Day 19: Aplenty
Megathread guidelines
- Keep top level comments as only solutions, if you want to say something other than a solution put it in a new post. (replies to comments can be whatever)
- You can send code in code blocks by using three backticks, the code, and then three backticks or use something such as https://topaz.github.io/paste/ if you prefer sending it through a URL
FAQ
- What is this?: Here is a post with a large amount of details: https://programming.dev/post/6637268
- Where do I participate?: https://adventofcode.com/
- Is there a leaderboard for the community?: We have a programming.dev leaderboard with the info on how to join in this post: https://programming.dev/post/6631465
Scala3
case class Part(x: Range, m: Range, a: Range, s: Range): def rating: Int = x.start + m.start + a.start + s.start def combinations: Long = x.size.toLong * m.size.toLong * a.size.toLong * s.size.toLong type ActionFunc = Part => (Option[(Part, String)], Option[Part]) case class Workflow(ops: List[ActionFunc]): def process(p: Part): List[(Part, String)] = @tailrec def go(p: Part, ops: List[ActionFunc], acc: List[(Part, String)]): List[(Part, String)] = ops match case o :: t => o(p) match case (Some(branch), Some(fwd)) => go(fwd, t, branch::acc) case (None, Some(fwd)) => go(fwd, t, acc) case (Some(branch), None) => branch::acc case (None, None) => acc case _ => acc go(p, ops, List()) def run(parts: List[Part], workflows: Map[String, Workflow]) = @tailrec def go(parts: List[(Part, String)], accepted: List[Part]): List[Part] = parts match case (p, wf) :: t => val res = workflows(wf).process(p) val (acc, rest) = res.partition((_, w) => w == "A") val (rej, todo) = rest.partition((_, w) => w == "R") go(todo ++ t, acc.map(_._1) ++ accepted) case _ => accepted go(parts.map(_ -> "in"), List()) def parseWorkflows(a: List[String]): Map[String, Workflow] = def generateActionGt(n: Int, s: String, accessor: Part => Range, setter: (Part, Range) => Part): ActionFunc = p => val r = accessor(p) (Option.when(r.end > n + 1)((setter(p, math.max(r.start, n + 1) until r.end), s)), Option.unless(r.start > n)(setter(p, r.start until math.min(r.end, n + 1)))) def generateAction(n: Int, s: String, accessor: Part => Range, setter: (Part, Range) => Part): ActionFunc = p => val r = accessor(p) (Option.when(r.start < n)((setter(p, r.start until math.min(r.end, n)), s)), Option.unless(r.end <= n)(setter(p, math.max(r.start, n) until r.end))) val accessors = Map("x"->((p:Part) => p.x), "m"->((p:Part) => p.m), "a"->((p:Part) => p.a), "s"->((p:Part) => p.s)) val setters = Map("x"->((p:Part, v:Range) => p.copy(x=v)), "m"->((p:Part, v:Range) => p.copy(m=v)), "a"->((p:Part, v:Range) => p.copy(a=v)), "s"->((p:Part, v:Range) => p.copy(s=v))) def parseAction(a: String): ActionFunc = a match case s"$v<$n:$s" => generateAction(n.toInt, s, accessors(v), setters(v)) case s"$v>$n:$s" => generateActionGt(n.toInt, s, accessors(v), setters(v)) case s => p => (Some((p, s)), None) a.map(_ match{ case s"$name{$items}" => name -> Workflow(items.split(",").map(parseAction).toList) }).toMap def parsePart(a: String): Option[Part] = a match case s"{x=$x,m=$m,a=$a,s=$s}" => Some(Part(x.toInt until 1+x.toInt, m.toInt until 1+m.toInt, a.toInt until 1+a.toInt, s.toInt until 1+s.toInt)) case _ => None def task1(a: List[String]): Long = val in = a.chunk(_ == "") val wfs = parseWorkflows(in(0)) val parts = in(1).flatMap(parsePart) run(parts, wfs).map(_.rating).sum def task2(a: List[String]): Long = val wfs = parseWorkflows(a.chunk(_ == "").head) val parts = List(Part(1 until 4001, 1 until 4001, 1 until 4001, 1 until 4001)) run(parts, wfs).map(_.combinations).sum
Nim
Part 1 was pretty straightforward. For part 2 I made an ItemRange type that’s just one integer range for each attribute. I also made a split function that returns two ItemRange objects, one for the values that match the specified rule, and the others for the unmatched values. When iterating through the workflows, I start a new recursion branch to process any matching values, and continue stepping through with the unmatched values until none remain or they’re accepted/rejected.
Python
6.528 line-seconds (ranks 8th hardest after days 17, 8, 12, 16, 14, 11 and 10).
import functools import operator import re import portion as P # noqa: N812 from .solver import Solver def isize(i: P.Interval): return sum(i_part.upper - i_part.lower - int(i_part.left == P.OPEN) + int(i_part.right == P.CLOSED) for i_part in i) class Day19(Solver): workflows: dict[str, list[str|tuple[str, str, int, str]]] parts: list[dict[str, int]] def __init__(self): super().__init__(19) def presolve(self, input: str): lines = input.splitlines() self.workflows = {} while lines: line = lines.pop(0) if not line: break name, program = line.split('{') instructions = program[:-1].split(',') self.workflows[name] = [] for item in instructions: match_condition = re.fullmatch(r'(\w+)([<>])(\d+):(\w+)', item) if match_condition: category, op, threshold, goto = match_condition.groups() self.workflows[name].append((category, op, int(threshold), goto)) else: self.workflows[name].append(item) self.parts = [] while lines: items = lines.pop(0)[1:-1].split(',') part = {} for category, value in (i.split('=') for i in items): part[category] = int(value) self.parts.append(part) def solve_first_star(self): return sum(sum(part.values()) for part in self.parts if self._count_options('in', 0, {c: P.singleton(v) for c, v in part.items()}) > 0) def solve_second_star(self): return self._count_options('in', 0, {c: P.closed(1, 4000) for c in self.parts[0].keys()}) def _count_options(self, workflow_name: str, workflow_index: int, ranges: dict[str, P.Interval]) -> int: if workflow_name == 'A': return functools.reduce(operator.mul, (isize(r) for r in ranges.values()), 1) if workflow_name == 'R': return 0 if any(isize(r) == 0 for r in ranges.values()): return 0 match self.workflows[workflow_name][workflow_index]: case (category, '>', threshold, goto): new_ranges_true = {c: r & P.open(threshold, P.inf) if c == category else r for c, r in ranges.items()} new_ranges_false = {c: r & P.openclosed(-P.inf, threshold) if c == category else r for c, r in ranges.items()} return (self._count_options(goto, 0, new_ranges_true) + self._count_options(workflow_name, workflow_index + 1, new_ranges_false)) case (category, '<', threshold, goto): new_ranges_true = {c: r & P.open(-P.inf, threshold) if c == category else r for c, r in ranges.items()} new_ranges_false = {c: r & P.closedopen(threshold, P.inf) if c == category else r for c, r in ranges.items()} return (self._count_options(goto, 0, new_ranges_true) + self._count_options(workflow_name, workflow_index + 1, new_ranges_false)) case next_workflow: return self._count_options(next_workflow, 0, ranges)
Nim
I optimized Part1 by directly referencing workflows between each rule (instead of doing a table lookup between them), in expectation of part 2 needing increased performance. But that turned out to not be needed 😋
I had to dig through my dusty statistics knowledge for part 2, and decided to try out Mermaid.js to create a little graph of the sample input to help visualize the solution.
After that it was pretty straightforward.