@selfage/puppeteer_test_executor

Execute bundled JS file with extra global functions exposed by Puppeteer

Usage no npm install needed!

<script type="module">
  import selfagePuppeteerTestExecutor from 'https://cdn.skypack.dev/@selfage/puppeteer_test_executor';
</script>

README

@selfage/puppeteer_test_executor

Install

npm install @selfage/puppeteer_test_executor

Overview

Written in TypeScript and compiled to ES6 with inline source map & source. See @selfage/tsconfig for full compiler options. Provides an executor which launches a single browser page via Puppeteer, exposes helper functions for testing to browser context/window scope, and executes a single/bundled JS file in that page.

Puppeteer is a Node library which provides a high-level API to control Chrome or Chromium over the DevTools Protocol.

Executing a JS file

You can use the CLI e.g. $ npx pexe file/to/be/run/in/browser.js -b base/dir. See below for detailed options.

$ npx pexe -h
Usage: pexe [options] <binFile> [pass-through-args...]

Execute the presumably bundled JavaScript file in browser context. The file ext can be neglected and is always fixed as .js. "--" is needed in between <binFile> and pass through arguments.

Options:
  -b, --base-dir <baseDir>  The base directory that all loaded files/resources should be relative to, from which all static files are being served in the server. If not
                            provided, it will be the current working directory.
  -p, --port <port>         The port number to start your local server. Default to 8000.
  -l, --outputToConsole     Whether to log logs in browser to console.
  -h, --help                display help for command

Or use programmatical API as below.

import { execute, OutputCollection } from '@selfage/puppeteer_test_executor';

// Runs in Node context.
let outputCollection: OutputCollection = execute(
  'file/to/be/run/in/browser.js',
  'base/dir',
  /* outputToConsole= */ true,
  /* port= */ 8080,
  /* argv= */ ['--case', 'AssertAddition']);
outputCollection.log; // Array<string>
outputCollection.warn; // Array<string>
outputCollection.error; // Array<string>
outputCollection.other; // Array<string>

As a prerequisite, you need a JS file that's meant to be run in browser context, which contains everything needed to render a page/perform whatever actions. It can be a bundled file using bundlers such as browserify, or a bootstrap file that loads other files.

The execute() function or the CLI will start a local Node server, embed the JS file into a temporary HTML file, and launch Puppeteer to navigate to that HTML file. When outputToConsole is true, it logs browser logs to Node console. Returned outputCollection will always collect all logs from the browser. pass-through-args or argv argument are made available to the JS file as a global variable.

The full JS file path in the example is ./base/dir/file/to/be/run/in/browser.js which is split into two parts: a base directory and a relative path. You might just want to use '.', which is the default, as the base directory. However the base directory is where all static files are being served from. E.g. if the JS file loads an image with src='/path/to/image.jpg', the server will try to find it at ./base/dir/path/to/image.jpg.

Note that, normally a browser page won't exit by itself even if everything have been executed, so execute() or the CLI will never resolve/return. The solution is to call exit() from the JS file.

JS file in browser context

The JS file executed by the way above has access to several more global/window variables & functions. Below is the full list copied from apis.ts which can be imported after installing @selfage/puppeteer_test_executor_api.

declare var argv: Array<string>;
declare function exit(): void;
declare function screenshot(
  relativePath: string,
  options?: { delay?: number; fullPage?: boolean; quality?: number }
): Promise<void>;
declare function fileExists(relativePath: string): Promise<boolean>;
declare function readFile(
  relativePath: string,
  // An encoding is mandatory because raw buffer data cannot be transmitted
  // correctly. Refer to BufferEncoding for all supported strings.
  encoding: string
): Promise<string>;
declare function writeFile(
  relativePath: string,
  // Represented as a binary string.
  data: string
): Promise<void>;
declare function deleteFile(relativePath: string): Promise<void>;
declare function setViewport(width: number, height: number): Promise<void>;

Argv

import '@selfage/puppeteer_test_executor_api'; // Import type definitions only.

// Runs in browser context.
parseArg(argv); // ['--case', 'AssertAddition']
// or parseArg(globalThis.argv);

If you have executed the JS file with an argv argument, the value, which is of Array<string>, can then be accessed in the JS file by argv, i.e. it's a global/window scope variable. It's intended to be used just as command line arguments as if it's an exetuable JS file used in Node CLIs.

Functions

import '@selfage/puppeteer_test_executor_api'; // Import type definitions only.

// Runs in browser context.
async function main(): Promise<void> {
  await fileExists('some/file.txt');
  exit();
}

Functions made available this way are more powerful than regular functions because they are actually run in Node context, as opposed to browser context, thanks to the magic exposeFunction(). I.e. you can interact with file systems within browser context directly, and even control browser behaviors.

exit() is a special and important function. It tells the browser page which runs the JS file to close itself. Normally a browser page won't close/exit at all even if everything has been executed, because it's waiting for user to interact with the page. But in testing scenarios, the JS file itself often controls all interactions and knows when to end tests. exit() can then be called upon all tests finished, which in turn resolves execute() function. You can see @selfage/test_runner as an example.