wavefunctioncollapse

Javascript port of https://github.com/mxgmn/WaveFunctionCollapse

• Interactive OverlappingModel demo
• Interactive SimpleTiledModel demo
• Animated SimpleTiledModel demo

Installing

With npm do:

npm install wavefunctioncollapse --production

Public API

OverlappingModel Constructor

new OverlappingModel(data, dataWidth, dataHeight, N, width, height, periodicInput, periodicOutput, symmetry[, ground])

• data : The RGBA data of the source image.
• dataWidth : The width of the source image.
• dataHeight : The height of the source image.
• N : Size of the patterns.
• width : The width of the generation (in pixels).
• height : The height of the generation (in pixels).
• periodicInput : Whether the source image is to be considered as periodic / as a repeatable texture.
• periodicOutput : Whether the generation should be periodic / a repeatable texture.
• symmetry : Allowed symmetries from 1 (no symmetry) to 8 (all mirrored / rotated variations)
• ground : Id of the specific pattern to use as the bottom of the generation (learn more)

var wfc = require('wavefunctioncollapse');

var imgData = ... // let's pretend this is an ImageData retrieved from a canvas context in the browser

var model = new wfc.OverlappingModel(imgData.data, imgData.width, imgData.height, 3, 48, 48, true, true, 4);

OverlappingModel Methods

model.graphics([array])

Retrieve the RGBA data of the generation.

• array : Array to write the RGBA data into (must already be set to the correct size), if not set a new Uint8Array will be created and returned. It is recommended to use Uint8Array or Uint8ClampedArray.

// create a blank ImageData
var imgData = canvasContext.createImageData(48, 48);

// write the RGBA data directly in the ImageData
model.graphics(imgData.data);

// print the ImageData in the canvas
canvasContext.putImageData(imgData, 0, 0);

SimpleTiledModel Constructor

new SimpleTiledModel(data, subsetName, width, height, periodicOutput)

• data : Tiles, subset and constraints definitions. The proper doc on this matter is yet to be written, check the example in the meantime.
• subsbetName : Name of the subset to use from the data. If falsy, use all tiles.
• width : The width of the generation (in tiles).
• height : The height of the generation (in tiles).
• periodicOutput : Whether the generation should be periodic / a repeatable texture.

var wfc = require('wavefunctioncollapse');

var data = ... // object with tiles, subsets and constraints definitions

var model = new wfc.SimpleTiledModel(data, null, 48, 48, false);

SimpleTiledModel Methods

model.graphics([array[, defaultColor]])

Retrieve the RGBA data of the generation.

• array : Array to write the RGBA data into (must already be set to the correct size), if not set a new Uint8Array will be created and returned. It is recommended to use Uint8Array or Uint8ClampedArray.
• defaultColor : RGBA data of the default color to use on untouched tiles.

// create a blank ImageData
var imgData = canvasContext.createImageData(48, 48);

// write the RGBA data directly in the ImageData, use an opaque blue as the default color
model.graphics(imgData.data, [0, 0, 255, 255]);

// print the ImageData in the canvas
canvasContext.putImageData(imgData, 0, 0);

Common Methods

model.generate([rng])

Execute a complete new generation. Returns whether the generation was successful.

model.generate(Math.random); // return true or false

• rng : A function to use as random number generator, defaults to Math.random.

model.iterate(iterations[, rng])

Execute a fixed number of iterations. Stop when the generation is successful or reaches a contradiction. Returns whether the iterations ran without reaching a contradiction.

• iterations : Maximum number of iterations to execute (0 = infinite).
• rng : A function to use as random number generator, defaults to Math.random.

model.iterate(30, Math.random); // return true or false

model.isGenerationComplete()

Returns whether the previous generation completed successfully.

model.clear()

Clear the internal state to start a new generation.

Changelog

2.0.0 (2019-07-06)

• Port of the newer, faster, implementation.
• This port is now written in ES6 instead of ES5 [breaking change].

1.0.0 (2016-10-14)

• Change and freeze the public API (with iteration support).
• First publication on NPM.

License

MIT