Benchmarks Elixir

Benchmarks to test some Elixir implementations.

asdf install
mix deps.get
mix run benchmarks/<name of the benchmark>

String templating at runtime

During runtime, you receive a string (the template) where variables are encoded in it. You also receive a list of bindings.

The expected output is a list of strings of the same size of the list of bindings. Each of those strings corresponds to the template, with the variables replaced by the binding values.

I noticed that this issue is often resolved in two steps : prepare the template in some way, and then apply the bindings. In many use-cases, those two steps are ran at different times. I therefore created a second bechmark, that only compares the methodes w.r.t. the time they take to apply the binding to the prepared template. I differentiate both benchmarks by calling them preparation included and preparation excluded.

Example

Template : "I like <%= music_genre %>, especially during <%= season %>."

Bindings : [[{:music_genre, "rock"}, {:season, "spring"}], [{:music_genre, "jazz"}, {:season, "summer"}]]

Expected output : ["I like rock, especially during spring.", "I like jazz, especially during summer."]

Use cases

1. You are designing a webpage editor. Therefore, some HTML pages are created by your users. When displaying those pages, you want to replace context variables in the page. You are in the preparation excluded scenario, since you can prepare the template when it is uploaded by the user, and then only the time to bind the variables matters.

2. You are provided with a template email that you need to send immediatly to a thousand people. That email template contains recipient-specific variables. Since you need to parse the template and send the emails immediately, you are in the preparation included scenario : the preparation time matters.