Codeceptjs Visual Testing

codeceptjs-visual-testing is a plugin (helper) for CodeceptJS E2E tests that helps verify that your UI doesn't unexpectedly change.

It makes use of pixelmatch and pngjs in order to avoid the ridiculous dependencies for ImageMagick and applies some basic heuristics to help prevent false positives.

Installation

1. Run

npm install codeceptjs-visual-testing

2. Add the following to your codeceptjs.conf.js file.

exports.config = {
  helpers: {
    VisualTesting: {
      require: 'codeceptjs-visual-testing',

      // This is where baseline screenshots will be stored alongside any
      // extra json data used by this helper.
      baseFolder: '../visual/',
      
      // If your test fails, an image will be generated in this directory
      // that highlights the differences between your app's current state
      // and the baseline image.
      diffFolder: '../output/visual/'
    }
  }
}

Usage

Example

Scenario('the dashboard page looks the same', async () => {
   I..amOnPage('/');

   await I.dontSeeVisualChanges('dashboard_basic', {
     // Optionally, substitute matched text in new screenshots using text from the baseline image (Default: []).
     preserveTexts: ['.date:last-of-type', '#random-string'],
     // Optionally, hide elements that can't be consistently reproduced (Default: []).
     hideElements: ['.some-randmom-popup'],
     // Optionally, allow for more variation in the screenshot without failing the test (Default: 1).
     allowedMismatchedPixelsPercent: 3
   })
})

Creating and Updating Baseline Images

1. Run UPDATE_VISUALS=1 codeceptjs run
2. Commit the files generated in your baseFolder

Disclaimer

Visual tests should generally be used sparingly and on pages unlikely to change. Altering the page style or layout will mean needing to update the corresponding visual tests. Similarly, unless you're using git lfs, the binary image files may eventually take up a lot of space in your Git repo history.

Testing

Run npm run test:docker to run this project's tests.

This project comes with a simple test app and is containerized with Docker for easier cross-platform development and automated CI tests.