npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

@guardian/ab-core

v8.0.0

Published

A client-side library for A/B & multivariate testing

Downloads

5,045

Readme

The Guardian's AB Testing Library

Getting Started

  1. Install the package with your manager of choice. e.g. pnpm add @guardian/ab-core
  2. Initialise the AB Class in your project
  3. Consume the API

Note There is also a (P)React framework-specific implementation, @guardian/ab-react

What does this client-side A/B testing library do?

This library:

  • Is initialised with an multivariate identifier (MVT ID)
    • This can be set by the server for fast user bucketing, e.g. as a cookie
  • Is built upon the A/B testing code from frontend
  • Has the ability to force variants for testing
  • Has simple integration with Ophan, with impression and success methods built in to the library
  • Can be integrated into a Typescript or Javascript project

There’s some background to the early requirements to the library and some documentation in Frontend, there’s some notes about the migration of A/B tests.

How it works

  1. Define the AB test: Each AB test and their variants are defined in code with configuration such as audience size & offset and impression & success listeners etc
  2. Initialise the library: The AB Test library is initialised with configuration values such as a user's MVT ID, an array of the above defined A/B tests etc
  3. Use the AB Test API: The intialisation returns an API that can be used to check if the current user is in a variant of a test along with a variety of other API methods

Initialising

import { AB } from '@guardian/ab-core';
import type { AbTestConfig } from '@guardian/ab-core';

const config: AbTestConfig = {
	mvtId: 999_999,
	pageIsSensitive: false,
	abTestSwitches: {},
	arrayOfTestObjects: [],
};

const abTests = new AB(config);

// Must be performed in the platform after initialisation to ensure tracking defined in ABTests is setup
// [tests] being an array of *runnable* ab tests
abTest.registerCompleteEvents([tests]);
abTest.registerImpressionEvents([tests]);
abTest.trackABTests([tests]);

// The API then provides access to the utility methods for use within modules
abTests.runnableTest(test);
abTests.firstRunnableTest([tests]);
abTests.isUserInVariant(testId, variantId);

The initialisation config object

| Config | Type | Example | | ------------------- | ------------------------------------------------- | --------------------------------------------------------------------------------- | | mvtMaxValue | number | 10_000 | | mvtCookieId | number | getCookie('mvt_id') | | pageIsSensitive | boolean | guardian.config.page.isSensitive | | abTestSwitches | Record | {'TestOne': true} | | forcedTestVariants | Optional: { [key: string]: { variant: string } }; | { TestOne: { variant: 'myCoolVariant' }, TestTwo: { variant: 'myCoolVariant' } }; | | forcedTestException | Optional: ABTest['id'] | | | arrayOfTestObjects | ABTest[] | | | ServerSideTets | ServerSideTests | | | errorReporter | ErrorReporterFunc | | | ophanRecord | OphanRecordFunction | |

The API

type CoreAPI = {
	allRunnableTests: (
		tests: ReadonlyArray<ABTest>,
	) => ReadonlyArray<Runnable<ABTest>> | [];
	runnableTest: (
		test: ABTest,
	) => Runnable<ABTest & { variantToRun: Variant }> | null;
	firstRunnableTest: (tests: ReadonlyArray<ABTest>) => Runnable<ABTest> | null;
	isUserInVariant: (testId: ABTest['id'], variantId?: Variant['id']) => boolean;
};

| API Method | Params | Returns | | ----------------- | ------------------------------ | ------------------------------------------------------------------------------------------ | | allRunnableTests | Array of ab tests | Array of Runnable tests or empty array | | runnableTest | A single AB test | A runnable ab test object with variantToRun property containing the variant to run or null | | firstRunnableTest | Array of AB Tests | A runnable ab test object or null | | isUserInVariant | A AB test ID, and a Variant ID | A boolean |

Ab Test Definition

Within your platforms, you should define the test this way. If the test needs to run across platforms, then the test definition needs to be the same (as well as the initialisation config).

