Is this a good way of implementing Blogging Functionality on GitHub Pages?

Question

I'm considering using GitHub Pages to implement a blog and I am kind of in the mood to reinvent the wheel. I've built my website from scratch and I don't want to use Jekyll because it doesn't seem like it can be integrated into an existing website with your own theme (correct me if I'm wrong). I want the following functionalities:

• Blog with articles but without the need to constantly update the html myself
• Ability to write in plain text on my computer and have my website parse the data and integrate into the website's html when I push to the git repository
• A feature on the website which shows a preview of the last three blog posts
• The ability to search for specific blog posts hosted on my page through google

I am considering both a static and dynamic solution. The static solution would be to write a python script that reads my blog in text format and updates the blog page to include this new article by adding more html to the blog html page. The dynamic solution would have my blog page fetch articles from the github repository and integrate them into my blog live. This would also allow for fetching more articles on demand (as the user scrolls down my page for example). Is this ridiculously over complicated or does this make sense to do as a little project? Any opinions would be appreciated!

Answer 1

I wrote a static site generator as a Grunt plugin a few years back, so I can probably offer some good advice here.

A static site generator will convert source files (typically Markdown or a similar format for the posts and pages and templates in your templating language of choice) into flat HTML which can then be pushed up to the server. It may also include scripts for pushing it to the place where the site is hosted. If you want to do this in Python there's plenty of appropriate modules - Jinja would be a good example of an appropriate templating system, there's several Markdown implementations, and if you need syntax highlighting Pygments is a solid choice.

It's actually possible to search on the client side. My site uses Lunr.js for this - during the build process it generates an index file for the search, and then that is loaded using jQuery with the rest of the page.

Infinite scroll should also be possible without server-side scripting as it just fetches the same content using Javascript anyway. Paul Irish's Infinite Scroll plugin is no longer maintained, but it would in theory be capable of this and there are probably alternatives.

You might want to check out the article I wrote about it for ideas. The static solution works well for me - it's cheap to host and easy to deploy. You can use Disqus or Facebook comments if need be too.

A simple static site generator is a practical weekend project, and it should be a good way of achieving what you want.