Jest Playwright

Running your tests using Jest & Playwright

npm install -D jest jest-playwright-preset playwright

Also you can use jest-playwright-preset with specific playwright packages: playwright-webkit, playwright-chromium and playwright-firefox

npm install -D jest jest-playwright-preset playwright-firefox

Requirements

• Node.js >= 10.15.0
• Playwright >=0.12.1

Usage

Update your Jest configuration, either:

with package.json:

"jest": {
  "preset": "jest-playwright-preset"
}

or with jest.config.js:

module.exports = {
  preset: 'jest-playwright-preset',
}

And add the Jest command as in the script section of your package.json:

{
  "scripts": {
    "test": "jest"
  }
}

Now you can use Playwright in your tests:

beforeAll(async () => {
  await page.goto('https://whatismybrowser.com/')
})

test('should display "google" text on page', async () => {
  const browser = await page.$eval('.string-major', (el) => el.innerHTML)
  expect(browser).toContain('Chrome')
})

Notes

It's recommend to use a separate Jest configuration jest.e2e.config.js for jest-playwright to gain speed improvments and by that to only use Playwright in the end-to-end tests. For that you have to use the -c flag when calling Jest and use the testMatch or testRegex in your Jest config to split them.

Be sure to remove any existing testEnvironment option from your Jest configuration. The jest-playwright-preset preset needs to manage that option itself.

Configuration

You can specify a jest-playwright.config.js at the root of the project or define a custom path using the JEST_PLAYWRIGHT_CONFIG environment variable. It should export a config object.

• launchOptions <[object]>. All Playwright launch options can be specified in config. Since it is JavaScript, you can use all stuff you need, including environment.
• launchType <LAUNCH | PERSISTENT | SERVER>. Method to launch browser instance. jest-playwright attaches Playwright to an existing browser instance by default.
• connectOptions <[object]>. All Playwright connect options can be specified in config.
• contextOptions <[object]>. All Playwright context options can be specified in config.
• browsers <[string[]]>. Define browsers to run tests in.
  • chromium Each test runs Chromium (default).
  • firefox Each test runs Firefox.
  • webkit Each test runs Webkit.
• devices <[(string | object)[] | RegExp]>. Define a devices to run tests in. Actual list of devices can be found here.
• exitOnPageError <[boolean]>. Exits process on any page error. Defaults to true.
• collectCoverage <[boolean]>. Enables the coverage collection of the saveCoverage(page) calls to the .nyc_output/coverage.json file.
• serverOptions <[object]>. All jest-process-manager options.
• selectors <[array]>. Define selectors. Each selector must be an object with name and script properties.

Device configuration

There are different ways to define browsers in your tests:

• You can use array of device names:

module.exports = {
  devices: ["iPhone 6", "Pixel 2"],
  ...
}

• You can use RegExp:

module.exports = {
  devices: /iPhone 8/,
  ...
}

• Also you can define custom device:

{
  // Name of device
  name: string
  // Page width and height
  viewport: {
    width: number
    height: number
  }
  // user agent
  userAgent: string
  // device scale factor
  deviceScaleFactor: number
  // is device is mobile
  isMobile: boolean
  // support of touch events
  hasTouch: boolean
}

Usage with query-selector-shadow-dom in jest-playwright.config.js:

const {
  selectorEngine,
} = require('query-selector-shadow-dom/plugins/playwright');

module.exports = {
  selectors: [
    {name: 'shadow', script: selectorEngine}
  ],
  ...
}

Notes

You can also specify browser with the BROWSER environment variable. You should do it only if you are using the whole playwright package. You can specify device with DEVICE environment variable.

Globals

• browserName <[string]> - name of the current browser (chromium, firefox or webkit)
• deviceName <[string]> - name of the current device
• browser <[Browser]> - Playwright browser instance
• context <[Context]> - a new Playwright context instance for each new test file
• page <[Page]> - Playwright page instance (since a new context for every test file also a new page for it)

All of them are available globally in each Jest test. If you are using ESLint and JavaScript, its recommend to use it in combination with the eslint-plugin-jest-playwright.

Put in debug mode

Debugging tests can be hard sometimes and it is very useful to be able to pause tests in order to inspect the browser. Jest Playwright exposes a method jestPlaywright.debug() that suspends test execution and gives you opportunity to see what's going on in the browser.

await jestPlaywright.debug()

Reset helper functions

Reset current page

beforeEach(async () => {
  await jestPlaywright.resetPage()
})

To create a new page for each test, you can use this snippet to have a new page object for each individual test.

Reset current context

beforeEach(async () => {
  await jestPlaywright.resetContext()
})

To create a new context for each test, you can use this snippet to have a new context object for each individual test.

Reset current browser

beforeEach(async () => {
  await jestPlaywright.resetBrowser()
})

You can use this snippet to reset current browser for each individual test. It will reset browser, context and page.

