gladvent - v2.2.0

Gladvent

Hex Package Hex.pm Hex Docs GitHub Workflow Status

An Advent Of Code runner for Gleam

This library is intended to be imported to your gleam project and used as a command runner for your advent of code project in gleam.

To add this library to your project run: gleam add gladvent and add import gladvent to your main gleam file.

This package works only on gleam’s erlang target!

Due to changes made in gleam 1.5, users that were calling gladvent via gleam run -m should upgrade to v2 and add a call to gladvent.run in their project main function.

Using the library

Multi-year support

Gladvent now comes with out-of-the-box multi-year support via the --year flag when running it.

For convenience it defaults to the current year. Therefore, passing --year=YEAR to either the run, run all or new commands will use the year specified or the current year if the flag was not provided.

Seeing help messages

General Workflow

Where X is the day you’d like to add:

Note: this method requires all day solutions be in src/aoc_<year>/ with filenames day_X.gleam, each solution module containing fn pt_1(String) -> Int and a fn pt_2(String) -> Int

  1. run gleam run new X
  2. add your input to input/<YEAR>/X.txt
  3. add your code to src/aoc_<YEAR>/day_X.gleam
  4. run gleam run run X

Available commands

This project provides your application with 2 command groups, new and run:

New

Run

The run command expects input files to be in the input/<year> directory, and code to be in src/aoc_<year>/ (corresponding to the files created by the new command).

Note:

Part 1: error: todo - unimplemented in module aoc_2024/day_1 in function pt_1 at line 2

Fetching problem inputs

When stubbing out a new day’s solution with the new command, you can use the --fetch flag to tell gladvent to fetch your problem input from the advent of code website.

Some things to note:

Reusable parse funtions

Gladvent supports modules with functions that provide a pub fn parse(String) -> a where the type a matches with the type of the argument for the runner functions pt_1 and pt_2. If this parse function is present, gladvent will pick it up and run it only once, providing the output to both runner functions.

An example of which looks like this:

pub fn parse(input: String) -> Int {
    let assert Ok(i) = int.parse(input)
    i
}

pub fn pt_1(input: Int) -> Int {
    input + 1
}

pub fn pt_2(input: Int) -> Int {
    input + 2
}

Note: gladvent now leverages gleam’s export package-interface functionality to type-check your parse and pt_{1|2} functions to make sure that they are compatible with each other.

Defining expectations for easy refactoring

One of the most satisfying aspects of advent of code (for me), second only to that sweet feeling of first solving a problem, is iteration and refactoring.

Gladvent makes it easy for you to define expected outputs in your gleam.toml for all your solutions so that you can have the confidence to refactor your solutions as much as you want without having to constantly compare with your submissions on the advent of code website.

Expectations in gleam.toml

Defining expectations is as simple as adding sections to your gleam.toml in the following format:

[gladvent.<year as int>]
1 = { pt_1 = <int or string>, pt_2 = <int or string> }
2 = { pt_1 = <int or string>, pt_2 = <int or string> }
3 = { pt_1 = <int or string>, pt_2 = <int or string> }
...

For example, to set the expectations for Dec 1st 2024 (2024 day 1) you would add something like:

[gladvent.2024]
1 = { pt_1 = 1, pt_2 = 2 }

When running, gladvent will detect whether a specific day has it’s expectations set and if so will print out the result for you.

Let’s say that your computed solution for 2024 day 1 is actually 1 for pt_1 and 3 for pt_2, the output will look like this:

Ran 2024 day 1:
  Part 1: ✅ met expected value: 1
  Part 2: ❌ unmet expectation: got 3, expected 2

Example inputs

Sometimes it’s helpful to run advent of code solutions against example inputs to verify expectations. Gladvent now provides a --example flag in both the new and run commands to conveniently support that workflow without needing to modify your actual problem input files. Example input files will be generated at and run from input/<year>/<day>.example.txt.

Note: gladvent will not compare your solution output against the expectations defined in gleam.toml when running in example mode.

Display execution time

Use the --timed flag when running your solutions to display how long each part took to solve.

For example:

Ran 2024 day 1:
  Part 1: ✅ met expected value: 1579939 (in 885 µs)
  Part 2: ✅ met expected value: 20351745 (in 605 µs)

Note: as the output of the parse function is reused for both parts, its execution time is not included in the displayed time.

FAQ

Why did you make this?

It seemed fun, I like small command line utilities and I wanted a way to get advent of code done in gleam without having the overhead of lots of copy-pasting and connecting things to get it to run.

Why run as a command line utility and not just use unit tests?

I thought a lot about that and I just prefer the overall interactivity of a CLI better, as well as allowing for endless runs or runs with configurable timeouts. Having it run as part of eunit doesnt provide as much flexibility as I would like. Other testing frameworks have been popping up but I leave the decision to use them up to you!

Why did you change your mind on fetching inputs?

I started to reflect a bit after gladvent’s users kept asking for the feature…

While my initial rationale was twofold:

  1. To encourage people to use the advent of code website, and I felt like fetching inputs somehow took away from that.
  2. To minimise the risk that people would use a tool I made to spam the advent of code website with requests.

Fetching inputs in a smart way (only ever if your input file does not already exist, so you should only need to do it once per day) still requires users to visit the advent of code website for the following (things gladvent will never do):

✨ Search Document