Cquant is node-addon-api based node extension, with the performance of C++, you can easily compute the major color of an image. The core algorithm of Cquant is based on leptonica.

CQuant-Web

• Supported Platform: Windows, Linux, macOS
• Node Supported Prebuild binary version: 6 | 8 | 10 | 11
• Electron Prebuild Supported Versions: v3 and v4.0.4

npm i cquant sharp // install cquant and sharp

This package is async. You can run multiple task without blocking the main loop.

const cquant = require('cquant')
// use sharp to convert image to RGB Buffer Array
const sharp = require('sharp')
sharp('path/to/image')
  .raw() // convert raw buffer like [RGB RGB RGB RGB]
  .toBuffer((_err, buffer, info) => {
    if (!_err) {
      let colorCount = 4

      cquant.paletteAsync(buffer, info.channels, colorCount).then(res => {

        console.log(res)
      }).catch(err => {

        console.log(err)
      })
    }
  })

  /**
   * 
   * @param buffer Image Buffer(RGB/RGBA)
   * @param depth 3 or 4 for RGB/RGBA
   * @param maxColor Color Amount You want
   * @param maxSub max sub-sample for image, 1 for no sub sample,0 for auto, by default it will scale to size of `1000x1000`
   * @param callback callback with err and result
   */
  function paletteAsync(buffer: Buffer, depth=3, maxColor=5, maxSub=0, callback:CallBackFunc): void;
  interface Color {
      R: number; /*red*/
      G: number; /*green*/
      B: number; /*blue*/
      count: number; /*count*/
  }
  declare type Palette = Color[];
  type CallBackFunc = (err, result: Palette) => void;
  function paletteAsync(buffer: Buffer, depth=3, maxColor=5, maxSub=0): Promise<Palette>;
  

test result will be diff based on your local machine

N/A: crashed

If you have lots of image to process, the best way to do it is using async.queue for parallel, and controllable

// test/example.js
const myQueue = async.queue(async (filePath) => {
  const img = await sharp(filePath)
    .raw() // to raw
    .toBuffer({ resolveWithObject: true })
  const palette = await cquant.paletteAsync(img.data, img.info.channels, 5)
  console.log(palette)
}, os.cpus().length - 1)

After running the install command make sure to use electron-rebuild to rebuild it for electron, usually it will just download the prebuild.

To be able to build from the source, you also need the standard build tool based on your OS.

• You can use this awesome app windows-build-tools to auto download all tools needed, if you don't have visual studio installed
• If you have already installed a visual studio like vs2017, basically you just need to enable c++ development(it's gonna be huge).

Basically you need run this command

npm run build

xVan Turing 2019