shnarazk / advent-of-code Goto Github PK

• 2022-08-20 (after solved 3 day puzzles)
  

• Not a cube: 6 rects. It has 8 triangles.
  

Anyhow

Rust side

Copy the idea used in the BQN program, which uses the upper bounds of robots.

records

introduces the upper bounds of robots as the BQN version

BQN side

The following change is a degression.

diff --git a/bqn/2022/day19.bqn b/bqn/2022/day19.bqn
index 033c888..cfd2262 100755
--- a/bqn/2022/day19.bqn
+++ b/bqn/2022/day19.bqn
@@ -15,13 +15,12 @@ Solve ⇐ { part 𝕊 line:
     𝕊 ⟨⟩: best;
     𝕊 to_visit:
       𝕊 ⟨⟩ { time‿resources‿robots 𝕊 next:
-          { 𝕊 i:
+          next { i 𝕊 next:
           upto< t ← time- w ← 1+⌈⌈´(i⊏masks)/robots÷˜(n ← i⊏bp)(0⊸⌈-)resources ?
-              { ¬∨´(∧´𝕩⊸≤)¨next ? next ⟨𝕩⟩⊸∾ ↩ ;@ } t∾(n-˜resources+w×robots)∾((1⊸+)⌾(i⊸⊑)robots)
+              { ¬∨´(∧´𝕩⊸≤)¨next ? ⟨𝕩⟩∾next ; next } t∾(n-˜resources+w×robots)∾((1⊸+)⌾(i⊸⊑)robots)
           ;
-              best ((¯1⊑resources)+(time-upto)ׯ1⊑robots)⊸⌈ ↩
-          }¨{robots<○(𝕩⊸⊑)limits }¨⊸/↕4
-          next
+              best ((¯1⊑resources)+(time-upto)ׯ1⊑robots)⊸⌈ ↩, next
+          }´{robots<○(𝕩⊸⊑)limits }¨⊸/↕4
       }´{⟨⊑𝕩,4↑1↓𝕩,5↓𝕩⟩}¨to_visit
   }
   (•Fmt n) lib.Debug Expand ⟨⟨0⟩∾⟨0,0,0,0⟩∾⟨1,0,0,0⟩⟩ # ⟨time,resources,robots⟩

2022-08-05

On the day solving the 3rd day's problem.

Is this a path-finding problem?

• use shape instead of count width and hight
• use Over
• define uncurry operator

But all operators are evaluated once. So hardly expect speedup.

• 2023 day18: 可視化するまで初期化に問題があるとは気づかなかった。
• 2023 day20: 可視化するまで問題の構造に合わせた戦略を取れなかった。
• 2023 day21: 可視化するまで処理過程のイメージが持てなかった。

see #21
But in Pharo.
Add --serialize option.

flowchart LR
    rust/serde -- serialize --> Id1[(JSON)]
Id1[(JSON)] --> Pharo/Roassal

[Metacello new
    baseline: 'Roassal';
    repository: 'github://pharo-graphics/Roassal';
    load: 'Full' ] on: MCMergeOrLoadWarning do: [:warning | warning load ]

| data nodes canvas lb elements |
data :='/Users/nash/Repositories/advent-of-code/day25.json' asFileReference readStreamDo: [ :stream |  STON fromStream: stream ].
nodes := data inject: (Dictionary new) into: [ :dic :element | dic at: (element at:1) update: [ :set | set add: (element at: 2);yourself] initial: [Set new add: (element at: 2); yourself ]; yourself].
data do: [ :each | | to | to := each at: 2.  (nodes keys includes: to) not  ifTrue: [ nodes at: to put: { }  ]  ].
canvas := RSCanvas new.
elements := nodes keys collect: [ :c | RSEllipse new size: 5; model: c] as: RSGroup.
elements @ RSDraggable .
canvas addAll:  elements .
lb := RSLineBuilder line.
lb canvas: canvas.
lb color: Color red.
lb connectFromAll: [:each | nodes at: each ].
RSForceBasedLayout new on: elements.
canvas @ RSCanvasController.
lb canvas.

under development

• check if it's really symmetric. -- I got it. It is!

• problem

• rust version

• bqn/2022/day07.bqn

• does not finish and

• is a cheat so far.

flow

$ cargo run -- 2022 23 0 0 > 2022-12-23.json

func _ready():
	var file = FileAccess.open("res://2022-12-23.json", FileAccess.READ)
	var content = file.get_as_text()
	var j = JSON.parse_string(content)
	for elf in j:
		var c = cube.instantiate()
		c.position.x = float(elf["x"])
		c.position.y = float(elf["y"])
		add_child(c)

Cuis requires Json package to handle it.

s := 'Repositories/GD4.1/2022-12-23.json' fileContents.
j := Json readFrom: (ReadStream on: s from: 0 to: s size).
(j at: 1 ) at: 'x'

The updating expression might contain self-reference, that disabled the early release of memory.

