easily mock a web application accessed through selenium. Contains also an environment for integrating with jest.
yarn add -D @4c/selenium-sandbox
First, we need to instrument our application to use the sandbox environment. Create a module (let's call it injectSandbox.js
)
import sandbox from '@4c/selenium-sandbox/browser';
const store = {
widgets: [],
};
// additional properties and methods that can be accessed from selenium. for example here we are passing a store object
const context = { store };
sandbox.setupTestContext(context);
sandbox.fetchMock.mock('path:/api/v1/widgets', () => store.widgets);
Note: documentation for
fetchMock
can be found here
Add utilities to an already existing selenium driver (v4):
import { augmentDriver } from '@4c/selenium-sandbox/webdriver';
const driver = augmentDriver(
myBaseSeleniumDriver,
'http://sandboxed-app-to-test',
);
... Alternatively you can use buildDriver to instantiate a reasonably opinionated driver:
import { buildDriver } from '@4c/selenium-sandbox/webdriver';
const driver = buildDriver(
baseUrl: 'http://sandboxed-app-to-test';
seleniumAddress: 'localhost:5000/wd/hub';
browserName: 'chrome';
screenSize: [1024, 768];
// optional
mobileEmulation: {
deviceName: 'iPhone 6/7/8',
};
);
To mock:
await driver.get('my/page');
await driver.executeInBrowser(context => {
context.store.widgets = [{label: '1', label: '2'}]
}
const button = await driver.find('//button[text()="Get Widgets"]');
await driver.click(button);
If a request needs to be mocked before page loading, you can use executeAtStartup
instead of executeInBrowser
. the code will be stored in storage and executed when the assets are loaded
Note: Since the function passed to
executeInBrowser
orexecuteAtStartup
needs to be stringified in order to be executed in the remote browser, it cannot access variables defined outside of its body.
You can also access fetchMock from the context to make additional mocks:
await driver.executeInBrowser(context => {
context.fetchMock.mock('path:/api/v1/users', () => ['user1', 'user2'])
}
Since fetchMock
is part of the context, it can be used to make assertions on the requests. There are two helper methods on the driver to facilitate that:
const request = await driver.getLastRequest();
invariant(request.headers.Authorization == 'Bearer XXX');
const requests = await driver.getRequests();
invariant(requests.every((req) => req.headers.Authorization == 'Bearer XXX'));
the driver has also the following utilities methods:
// more resilient than element.click, will retry several times to avoid flakiness
await driver.click(element);
// accepts an xpath query and returns the first match. It will wait that all
// images are loaded and that the element is indeed visible
await driver.find('//button[text()="Get Widgets"]');
// unlike the base webdriver, allows to pass a path which is concatenated
// to `baseUrl`
await driver.get('my/page');
// the name says it all. Useful to make sure the page has finished rendering
await driver.waitForAllImages();
There is also support for easy integration with Jest.
add the following line to your jest config:
{
// ...
testEnvironment: '@4c/selenium-sandbox/jest/environment.js',
setupFilesAfterEnv: ['@4c/selenium-sandbox/jest/setup.js'],
testEnvironmentOptions: {
baseUrl: 'http://sandboxed-app-to-test',
seleniumAddress: 'localhost:5000/wd/hub',
browserName: 'chrome',
screenSize: [1024, 768],
// optional
mobileEmulation: {
deviceName: 'iPhone 6/7/8',
},
// optional, default 10000
waitTimeout: 50000,
},
}
this will:
- declare a global
browser
variable as described above - set a 10s timeout which is a better fit for selenium tests
- add a snapshot serializer specific to fetchMock requests
Type definitions come out of the box. For full jest support, you should add the following declaration to be used in your test file:
import { AugmentedDriver } from '@4c/selenium-sandbox/webdriver';
// define a context type that reflects the context passed in `setupTestContext`
type Context = {
store: {
widgets: {}[];
};
};
declare const browser: AugmentedDriver<Context>;