I was unable to upload even the shortest video because it was too long for my instance. Therefore, please enjoy the following:

  1. Partial visualization of test data. I cut this short because it took 40 seconds to do just a few (out of 81) paths: https://youtube.com/shorts/7UvzgSsMQNA
  2. Partial visualization of full data. I cut this short because I didn’t want to wait 40 minutes. It’s sped up 2x by making it 60fps (each step is approximately one frame) https://youtu.be/cv9qSdrV2Z4
  3. Full visualization, but it only shows the end paths, not individual steps: https://youtube.com/shorts/ozQ77ikI7JI

Unfortunately youtube is forcing my videos to be shorts due to aspect ratio and length, I don’t know if I can force them to a regular video

  • @[email protected]OP
    link
    fedilink
    26 days ago

    Here’s the code. Everything prefixed with ‘RS’ is from Roassal. Alternatively, just look at where I use canvas. Keep in mind this is quick spaghetti code, and it includes the full solution for day 10 part 2.

    day10p1vis: in
    	| input starts canvas |
    	
    	canvas := RSCanvas new.
    	input := CTNewArray2D
    		fromArray: ((in copyWithRegex: '\n|\r' matchesReplacedWith: '') 
    			collect: [ :v | v asString asInteger ] as: Array)
    		width: (in lineNumber: 1) size.	
    	
    	input withIndexesDo: [ :x :y :val |
    		canvas add: (RSLabel new size: 5; text: val; x: x * 12; y: y * 12) ].
    	canvas signalUpdate.
    	(canvas @ RSCanvasController) open.
    	14 seconds wait.
    	starts := (input contents indicesOf: 0) collect: [:v | input indexToCoord: v].
    	
    	d9heads := OrderedCollection new.
    	^ starts sumNumbers: [ :s | | thead |
    		canvas signalUpdate.
    		0.08 seconds wait.
    		d9heads removeAll.
    		
    		thead := RSBox new size: 12; position: s * 12; color: Color red translucent.
    		canvas add: thead.
    	
    		self d9graphvis: input start: s dir: -1@0 canvas: canvas.
    	]
    
    d9graphvis: input start: start dir: dir canvas: canvas
    	| next sum direc bounded bnext |
    	
    	bounded := [ :p | (p x between: 1 and: input width) 	and: [ p y between: 1 and: input height ] ].
    	
    	direc := dir.
    	sum := 0.
    	4 timesRepeat: [ 
    		next := start + direc.
    		bnext := RSBox new size: 12; position: next*12; color: Color blue translucent.
    		canvas add: bnext.
    		"canvas signalUpdate."
    		"0.034 seconds wait."
    		sum := sum + ((bounded value: next) ifTrue: [
    			((input at: next) = 9 and: [(input at: start) = 8])
    				ifTrue: [ bnext color: Color green translucent. 1 ]
    				ifFalse: [ (input at: next) = ((input at: start) + 1)
    					ifTrue: [ | t | 
    						t :=  self d9graphvis: input start: next dir: direc canvas: canvas.
    						t = 0 ifTrue: [ self canvasRemove: { bnext } canvas: canvas ].
    						t ]
    					ifFalse: [ self canvasRemove: { bnext } canvas: canvas. 0 ] 
    				]
    			]
    		ifFalse: [ self canvasRemove: { bnext } canvas: canvas. 0 ]).
    		direc := direc rightRotated.
    		"canvas signalUpdate."
    		"0.034 seconds wait."
    	].
    
    	^ sum.
    
    • @mykl
      link
      15 days ago

      Great, thanks! I’ve always had Roassal at the back of my mind as a way of generating visualisations for AOC, but never got round to it. This might kick-start me, but maybe for next year at this point :-)