https://uiua.org/pad?src=4oaSKOKIteKnuy7iio8pIDIgMF8yIHsiYWJjZSIgIiIgInh5In0KLgrihpIoK8OXIMKvMV8xKQrihpIoXCviiKco4oqC4oaSISlbXeKItSjilqE9MOKHoSkpCuKGmeKItltd

https://uiua.org/pad?src=4oaSKOKIteKnuy7iio8pIDIgMF8yIHsiYWJjZSIgIiIgInh5In0KLgrihpIoK8OXIMKvMV8wKQrihpIoXCviiKco4oqC4oaSISlbXeKItSjilqE9MOKHoSkpCuKKguKNmzLihpniiLZbXQrihpIv4oqCCuKGkigvKOKKguKItyEpKQrihpgx4oqV4pah

https://uiua.org/pad?src=4oaS4oaS4oaSKOKIteKnuy7iio8p4oi24oaS4oi2LuKGki7ihpLihpIuIDIgMF8yIHsiYWJjZSIgIiIgInh5In0K4oaS4oaSKCvDlyDCrzFfMCkK4oaS4oaSKFwr4oinKOKKguKGkiEpW13iiLUo4pahPTDih6EpKQrihpIo4oqC4o2bMuKGmeKItltdKQrihpLihpIv4oqCCuKGkuKGkigvKOKKguKItyEpKQrihpIo4oaYMeKKleKWoSkK4o2c4oqPXiDiiLYK4oaSOw==

https://uiua.org/pad?src=KOKHjCkyXzBfMiB7ImFiY2UiICIiICJ4eSJ9CuKGkuKGkuKGkijiiLXip7su4oqPKeKGkigu4oeJ4oqiKOKGmDEpKSDihpLihpLiiLbihpLiiLcuCuKGkuKGkigrw5cgwq8xXzApCuKGkuKGkihcK-KIpyjiioLihpIhKVtd4oi1KOKWoT0w4oehKSkK4oaSKOKKguKNmzPihpniiLZbXSkK4oaS4oaSKC8n4oqC4oi3ISkK4oaS4oqc4pahCuKNnCjiiqEwKQriiLbih4kn4oqhMCfihpgxCuKGkifihpgx4pah4o2cKOKKoTEpKOKWoeKKguKItyHiiLYpCuKGkjvijZziio8=

Not make sense to define type Answer.

• rename to advent-of-code
• add a new branch or directory and set it as default
• cargo clippy
• cargo update
• documentation on my framework

• 2023 day18: 可視化するまで初期化に問題があるとは気づかなかった。
• 2023 day20: 可視化するまで問題の構造に合わせた戦略を取れなかった。
• 2023 day21: 可視化するまで処理過程のイメージが持てなかった。

see #21
But in Pharo.
Add --serialize option.

flowchart LR
    rust/serde -- serialize --> Id1[(JSON)]
Id1[(JSON)] --> Pharo/Roassal

[Metacello new
    baseline: 'Roassal';
    repository: 'github://pharo-graphics/Roassal';
    load: 'Full' ] on: MCMergeOrLoadWarning do: [:warning | warning load ]

Where should I put them?

Added https://github.com/shnarazk/learn-bqn

!/usr/bin/env bqn
# Advent of Code 2022 Day 01

input ← ⟨199, 200, 208, 210, 200, 207, 240, 269, 260, 263⟩
•Show +´{(0⊸⊏𝕩)<(1⊸⊏𝕩)}˘(2⊸↕)input
•Show +´{(0⊸⊏𝕩)<(1⊸⊏𝕩)}˘(2⊸↕)+´˘3↕ input

• problem

• rust version

• bqn/2022/day05.bqn

• It's a cheat

• under construction

progress!("{x}");
is expanded to
dbg!("\[33m??{x}\[33m");

• Dim2<usize>, Vec2<usize>
• Vec2<isize>, Dim2<isize>
• Dim3<usize>, Vec3<usize>
• Vec3<isize>, Dim3<isize>

https://codegolf.stackexchange.com/questions/237866/tips-for-golfing-in-bqn

y2021 day5

xで順序付け

2本の直線と二つの線分の全交点を求める問題：線分と直線の交差はするかしないかだけ

The next drop traces the same path as the previous one!
So it'll help to speed up the BQN version.

old: ./aoc.bqn 2022 14 2  95.91s user 0.07s system 99% cpu 1:36.12 total
new: ./aoc.bqn 2022 14 2  95.73s user 0.19s system 99% cpu 1:36.12 total

On the flip side:

cargo run --release -- 2022 14 2  0.17s user 0.02s system 65% cpu 0.297 total

Again!

Stop checking all points in the solution space.
It should be on the edge of one of the sensing spaces.
So the complexity should be $O(4\times\mbox{sensor-range})$ instead of $O(400000^2)$.
Make an iterator traversing the edge of a sensor.

This was a good idea. The checking cost seems high.