npm

Bring the best of OSS JavaScript development to your projects with npm Orgs - private packages & team management tools.Learn more ¬Ľ

fuzi

1.2.0-alpha¬†‚Äʬ†Public¬†‚Äʬ†Published

ūüöß Alpha Version (Work in Progress)

API subject to change


Fuzi!

ūüźĽ fuzzy image diff'ing for the terminal

Build Status Coverage Status NPM Version XO code style

Support

Support the development of Fuzi by becoming a patreon.

Example

Need to check the difference between two images? Ask Fuzi! ūüźĽ

$ fuzi red-square.png green-circle.jpg -v

Fuzi! in action

Features

  • Diff images from the CLI
  • Diff in API Mode for use in test frameworks
  • Images do not have to match
    • Size can be different
    • Dimensions can be different
    • File type can be different
  • Supports file-types: PNG, JPG
  • Configurable settings for tweaking each tests

About

Fuzi is an image diffing tool for Ref-Testing in a CI environment, (Travis, Circle, etc). Fuzi outputs graphics to your terminal, making it quicker to understand why a test is failing.

CLI plus API

Fuzi can diff in the CLI so that you can experiment and get used to it's inputs and outputs. Fuzi diff as a Node API to use in your testing framework.

Call it a "PASS"

When your tests fail in the cloud, Fuzi gives you the tolerance settings required to make the test pass the next time it runs. Phew!

Test details in the Terminal

Installation

For global use:

$ npm i -g fuzi
$ yarn --global add fuzi

For local testing:

$ npm i --save-dev fuzi
$ yarn add fuzi

How does Fuzi work?

Grid

Rather than try to compare the whole image, Fuzi splits each image into a spatial grid and compares the two. This helps to evaluated which parts of the image match, and which are different.

Channels

Fuzi splits your images into 4 channels to finds the average value for each grid square. These chanels are:

  • Hue
  • Saturation
  • Luminance
  • Alpha

Tolerance

Fuzi then takes the tolerance parameter you provide, and compares the average chanel values in each square to see if they are above or below your threshold.

The thresholds default to 0:

const opts = {
    tolerance: {
        hue: 0,
        sat: 0,
        lum: 0,
        alp: 0
    }
}

PASS or FAIL

If any grid square's average channel value is above the tolerance (for any channel), your images are considered to be different.

This results in a FAIL.

If all grid squares average channel values are below your provided tolerance threshold, then your images are considered to the alike.

Using Fuzi with your Test Framework

import test from 'ava'
import fuzi from 'fuzi'
 
test('Different images fail', async t => {
    const img1 = 'fixtures/green-circle.png'
    const img2 = 'fixtures/red-square.png'
 
    const opts = {
        grid: {
            columns: 8,
            rows: 8
        },
        tolerance: {
            hue: 0,
            sat: 0,
            lum: 0,
            alp: 0
        }
    }
 
    const result = await fuzi(img1, img2, opts)
    t.true(result.fail)
})
import test from 'ava'
import fuzi from 'fuzi'
 
test('Identical images pass', async t => {
    const img1 = 'fixtures/green-circle.png'
    const img2 = 'fixtures/green-circle.png'
    const opts = {
        grid: {
            columns: 8,
            rows: 8
        },
        tolerance: {
            hue: 0,
            sat: 0,
            lum: 0,
            alp: 0
        }
    }
    const result = await fuzi(img1, img2, opts)
    t.true(result.pass)
})

install

npm i fuzi

Downloadsweekly downloads

9

version

1.2.0-alpha

license

MIT

homepage

github.com

repository

Gitgithub

last publish

collaborators

  • avatar
Report a vulnerability