title: Jekyll Related Posts without Plugin time: 2014.07.15 22:57:38 layout: post tags:
site.related_posts may return recent posts when there aren't enough posts sharing the same tags with current post.What will you learn in this post?
This post introduces how to implement related posts function with Jekyll. Since GitHub disables plugin for security reasons, plugins like <a href="https://github.com/LawrenceWoodman/related_posts-jekyll_plugin" targe="_blank">LawrenceWoodman / related_posts-jekyll_plugin</a> cannot be used. Default
site.related_postsmay return recent posts when there aren't enough posts sharing the same tags with current post.
Jekyll has native support for related posts function, which can be used with Liquid as site.related_posts.
However, as we can see from its <a href="https://github.com/jekyll/jekyll/blob/df8458275de4dc3d0f9b92c5247ff20832d1cc8e/lib/jekyll/related_posts.rb" target="_blank">source code</a>, it lists posts with mutual tags without ordering by how many mutual tags they share. In this way, if A post is tagged with a, b and c, B post is tagged with a and C post is tagged with a and c, A's related posts may contain only B but not C if you set the limit to be 1. Apparently, this doesn't make sense.
<a href="https://github.com/LawrenceWoodman/related_posts-jekyll_plugin" targe="_blank">LawrenceWoodman / related_posts-jekyll_plugin</a> is an easy-to-use Jekyll plugin to implement related posts feature.
It solves the above problem by calculating how many mutual tags they share and sort by it, as we can see from the <a href="https://github.com/LawrenceWoodman/related_posts-jekyll_plugin/blob/master/_plugins/related_posts.rb" target="_blank">source code</a> below.
{% highlight ruby %} def related_posts(posts) return [] unless posts.size > 1 highest_freq = Jekyll::Post.tag_freq(posts).values.max related_scores = Hash.new(0) posts.each do |post| post.tags.each do |tag| if self.tags.include?(tag) && post != self cat_freq = Jekyll::Post.tag_freq(posts)[tag] related_scores[post] += (1+highest_freq-cat_freq) end end end
Jekyll::Post.sort_related_posts(related_scores)
end {% endhighlight %}
But just as the default Jekyll behaves, it returns recent posts if there aren't enough posts sharing the same tags. I think this doesn't make sense. After all, why should a post called How to become a good programmer be related to I was bitten by a bug yesterday just because there are no other related posts?
So, I improved it with the following code and <a href="https://github.com/Ovilia/related_posts-jekyll_plugin/commit/4562de1b048a5dbce5f533d343be30254427160e" target="_blank">pulled a request</a>.
{% highlight ruby %} related_scores.each do |post, score| if score < 0 related_scores.delete(post) end end {% endhighlight %}
I have almost no experience with Ruby, but this isn't too hard for me.
After I've done with the previous improvement and pushed my blog code to GitHub with joy, I found it didn't word! :scream:
I thought it was the cache, but it still doesn't change when I tried to clear the cache and even visit it using a different browser and a different device.
Finally, I realized that this is caused by GitHub's disabling plugins for security concerns. I knew this before, but didn't know that this is the reason I was looking for.
Then, I improved the default one and return only those with mutual tags. It still cost me much time, since the syntax of Liquid is so strange! :weary:
<script src="https://gist.github.com/Ovilia/ea95e762544d84f00281.js"></script>You can't use !abc to check if abc is false. Instead, you should use abc == false.
You can't use abc++ or even abc = abc + 1 to increment a variable. Instead, you should use something as strange as {% capture abc %}{{ abc }}{% endcapture %} appending a character `at a time, and then checkabc.size` to get its length as a string and this is how you increment a variable! :joy: But, believe me, this is not the strangest syntax when you use Liquid.
Who can tell me why this syntax is so strange!