This post is about my blog, which is powered by Pelican and hosted through GitHub using GitHub Pages. I’ll describe the workflow I use when deploying new posts.
For those of you not familiar with these technologies, Pelican is a static site generator — meaning you can write your content in a format such as Markdown/Jupyter notebooks, and Pelican will automatically generate the HTML files for you; and GitHub Pages is a service provided by GitHub for hosting a website under the <your-username>.github.io URL.
Using Pelican and GitHub Pages is quite easy. There’s one annoying little thing though… GitHub Pages assumes the
master
branch contains the root folder to be served to the world. If you're using Pelican's default settings, the output
folder is the folder you want to serve. output
contains the generated website's files. A natural choice of how to organize the files inside the repository would be to define pelican's root folder - the parent of output
- as the root of the repository. But GitHub Pages needs output
to be the root. Bummer...There are those who solve it using two separate repositories: one for the website “source” files, and one for the output which will be served using GitHub Pages.
I personally don’t like breaking my blog into two repositories. I want to keep everything in one place, so I chose to solve the problem using branches and git hooks.
The first step is to create two branches:
will contain the blog's "source" files, namely - all the files such as thesource
folder which contains the actual posts, andcontent
file.pelicanconf.p
will contain only the contents ofmaster
.output
These branches will obviously live in my GitHub Pages repo (https://github.com/yoel-zeldes/yoel-zeldes.github.io), and since the
master
branch contains output
's contents, a user navigating to yoel-zeldes.github.iowill be presented with the goodness of my blog.Unfortunately, manually maintaining these two branches is cumbersome. Git hooks to the rescue!
Git has a mechanism to execute custom scripts when certain important actions occur. In my case, whenever I push a commit to the
source
branch, I'd like the master
branch to get updated with the new contents of output
. This can be done using the pre-push hook, which is executed - you guessed it - just before a push occurs.All you have to do is create a file named
.git/hooks/pre-push
with the following content:#!/bin/sh
while read local_ref local_sha remote_ref remote_sha
do
if [ "$remote_ref" = "refs/heads/source" ]
then
echo 'pushing output folder (production version) to master...'
pelican content -o output -s publishconf.py
echo anotherdatum.com > output/CNAME
ghp-import output
git push --no-verify git@github.com:yoel-zeldes/yoel-zeldes.github.io.git gh-pages:master
pelican content -o output
fi
done
exit 0
- The first thing the script does is iterating over the commits that are about to be pushed. Specifically, only commits that are pushed to the
branch are of interest to us.source
- If commits are pushed to
, it executessource
command usingpelican
. This will generate the production version of the blog intopublishconf.py
.output
- It then creates a
file, which is needed since I use a custom domain (http://anotherdatum.com).CNAME
- The GitHub Pages Import tool is used for copying the contents of
to a branch namedoutput
.gh-pages
is pushed to the remotegh-pages
branch.master
skips the pre-push hook so this script won't run again.--no-verify
is executed again to generate the development version of my blog, so I'll be able to write the next post.pelican
Now, whenever I push to
source
, and only to source
, master
gets updated with the new contents. Cool!One last small detail: I added
output
to the .gitignore
file. This way, the source
branch won't include this folder. We don't really want to put it under version control - it would be like putting other types of generated files such as .pyc
or .o
under version control.Originally published at anotherdatum.com.