Advent of Code 2023

2015 / 2017 / 2018 / 2019 / 2020 / 2021 / 2022 / 2023

It's the most wonderful time of the year!

My Advent of Code 2023 Haskell solutions here, along with an automated fetching, testing, running environment (powered by the advent-of-code-api library). The interactive development environment and runner/bench marker/viewer/tester has been pulled out here, so this is implemented as "fork" of it with my own solutions and reflections.

Check out the reflections --- more info below!

Reflections and Benchmarks

Challenge	Reflections	Code	Benchmarks
Day 1		x	x
Day 2		x	x
Day 3		x	x
Day 4		x	x
Day 5		x	x
Day 6		x	x
Day 7
Day 8		x	x
Day 9		x	x
Day 10
Day 11		x	x
Day 12
Day 13		x	x
Day 14		x	x
Day 15		x	x
Day 16		x	x
Day 17
Day 18
Day 19
Day 20
Day 21
Day 22
Day 23
Day 24
Day 25

:~> type

If you're looking at my actual github solutions, you'll notice that this year the solutions are implemented in terms of a :~> record type:

data a :~> b = MkSol
    { sParse :: String -> Maybe a    -- ^ parse input into an `a`
    , sSolve :: a      -> Maybe b    -- ^ solve an `a` input to a `b` solution
    , sShow  :: b      -> String     -- ^ print out the `b` solution for submission
    }

An a :~> b is a solution to a challenge expecting input of type a and producing answers of type b. It also packs in functions to parse a String into an a, and functions to show a b as a String to submit as an answer.

This helps separate out parsing, solving, and showing, allowing for some cleaner code and an easier time planning the solution.

Such a challenge can be "run" on string inputs by feeding the string into sParse, then sSolve, then sShow:

-- | Run a ':~>' on some input, retuning 'Maybe'
runSolution :: Challenge -> String -> Maybe String
runSolution MkSol{..} s = do
    x <- sParse s
    y <- sSolve x
    pure (sShow y)

In the actual library, runSolution returns an Either so one can debug which stage the error happened in.

You might also notice the function dyno_, used like dyno_ "limit" 10000. This is how parameters are implemented in problems that vary between test data and actual input. For example, 2018 Day 6 involved finding points that had a total distance of less than 10000, but for the test input, we found the points that had a total distance of less than 32. So, I have a system that lets me write dyno_ "limit" 10000 in my code instead of hard-coding in 10000. This 10000 would be replaced by 32 when running with test data.

Interactive

The AOC.Run.Interactive module has code (powered by advent-of-code-api) for testing your solutions and submitting within GHCI, so you don't have to re-compile. If you edit your solution programs, they are automatically updated when you hit :r in ghci.

ghci> execSolution_   $ solSpec 'day02a   -- get answer for challenge based on solution
ghci> testSolution_   $ solSpec 'day02a   -- run solution against test suite
ghci> viewPrompt_     $ solSpec 'day02a   -- view the prompt for a part
ghci> waitForPrompt_  $ solSpec 'day02a   -- count down to the prompt for a part
ghci> submitSolution_ $ solSpec 'day02a   -- submit a solution

These are loaded with session key stored in the configuration file (see next section).

Executable

Comes with test examples given in problems.

You can install using stack:

$ git clone https://github.com/egnwd/advent
$ cd advent
$ stack setup
$ stack install

The executable aoc2023 includes a testing and benchmark suite, as well as a way to view prompts within the command line:

$ aoc2023 --help
aoc2023 - Advent of Code 2023 challenge runner

Usage: aoc2023 [-c|--config PATH] COMMAND
  Run challenges from Advent of Code 2023. Available days: 1, 2, 3 (..)

Available options:
  -c,--config PATH         Path to configuration file (default: aoc-conf.yaml)
  -h,--help                Show this help text

Available commands:
  run                      Run, test, and benchmark challenges
  view                     View a prompt for a given challenge
  submit                   Test and submit answers for challenges
  test                     Alias for run --test
  bench                    Alias for run --bench
  countdown                Alias for view --countdown

$ aoc2023 run 3 b
>> Day 03b
>> [✓] 243

You can supply input via stdin with --stdin:

$ aoc2023 run 1 --stdin
>> Day 01a
+1
+2
+1
-3
<Ctrl+D>
[?] 1
>> Day 01b
[?] 1

Benchmarking is implemented using criterion

$ aoc2023 bench 2
>> Day 02a
benchmarking...
time                 1.317 ms   (1.271 ms .. 1.392 ms)
                     0.982 R²   (0.966 R² .. 0.999 R²)
mean                 1.324 ms   (1.298 ms .. 1.373 ms)
std dev              115.5 μs   (77.34 μs .. 189.0 μs)
variance introduced by outliers: 65% (severely inflated)

>> Day 02b
benchmarking...
time                 69.61 ms   (68.29 ms .. 72.09 ms)
                     0.998 R²   (0.996 R² .. 1.000 R²)
mean                 69.08 ms   (68.47 ms .. 69.99 ms)
std dev              1.327 ms   (840.8 μs .. 1.835 ms)

Test suites run the example problems given in the puzzle description, and outputs are colorized in ANSI terminals.

$ aoc2023 test 1
>> Day 01a
[✓] (3)
[✓] (3)
[✓] (0)
[✓] (-6)
[✓] Passed 4 out of 4 test(s)
[✓] 416
>> Day 01b
[✓] (2)
[✓] (0)
[✓] (10)
[✓] (5)
[✓] (14)
[✓] Passed 5 out of 5 test(s)
[✓] 56752

This should only work if you're running aoc2023 in the project directory.

To run on actual inputs, the executable expects inputs to be found in the folder data/XX.txt in the directory you are running in. That is, the input for Day 7 will be expected at data/07.txt.

aoc2023 will download missing input files, but requires a session token. This can be provided in aoc-conf.yaml:

session:  [[ session token goes here ]]

Session keys are also required to download "Part 2" prompts for each challenge.

You can "lock in" your current answers (telling the executable that those are the correct answers) by passing in --lock. This will lock in any final puzzle solutions encountered as the verified official answers. Later, if you edit or modify your solutions, they will be checked on the locked-in answers.

These are stored in data/ans/XXpart.txt. That is, the target output for Day 7 (Part 2, b) will be expected at data/ans/07b.txt. You can also manually edit these files.

You can view prompts: (use --countdown to count down until a prompt is released, and display immediately)

$ aoc2023 view 3 b
>> Day 03b
--- Part Two ---
----------------

Amidst the chaos, you notice that exactly one claim doesn't overlap by
even a single square inch of fabric with any other claim. If you can
somehow draw attention to it, maybe the Elves will be able to make
Santa's suit after all!

For example, in the claims above, only claim `3` is intact after all
claims are made.

*What is the ID of the only claim that doesn't overlap?*

You can also submit answers:

$ aoc2023 submit 1 a

Submissions will automatically run the test suite. If any tests fail, you will be asked to confirm submission or else abort. The submit command will output the result of your submission: The message from the AoC website, and whether or not your answer was correct (or invalid or ignored). Answers that are confirmed correct will be locked in and saved for future testing against, in case you change your solution.

All networking features are powered by advent-of-code-api.