Leveraging Elixir's Enum Module

I was recently tasked with counting the occurrences of tags for some blog posts. The blog posts where in Markdown with some YAML like frontmatter. I'd already parsed out the front matter and each post had a tags property with a string of tags separated by a comma.

My initial thought was stick it in a GenServer and iterate over the posts using the GenServer's state to keep count of things. I then reasoned that could be done just as simple with an Elixir module as well. Both could work in theory, but neither is great--or leveraging the Elixir language.

Once it was too late to change my mind it hit me--using Enum.map and Enum.reduce was the obvious answer and far simpler answer! Frustrated I decided to write a quick module to see if I was write (if too late to matter!).

We'll assume we've already parsed all the posts and have them in a nice list with a key "tags".

posts = [
  %{"tags" => "Web, Development, WebAssembly, Performance"},
  %{"tags" => "Web, Development"},
  %{"tags" => "Web, Performance"}
]

First, we'll clean up our data so that we have a list of just the tags.

posts
|> Enum.flat_map(fn x -> [x["tags"]] end)
|> Enum.map(fn x -> String.split(x, ", ") end)
|> List.flatten()

# ["Web", "Development", "WebAssembly", "Performance", "Web", "Development", "Web", "Performance"]

Perfect! Now two simple functions will be combined to count the tags. count/1 uses Enum.reduce to go over our list. Initially, I've had issues finding uses for reduce, but once I realized the accumulator could be a simple Map, it opened up a bunch of opportunities. This will iterate over our list of tags, for each one invoke our update_count/2 function along with the accumulator. After each item in the list it'll store the result of our function back into the accumulator and pass that on to the next item. Once done it'll return our accumulator.

def count(tags) do
  Enum.reduce(tags, %{}, &update_count/2)
end

The real magic happens in update_count/2! We'll take our accumulator and tag, and call Map.update on our accumulator. After converting our tag from a string to an atom, we'll either add it to the map if it doesn't exist with a default value of 1, otherwise we'll update the existing value by increasing it by 1. This way as the same tag keeps showing up we'll keep increasing it's value in our map.

def update_count(tag, acc) do
  Map.update(acc, String.to_atom(tag), 1, &(&1 + 1))
end

After all is said and done we get a nice map with the data we need.

%{Development: 2, Performance: 2, Web: 3, WebAssembly: 1}

I often forget to look at the tools Elixir provides you to solve many problems, Enum and Map are powerful tools that should be leveraged whenever possible! The cherry on top is this solution is also far more functional in nature, versus a rather procedural approach I first imagined.