I’ve decided it’s time to use GitHub Gist for posting code snippets in my blog. My original motivation was better syntax highlighting. I was using WP Code Highlight, which isn’t bad. It’s easy to use and does what it promises. But it doesn’t add line numbers and the syntax highlighting doesn’t seem optimal for R. GitHub Gist on the other hand handles R code beautifully.

Now highlighted code is nice, but what’s really cool about Gist is that the code is actually stored in a repository. This means I can update the code without having to update my WordPress post. And it’s easier for other people to use should they ever want to use my code. And there are other benefits as well. I guess the only bad thing is that GitHub has my code, which would be bad if they went belly-up. I’m pretty sure that won’t happen though, because I don’t want to think about it.

So here’s how you use GitHub Gist with WordPress to host R code. First create an account on GitHub and go to GitHub Gist. Where it says “Name of file”, type the name of your file and add “.r” to the end. Next select R from the language pulldown, though Gist will probably select it for you if you typed a file name ending in “.r”. Now copy and paste your code into the editor and click “create public gist”. You’re done with GitbHub

To make this block of code appear in your WordPress post, copy the code in the “Embed this gist” field and paste into your post. (Make sure you’re in the Text editor instead of the Visual editor). It would be really nice if that was the end of it, but it’s not. WordPress strips out the embed code and nothing happens. The workaround is to add a little plug-in called Raw HTML. As the description says, “This plugin adds a set of shortcodes that you can use to “protect” specific parts of your post and prevent WP from messing with them.” Awesome, just what we need. So the final step is to surround the Gist embed code with [raw]…[/raw] shortcode. (A tip of my hat to this site for telling me about Raw HTML.

Let’s try this out. Here’s some code that samples 30 values from a N(15,2) distribution, calculates a 95% confidence interval, and plots the interval as a line. It does this 40 times and colors the CI’s that don’t capture 15 red.

To make that appear in my post, I added the following:

[raw]
<script src=”https://gist.github.com/clayford/7842136.js”></script>
[/raw]