If there's one thing I like about Python is how I can use it to automate boring tasks for me. Today I used it to help me manage my own blog!

Let me start by describing what I needed to do and then I will go over how Python helped me do it.

The problem

The fact that I am using Grav to manage the contents of my website and the fact that I have my website both in Portuguese and in English means I end up having two text files per page. For example, if you follow this GitHub link you can see the subfolder that holds the content for this particular blog post. In there you can find three main files, which are

• item.en.md
• item.pt.md
• frontmatter.yaml

The item files contain the blog post itself, and the extension .en.md or .pt.md identifies the language in which they are written. The frontmatter.yaml file has "headers" that help customize my blog. I set those headers with a special syntax called YAML; with it I set the post tags, the title of the post, the slug (the URL suffix, yamlutils for this blog post), the date in which this was published, etc.

For most of it, these things do not depend on the language the user is viewing the page on, for example I keep the slug the same for the English and Portuguese versions, cf.

• https://mathspp.com/en/blog/yamlutils
• https://mathspp.com/pt/blog/yamlutils

Those headers that are the same, regardless of the language, are kept in the frontmatter.yaml file, which looks something like this:

On the other hand, there are other things that do depend on the language. For example, the title of the page that is displayed in the beginning of the post and in the browser tab. Those are specified in the beginning of the item files in between two sets of ---, which makes GitHub render them as such:

You can clearly see that while the structure of the boxes is the same, their contents are in different languages.

The problem is that for a long time I didn't use the frontmatter.yaml file to store the headers that were identical, meaning I have plenty of pages and blog posts that have duplicated headers in the .pt.md and .en.md files... this is not a nice situation to be in because it makes it harder to have everything updated and synced correctly. Of course I could sort this out by hand... but that would be really boring!

Python to the rescue

Being a scripting language, Python excels at this type of tasks. I realized I could easily write a Python script that would traverse my blog directory, looking for pairs of .pt.md and .en.md pages, finding the headers those pages have in common, and then updating the corresponding frontmatter.yaml (or creating a new one if needed).

That is how my little YAMLUtils project was born. The yamlutils.py script takes a folder path as a command line argument (and optionally a -r flag to traverse the directories recursively) and then does what I just described, merging all the YAML headers it can. I used it on my blog with python yamlutils.py pages/ -r and you can see what it did in this commit. Imagine having to do all that by hand!

More specifically, my script looks at the path you specified (and recurses if you set -r) and then, for each folder:

• looks for all files with a name matching the pattern *.*.md, which it interprets as being name.language.md;
• one by one, opens those .md files, extracts the YAML headers from within the --- and computes the headers they have in common;
• tries to see if there already is a frontmatter.yaml or frontmatter.yml file available (with preference for the .yaml) and loads the YAML headers there;
• updates the headers that were already in a frontmatter file (if any) and writes the updated version in the frontmatter file, creating a frontmatter.yaml file if needed;
• goes back to the *.*.md files, removing from their headers everything that was saved in the frontmatter file.

Was it worth the trouble?

Yes.

It took me around two hours to put the script together and to test it plus ten minutes to fix a bug that I only discovered when I later applied it to the blog (cf. this and this buggy usages). Doing all this by hand would easily have taken me over two hours. Plus, now I have a cool new script that I can extend later (and probably will) to do some more YAML management and I programmed a bit (which is always fun!) instead of copying and pasting around YAML headers.

Let me know in the comments below how Python has helped you automate any of your tasks!

For your convenience, the contents of my script at the time of writing this blog post:

Become a better Python 🐍 developer 🚀

+35 chapters. +400 pages. Hundreds of examples. Over 30,000 readers!

My book “Pydon'ts” teaches you how to write elegant, expressive, and Pythonic code, to help you become a better developer. >>> Download it here 🐍🚀.