Advent of Code 2017

Warning: Spoilers

Reflections and Benchmarks

I try to reflect on every day's puzzle, attempting to describe my thought processes and how my solutions all work. Benchmarks also included.

Day 1 Reflections (code) (benchmarks)
Day 2 Reflections (code) (benchmarks)
Day 3 Reflections (code) (benchmarks)
Day 4 Reflections (code) (benchmarks)
Day 5 Reflections (code) (benchmarks)
Day 6 Reflections (code) (benchmarks)
Day 7 Reflections (code) (benchmarks)
Day 8 Reflections (code) (benchmarks)
Day 9 Reflections (code) (benchmarks) (stream)
Day 10 Reflections (code) (benchmarks) (stream)
Day 11 Reflections (code) (benchmarks)
Day 12 Reflections (code) (benchmarks)
Day 13 Reflections (code) (benchmarks)
Day 14 Reflections (code) (benchmarks)
Day 15 Reflections (code) (benchmarks)
Day 16 Reflections (code) (benchmarks)
Day 17 Reflections (code) (benchmarks)
Day 18 Reflections (code) (benchmarks)
Day 19 Reflections (code) (benchmarks)
Day 20 Reflections (code) (benchmarks)
Day 21 Reflections (code) (benchmarks)
Day 22 Reflections (code) (benchmarks)
Day 23 Reflections (code) (benchmarks)
Day 24 Reflections (code) (benchmarks)
Day 25 Reflections (code) (benchmarks)

Executable

Comes with test examples given in problems.

You can install using stack:

$ git clone https://github.com/mstksg/advent-of-code-2017
$ cd advent-of-code-2017
$ stack setup
$ stack install

The executable aoc2017 includes a testing and benchmark suite

$ aoc2017 --help
aoc2017 - Advent of Code 2017 challenge runner

Usage: aoc2017 DAY [PART] [-t|--tests] [-b|--bench]
  Run challenges from Advent of Code 2017

Available options:
  DAY                      Day of challenge (1 - 25), or "all"
  PART                     Challenge part (a, b, c, etc.)
  -t,--tests               Run sample tests
  -b,--bench               Run benchmarks
  -h,--help                Show this help text

$ aoc2017 5 b
>> Day 05b
>> [✓] 27720699

Benchmarking is implemented using criterion

$ aoc2017 2 --bench
>> Day 02a
benchmarking...
time                 729.1 μs   (695.0 μs .. 784.2 μs)
                     0.967 R²   (0.926 R² .. 0.995 R²)
mean                 740.4 μs   (711.9 μs .. 783.6 μs)
std dev              116.8 μs   (70.44 μs .. 172.8 μs)
variance introduced by outliers: 89% (severely inflated)

>> Day 02b
benchmarking...
time                 782.4 μs   (761.3 μs .. 812.9 μs)
                     0.983 R²   (0.966 R² .. 0.998 R²)
mean                 786.7 μs   (764.1 μs .. 849.4 μs)
std dev              110.8 μs   (42.44 μs .. 228.5 μs)
variance introduced by outliers: 86% (severely inflated)

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

$ aoc2017 1 --tests
[9] [!35732] $ aoc2017 1 --tests
>> Day 01a
[✓] (3)
[✓] (4)
[✓] (0)
[✓] (9)
[✓] Passed 4 out of 4 test(s)
[✓] 1097
>> Day 01b
[✓] (6)
[✓] (0)
[✓] (4)
[✓] (12)
[✓] (4)
[✓] Passed 5 out of 5 test(s)
[✓] 1188

This should only work if you're running aoc2017 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.

aoc2017 will download missing input files, but requires a session token. This can be provided in aoc2017-conf.yaml:

session:  [[ session token goes here ]]

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 store 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.

tommyengstrom/advent-of-code-2017

Advent of Code 2017

Reflections and Benchmarks

Executable