Day 14: Restroom Redoubt
Puzzle DescriptionAdvent of Code 2024, Day 14
Scala Center PageScala Center Advent of Code 2024, Day 14
Solution Summary
- Parse input into list of robots
- Perform steps
- For part 1, we just have to do it 100 times
- For part 2, it’s technically unbounded
- For part 2, detect easter egg
Part 1
First, I’ll define Vec2i
:
case class Vec2i(x: Int, y: Int)
To accommodate testing, let’s define a case class GridSize
and a given
so we can change the size depending on the input we are using.
case class GridSize(x: Int, y: Int)
given gridSize: GridSize = GridSize(101, 103)
Now we can make a Robot
:
case class Robot(pos: Vec2i, velocity: Vec2i)
Let’s parse the input now:
def parse(str: String): List[Robot] = {
str.linesIterator.map { case s"p=$px,$py v=$vx,$vy" => Robot(Vec2i(px.toInt, py.toInt), Vec2i(vx.toInt, vy.toInt)) }.toList
}
An observation of the movement rules is that when we move forward n
steps, we don’t need to calculate in between moves. A modulo will suffice.
Of course the modulo in scala doesn’t work how we want, so we can define rem
:
extension (self: Int) {
infix def rem(that: Int): Int = {
// these don't cancel out due to floor div
self - that.abs * (self.toDouble / that.toDouble.abs).floor.toInt
}
}
With this we can add stepN
to Robot
:
case class Robot(pos: Vec2i, velocity: Vec2i) {
def stepN(n: Int = 1)(using size: GridSize): Robot = {
copy(pos = pos.copy(x = (pos.x + n * velocity.x) rem size.x, y = (pos.y + n * velocity.y) rem size.y))
}
}
To ease function calling on the List[Robot]
, let’s add extensions:
extension (robots: List[Robot])(using size: GridSize) {
def stepN(n: Int = 1): List[Robot] = robots.map(_.stepN(n))
}
Let’s also add the safety score as an extension as well:
extension (robots: List[Robot])(using size: GridSize) {
// ...
def safety: Int = {
val middleX = size.x / 2
val middleY = size.y / 2
robots.groupBy { robot =>
if (robot.pos.x == middleX || robot.pos.y == middleY) {
-1
} else if (robot.pos.x < middleX) {
if (robot.pos.y < middleY) {
0
} else {
2
}
} else {
if (robot.pos.y < middleY) {
1
} else {
3
}
}
}.removed(-1).values.map(_.length).product
}
}
Now we can finish part 1:
def part1(str: String): Int = {
val input = parse(str)
input.stepN(100).safety
}
Part 2
… What? Find an image????
I was completely and entirely lost on this one because it gives you 0 information other than that the pattern is a christmas tree. My original solution was based off scarf’s where you just generate images to a file and eyeball it. However, there are better solutions from others I’ve read. My favorite I’ve seen is from a solution from Philippus, where you test each map to see if there are 15 or more robots inline vertically or horizontally.
An important part of this problem is that the tree is actually framed by a rectangle. Testing for this rectangle is way easier than testing for the christmas tree.
First, we need to be able to actually see our output:
extension (robots: List[Robot])(using size: GridSize) {
// ...
def robotMap: Vector[Vector[Int]] = {
val goodGrid = mut.ArrayBuffer.fill(size.y, size.x)(0)
robots.foreach { robot =>
goodGrid(robot.pos.y)(robot.pos.x) = goodGrid(robot.pos.y)(robot.pos.x) + 1
}
goodGrid.map(_.toVector).toVector
}
def pretty: String = {
val grid = robots.robotMap
grid.map(_.map {
case 0 => '.'
case _ => '#'
}.mkString("", "", "")).mkString("", "\n", "")
}
}
Now, let’s implement Philippus’s solution.
extension (robots: List[Robot])(using size: GridSize) {
// ...
def findEasterEgg(): Unit =
for (i <- 0 to 10000) {
val newRobots = robots.stepN(i)
if (newRobots.groupBy(_._1.x).map(_._2.length).max > 15 && newRobots.groupBy(_._1.y).map(_._2.length).max > 15) {
val str = newRobots.pretty
println(s"$i seconds:\n" + str)
}
}
}
Unfortunately, this still requires some human double checking, but this is much better than manually checking every image. On my input it only outputs 5 images.
We can make this entirely programatic by making our test a little stricter: this tests if bots are inline with gaps, but the border on a tree is completely solid.