import { ABTest } from '@guardian/ab-core';

export const abTestTest: ABTest = {
    id: 'AbTestTest', // This ID must match the Server Side AB Test
    start: '2020-05-20',
    expiry: '2020-12-01', // Remember that the server side test expiry can be different
    author: '[email protected]',
    description: 'This Test'
    audience: 0.0001, // 0.01% (1 is 100%)
    audienceOffset: 0.5, // 50% (1 is 100%). Prevent overlapping with other tests.
    successMeasure: 'It works',
    audienceCriteria: 'Everyone',
    idealOutcome: 'It works',
    showForSensitive: true, // Should this A/B test run on sensitive articles?
    canRun: () => true, // Check for things like user or page sections
    variants: [
        {
            id: 'control',
            test: (): void => {}, // You can define what you want your variant to do in here or use the isUserInVariant API
            impression: (impression) => {
                // This will be immediate.
                // You could also use eventListeners as below
                // Make sure abTest.registerCompleteEvents([tests]); and abTest.registerImpressionEvents([tests]); have been called
                impression();
            },
            success: (success) => {
                // https://developer.mozilla.org/en-US/docs/Web/Guide/Events/Creating_and_triggering_events
                window.addEventListener('guAbTestEvent', (e) => {
                    const detail = e?.detail;
                    if (
                        detail?.abTest === 'abTestTest' &&
                        detail?.variant === 'control' &&
                        detail?.event === 'success') {
                            success();
                        }
                });

            },
        },
        {
            id: 'variant',
            test: (): void => {},
            impression: (impression) => {
                impression();
            },
            success: (success) => {
                //...
            },
        },
    ],
};

// If you're using event listeners for the impression and success events you can call them with CustomEvents
// Say a user clicked something
window.dispatchEvent(new CustomEvent('guAbTestEvent', {
  detail: {
    abTest: 'abTestTest',
    variant: 'control'
    event: 'success'
  }
}))

Example of the AB Test config in Frontend and DCR

The initialisation values are populated on these platforms like so:

| Config | Note | | ------------------- | --------------------------------------------------------------------------------------------------------------------------------- | | mvtMaxValue | MVT % is calculated from 0 to mvtMaxValue | | mvtCookieId | The user's MVT ID to calculate what tests and variants they fall into | | pageIsSensitive | Sensitive pages must have explicit settings in AB tests | | abTestSwitches | An object containing all of the boolean values of abTestSwitches, in Frontend from page.config.switches.abTests | | forcedTestVariants | In Frontend this might be set by the URL override, but otherwise can be used to force a user into a test and variant at init time | | forcedTestException | Can be used to force a user out of a test (in Frontend, again with url override) | | arrayOfTestObjects | Pass all tests definitions into the config | | ServerSideTets | ServerSideTests are accessed via client-side config in Frontend and DCR | | errorReporter | Pass an error reporter, probably Sentry | | ophanRecord | Probably Ophan's 'record' function |

Frontend: Difference and Integration with DCR

Integration between Frontend and DCR

  • There is currently a requirement to copy and paste the AB test definitions between the two platforms. Each platform has an experiments folder (Frontend, DCR) and the test definition and structure of those folders should match. The difference will be where to import - in Frontend in ab-test.ts and in DCR in ab-tests.ts.
  • For Frontend and DCR, you will need to have a switch as you would usually do in Frontend. This will be passed through to DCR in the backend and be accessible to the client-side code.

Differences of this library vs Frontend implementation

  • There is no local storage functionality in this library, unlike previously where you could store the A/B test in Frontend. To persist an opted-in test, the MVT cookie will need to be set to the correct value.
  • Forced Tests - used by the url-opt-in mechanism ignores canRun on both the test and the variant, so it will always run when forced. Previously on Frontend it still listened to the variant's canRun.
  • Handles only concurrent tests, this library does not concern itself with epics or banner tests. There is no mechanism for A/B tests to interact or wait for one another outside of audience size and offsets.
  • The public API is reduced to only what was used in Frontend
  • Some public methods have been renamed like isUserInVariant (this does make it difficult to copy and paste an implementation between Frontend and DCR right now until this library is integrated with Frontend)

MVTId calculator

Use this simple calculator to see what MVT ID your test variant will fall into.

Testing with Jest

Uses Jest, see .test.ts files.

What's Next

  • [x] Integrate into Frontend
  • [ ] Review usability across other platforms and required APIs
  • [ ] Investigate tree-shakeable-ness of methods
  • [ ] Investigate exposing API methods outside of the configuration, to allow usage inside of modules without passing a prop