use each folder as a part/chapter/subsection further down. In the folder chapter-1, you can optionally include a README.md file, which will act as the chapter cover for this folder's chapter. Note that this README.md is different from the README.md in the root of the book base folder.

It is not advised to include space in any of the file names. I personally advise using TextOne.md if you want it to be shown as an article named Text One in the book. The script by default will take care of that for you.

Everything not mentioned yet

The basic functionality this script provides is generating SUMMARY.md and split file names. There are something additional you can do.

When you supply command line arguments, do make sure to stick to something like

$ java Gen demo true

In the specified order, where demo specifies the base_to_root relative path, and true specifies that you want the additional file content processing mentioned below to be carried out. Both are optional, but you have to specify demo (default value is . if you are already in the book base folder) before you can supply true.

Now, about file content processing, these are purely personal preferences. I don't like how newline is handled in github markdown (newline alone does not give you a new line, you have to append two spaces to the end of the line), so I delt with that. Also, I constantly insert dropbox screenshot links like this:

https://www.dropbox.com/s/gtk56vs39qvjtzi/Screenshot%202018-04-04%2000.32.28.png?dl=0

in my markdown files. Instead of writing ![]() or <img src="" width="500"> yourself, just leave the line like that, and it will work.

Such linkes are automatically put in your clipboard when you have dropbox on your macbook installed and this configuration enabled:

Please Google the corresponding configuration if you are using Window.

Additionally, you can optionally append a whitespace-separated width-scaling attribute after then url. An example in a nutshell:

https://www.dropbox.com/s/gtk56vs39qvjtzi/Screenshot%202018-04-04%2000.32.28.png?dl=0 800

will be converted to

<img src="https://www.dropbox.com/s/gtk56vs39qvjtzi/Screenshot%202018-04-04%2000.32.28.png?raw=1" width="800">

One final thing, about how the file names is split. You can supply a regex.md at the folder of Gen script so that you can specify the regex rules you want applied to name splitting. You will supply pattern and seperate pair in the file as alternating lines, which will be used like:

file_name.replaceAll (pattern, seperator);

I have to say this is a limitted feature. You have to store something like

pattern1
seperator1
pattern2
seperator2
...

in the regex.md file in this line-alternating style. Each of the rule pair will be applied sequentially, so you have to organize your logic if you have more than one rules.
If this file is not supplied, the default rule is applied, which is virtually equivalant to having this regex.md file:

(?<=[^A-Z&&\S])(?=[A-Z])
 

The second line only has a space. This rule actually do the TextOne -> Text One thing mentioned above.

Setting up Your Gitbook

This is not relevant to this script, but just put here to save you a little Googling.

To have a Gitbook, there are various ways to go, please refer to the full documentation. My personal favorite is to sync a book with a repo, and afterwards each of your github commit will now trigger an update to the book automatically.

To do that, first use my script to set up the book base folder as discussed above. Or manually according to the documentation if you prefer. Do remember that your github repo should have the properly set book.json at the root of the repo, so gitbook server knows how to reach the book base folder from your repo root.

I could elaborate, but this page really says it all.