So we modify our findEasterEgg
function like so:
def longestSubseq[A](seq: List[A], item: A): Int =
seq.foldLeft((0, 0)):
case ((max, count), a) =>
if item == a then
(math.max(max, count + 1), count + 1)
else
(max, 0)
._1
extension (robots: List[Robot])(using size: GridSize) {
def findEasterEgg: Int =
(0 to 10000).find { i =>
val newRobots = robots.stepN(i)
if (newRobots.groupBy(_.pos.x).map(_._2.length).max > 15 && newRobots.groupBy(_.pos.y).map(_._2.length).max > 15) {
val xLineMembers = newRobots.groupBy(_.pos.x).maxBy(_._2.length)._2.map(_.pos.y).toSet
val yLineMembers = newRobots.groupBy(_.pos.y).maxBy(_._2.length)._2.map(_.pos.x).toSet
val xLine = (0 until size.y).map(it => xLineMembers.contains(it)).toList
val yLine = (0 until size.x).map(it => yLineMembers.contains(it)).toList
longestSubseq(xLine, true) > 15 && longestSubseq(yLine, true) > 15
} else false
}.getOrElse(-1)
}
This checks the line with the most members and finds the length of the longest subsequence. If these are both higher than 15 then we found the christmas tree.
This works - no need to double check with human eyes.
We can now solve part 2 fully programatically:
def part2(str: String): Int =
val input = parse(str)
input.findEasterEgg
Final code:
case class Vec2i(x: Int, y: Int)
case class GridSize(x: Int, y: Int)
given gridSize: GridSize = GridSize(101, 103)
extension (self: Int) {
infix def rem(that: Int): Int = {
// these don't cancel out due to floor div
self - that.abs * (self.toDouble / that.toDouble.abs).floor.toInt
}
}
case class Robot(pos: Vec2i, velocity: Vec2i) {
def stepN(n: Int = 1)(using size: GridSize): Robot = {
copy(pos = pos.copy(x = (pos.x + n * velocity.x) rem size.x, y = (pos.y + n * velocity.y) rem size.y))
}
}
def longestSubseq[A](seq: List[A], item: A): Int =
seq.foldLeft((0, 0)):
case ((max, count), a) =>
if item == a then
(math.max(max, count + 1), count + 1)
else
(max, 0)
._1
extension (robots: List[Robot])(using size: GridSize) {
def stepN(n: Int = 1): List[Robot] = robots.map(_.stepN(n))
def safety: Int = {
val middleX = (size.x / 2)
val middleY = (size.y / 2)
robots.groupBy { robot =>
if (robot.pos.x == middleX || robot.pos.y == middleY) {
-1
} else if (robot.pos.x < middleX) {
if (robot.pos.y < middleY) {
0
} else {
2
}
} else {
if (robot.pos.y < middleY) {
1
} else {
3
}
}
}.removed(-1).values.map(_.length).product
}
def pretty: String = {
val grid = robots.robotMap
grid.map(_.map {
case 0 => '.'
case _ => '#'
}.mkString("", "", "")).mkString("", "\n", "")
}
def findEasterEgg: Int =
(0 to 10000).find { i =>
val newRobots = robots.stepN(i)
if (newRobots.groupBy(_.pos.x).map(_._2.length).max > 15 && newRobots.groupBy(_.pos.y).map(_._2.length).max > 15) {
val xLineMembers = newRobots.groupBy(_.pos.x).maxBy(_._2.length)._2.map(_.pos.y).toSet
val yLineMembers = newRobots.groupBy(_.pos.y).maxBy(_._2.length)._2.map(_.pos.x).toSet
val xLine = (0 until size.y).map(it => xLineMembers.contains(it)).toList
val yLine = (0 until size.x).map(it => yLineMembers.contains(it)).toList
longestSubseq(xLine, true) > 15 && longestSubseq(yLine, true) > 15
} else false
}.getOrElse(-1)
def robotMap: Vector[Vector[Int]] = {
val goodGrid = mut.ArrayBuffer.fill(size.y, size.x)(0)
robots.foreach { robot =>
goodGrid(robot.pos.y)(robot.pos.x) = goodGrid(robot.pos.y)(robot.pos.x) + 1
}
goodGrid.map(_.toVector).toVector
}
}
def part1(str: String): Int =
val input = parse(str)
input.stepN(100).safety
def part2(str: String): Int =
val input = parse(str)
input.findEasterEgg
My Solution on GitHub