Tracking the coverage

It's possible to track the coverage of the end-to-end tests with the babel-plugin-istanbul Babel plugin configured. It needs to be included in the web application which you are gonna test otherwise it won't work. To use it, you have to set collectCoverage in the jest-playwright.config.js to true. Per default the test coverage will be automatically saved after each navigation change (beforeunload event). If a certain code path is not covered, you can manually call and add the corresponding saveCoverage(page) call to your tests like that:

await jestPlaywright.saveCoverage(page)

By using coverage collection, it will write the coverage data to the .nyc_output/coverage.json file which can be transformed using nyc to the lcov format:

npx nyc report --reporter=lcovonly

or to HTML:

npx nyc report --reporter=html

which will create a HTML website in the coverage directory.

Skip tests for specific browsers and devices

It's possible to skip tests for browsers or combination of browsers and devices

it.jestPlaywrightSkip(
  { browsers: ['chromium'] },
  'should skip this one',
  async () => {
    const title = await page.title()
    expect(title).toBe('Google')
  },
)

Start a server

Jest Playwright integrates a functionality to start a server when running your test suite, like jest-puppeteer. It automatically closes the server when tests are done.

To use it, specify a server section in your jest-playwright.config.js.

// jest-playwright.config.js
module.exports = {
  serverOptions: {
    command: 'node server.js',
    port: 4444,
  },
}

Other options are documented in jest-process-manager.

Using with different jest environments

The default jest-playwright environment is node, but you can use a browser-like environment through jest-playwright-jsdom

expect-playwright

There is a utility package expect-playwright which simplifies the expect statements in combination with Playwright to make e.g. shorter text comparisons.

ESLint globals / 'page' is not defined

There is an ESLint plugin available eslint-plugin-jest-playwright available which includes the globals for using jest-playwright.

Unstable and experimental API

You can run tests for multiple browsers and devices:

• You must have installed the playwright package
• You must define browsers to test with your jest-playwright.config.js:

module.exports = {
    browsers: ["chromium", "webkit"],
    devices: ["iPhone 6", "Pixel 2"],
    ...
}

It will run your tests for:

• Chromium browser and iPhone 6 device;
• Chromium browser and Pixel 2 device;
• Webkit browser and iPhone 6 device;
• Webkit browser and Pixel 2 device;

If there is no defined browsers in config it will run tests for chromium browser.

Usage with jest-circus

You can use jest-playwright with the jest-circus runner for taking screenshots during test failures for example:

jest.config.json

"testRunner": "jest-circus/runner",
"testEnvironment": "./CustomEnvironment.js"

CustomEnvironment.js

const PlaywrightEnvironment = require('jest-playwright-preset/lib/PlaywrightEnvironment')
  .default

class CustomEnvironment extends PlaywrightEnvironment {
  async setup() {
    await super.setup()
    // Your setup
  }

  async teardown() {
    // Your teardown
    await super.teardown()
  }

  async handleTestEvent(event) {
    if (event.name === 'test_done' && event.test.errors.length > 0) {
      const parentName = event.test.parent.name.replace(/\W/g, '-')
      const specName = event.test.name.replace(/\W/g, '-')

      await this.global.page.screenshot({
        path: `screenshots/${parentName}_${specName}.png`,
      })
    }
  }
}

module.exports = CustomEnvironment

Usage with Typescript

Example Jest configuration in combination with ts-jest:

module.exports = {
  preset: 'jest-playwright-preset',
  transform: {
    '^.+\\.ts$': 'ts-jest',
  },
}

Types are also available, which you can either use via directly in your test:

/// <reference types="jest-playwright-preset" />
/// <reference types="expect-playwright" />

or at your central tsconfig.json either via files:

{
  "files": [
    "./global.d.ts",
    "node_modules/jest-playwright-preset/types/global.d.ts",
    "node_modules/expect-playwright/global.d.ts"
  ]
}

or via types:

{
  "compilerOptions": {
    "types": ["jest-playwright-preset", "expect-playwright"]
  }
}

It's important to not change the testEnvironment to node. Otherwise it won't work.

Known issues

Error reporting with Jest

If you face into error messages like UnhandledPromiseRejectionWarning: Error: Protocol error (Runtime.callFunctionOn): Target closed. or

Timeout - Async callback was not invoked within the 20000ms timeout specified by jest.setTimeout.Timeout - Async callback was not invoked within the 20000ms timeout specified by jest.setTimeout.Error:

and your Jest error reporting will only show that an entire test (it() function) has failed, then you need to increase the Jest timeout because the Playwright timeout is greater than the Jest timeout. So Jest in the end will simply stop the execution and no verbose (which exact line) error reporting can be generated.

To fix this behavior simply call

jest.setTimeout(35 * 1000)

in your tests at the top. (30 seconds is the default Playwright timeout for waiting for an specific element.)

New Browser instance for each test

If for your individual tests a new entire browser instance spins up each time and it won't be reused, then you probably run them in parallel. If you run them in a synchronous way with the --runInBand CLI option for Jest, then the same browser instance will be re-used and this should fix the issue.

Examples

Demonstration the usage of jest-playwright for various test cases can be found in playwright-jest-examples

Inspiration

Thanks to Smooth Code for the great jest-puppeteer.

License

MIT