This project holds all the data for the https://www.bookstackapp.com/
This site is built using Hugo. Images are stored using git-lfs
.
The "extended" version of hugo must be used (Has support for certain features like WebP resize).
- Blog Posts -
content/posts
- Docs -
content/docs
- Theme -
themes/bookstack
The website also contains a repository of customization hacks which can be found at the /hacks part of the site.
These are managed via a separate "Hacks" repo which is part of this repository as a git submodule. A symlink is then used to point the content/hacks
directory to the hacks/content
directory, where hacks
is the submodule location.
The theme is custom made with snippets taken from the hugo capser theme.
SCSS is used for the styling. Install NPM dependencies via npm install
or yarn
then you can use npm run build
to build the css and site once or npm run dev
to watch for changes.
Search is performed using webidx, which essentially builds a sqlite database search index, that is then loaded to browser upon search then queried locally in-browser via sql.js.
This files required are all in this repo, and hacked to suit our use-case.
The script to build the index is located at search/webidx.pl
, and can be ran via the npm script
npm run build:search
Note: you may need to install some dependencies to run the script see the search/webidx.pl
for more information.
The above command will build the sqlite index database to static/search.db
, intended to be deployed to production. There is also a npm run build:search:compress
command to compress the database file using brotli and gzip (requires both to be installed). In production use, these compressed files should be deployed then served from their compressed state where possible. Here's relevant config for nginx:
location ~* \.(db)$ {
brotli_static on;
gzip_static on;
}
Much of the search UI handling logic can be found in our themes/bookstack/static/js/scripts.js
file.