featureflow-javascript-sdk

Featureflow JavaScript/TypeScript Client SDK for Browser Environments

This 2.x.x Client is for modern browser environments, for legacy support, please use 1.3.18

Note: This SDK is designed for JavaScript in browser environments. For server-side JavaScript/Node.js applications, please use the featureflow-node-sdk.

Get your Featureflow account at featureflow.io

Get Started

The easiest way to get started is to follow the Featureflow quick start guides

Examples

To see featureflow running in action, you can run the example in this repo:

1. Clone this repository
2. Update example/src/index.js const FF_KEY = 'your-javascript-environment-sdk-key';
3. Run $ npm install and $ npm run example

or See the example at Codepen.

Change Log

Please see CHANGELOG.

Installation

Using NPM

npm install --save featureflow-client

Usage

Adding Featureflow

The SDK supports multiple integration patterns:

Using CommonJS (for bundlers that support CommonJS syntax)

const Featureflow = require('featureflow-client');

Using ES2015 modules (modern JavaScript environments and most bundlers)

import Featureflow from 'featureflow-client';

Using TypeScript (same as ES modules, with type support)

import { init, events } from 'featureflow-client';
import type { FeatureflowClient, FeatureflowUser, Config, EvaluatedFeatures } from 'featureflow-client';

Webpack

If you are using webpack, you can require the code using

var Featureflow = require('featureflow-client');

or using es6 syntax

import Featureflow from 'featureflow-client';

CDN

Include the following script in HTML file. This will expose the global variable Featureflow

<script crossorigin="anonymous" src="https://cdn.featureflow.io/featureflow.js"></script>

Quick start

Get your environment's Featureflow Javascript API key and initialise a new Featureflow client

// Using init() - Promise-based initialization
var FF_JS_API_KEY = '<Your javascript api key goes here>';
const featureflow = await Featureflow.init(FF_JS_API_KEY);

TypeScript users can import the type for better autocomplete and type safety:

import { init, type FeatureflowClient } from 'featureflow-client';

const featureflow: FeatureflowClient = await init(FF_JS_API_KEY);

This will load the value of each feature for the current environment specified by the api key. These values can be toggled on and off at https://<your-org-key>.featureflow.io

Note

You are responsible for passing the featureflow client instance around your application

In your code, you can test the value of your feature where the value of my-feature-key is equal to 'on'

if (featureflow.evaluate('my-feature-key').is('on')){
  // this feature code will be run because 'my-feature-key' is set to 'on'
}

Because the default variants for any feature are 'on' and 'off', we have provided two helper methods .isOn() and .isOff()

if(featureflow.evaluate('my-feature-key').isOn()){
  // this feature code will be run because 'my-feature-key' is set to 'on'
}

if(featureflow.evaluate('my-feature-key').isOff()){
  // this feature code won't be run because 'my-feature-key' is not set to 'off'
}

You can include user context information when you initialise featureflow, these attributes can be used in feature targeting rules:

// User with id (required) and attributes (optional)
var user = {
  id: 'user123',
  attributes: {
    tier: 'gold',
    country: 'australia',
    roles: ['role1', 'role3']
  }
};
const featureflow = await Featureflow.init(FF_KEY, user);

// Or without user - anonymous user will be auto-generated
const featureflow = await Featureflow.init(FF_KEY);

TypeScript example:

import { init, type FeatureflowClient, type FeatureflowUser } from 'featureflow-client';

const user: FeatureflowUser = {
  id: 'user123',
  attributes: {
    tier: 'gold',
    country: 'australia',
    roles: ['role1', 'role3']
  }
};

const featureflow: FeatureflowClient = await init(FF_KEY, user);

Additional configuration can be set during init also. You can set offline mode for test environments or local development and provide a default set of feature values, for example:

const featureflow = await Featureflow.init(FF_KEY, user, {
    offline: true,
    defaultFeatures: {
      'feature-1': 'on',
      'feature-2': 'red',
      'test': 'off'
    },
});

Further documentation can be found in the Featureflow documentation

Anonymous User and Featureflow Server SDKs

In some instances you will want to split test a feature with anonymous users that spans both client and server code. The featureflow client generates a unique anonymous key for the user which can be shared with your server requests. There are a couple of ways you can do this.

1. Access the cookie ff-anonymous-id

Anytime the anonymous user is updated the featureflow client will set the ff-anonymous-id cookie. If request server is on the same [sub]domain and it doesn't have a user id you should use this cookie.

2. Pass the value into the request yourself

If you cannot use cookies (e.g. api on a separate [sub]domain), you can pass the anonymous id directly into the request.

Here are some examples of how you can do this:

1. Using HTTP Header: headers["X-Featureflow-Anonymous-Id"] = featureflow.getAnonymousId()
2. Adding a query param: '?ff-anonymousid='+featureflow.getAnonymousId()

API and Configuration

Globals

Featureflow.init(apiKey, [user], [config])

Returns a Promise<featureflow> instance that resolves when initialization is complete.

Params	Type	Default	Description
apiKey*	string	Required	The Featureflow Javascript API key for the environment or project you are targeting
user	FeatureflowUser	Auto-generated anonymous user	See the user object below. If not provided, an anonymous user will be automatically generated
config	Config	{}	See the config object below
return	Promise<featureflow>		Promise that resolves with a Featureflow Instance (see below)

Example:

import { init, type FeatureflowClient, type FeatureflowUser, type Config } from 'featureflow-client';

// Using async/await with TypeScript typing
const featureflow: FeatureflowClient = await init(FF_KEY, user, config);

// Or using .then()
Featureflow.init(FF_KEY, user, config).then((featureflow: FeatureflowClient) => {
  // use featureflow with full type safety
});

Featureflow Instance

These properties are available on the return of Featureflow.init(...)

featureflow.evaluate(featureKey)

Returns an object that can be used to help evaluate feature values in an expressive way.

Params	Type	Default	Description
featureKey*	string	Required	The feature key you are targeting
return	Evaluate		An Evaluate object with helper methods (.is(), .isOn(), .isOff(), .value())

TypeScript example:

import type { Evaluate } from 'featureflow-client';

const evaluation: Evaluate = featureflow.evaluate('my-feature');
const value: string = evaluation.value(); // 'on' | 'off' | variant string
const isOn: boolean = evaluation.isOn();
const isOff: boolean = evaluation.isOff();
const matches: boolean = evaluation.is('red');

featureflow.evaluate(featureKey).is(value)

Evaluates the value of a feature for the given user.

Params	Type	Default	Description
featureKey*	string	Required	The feature key you are targeting
value*	string	Required	The value you are testing against
return	boolean		true if the feature's value is equal to the value provided, otherwise false

featureflow.evaluate(featureKey).isOn()

Evaluates the value of a feature for the given user is equal to 'on'.

Params	Type	Default	Description
featureKey*	string	Required	The feature key you are targeting
return	boolean		true if the feature's value is equal to 'on' provided, otherwise false

featureflow.evaluate(featureKey).isOff()

Evaluates the value of a feature for the given user is equal to 'off'.

Params	Type	Default	Description
featureKey*	string	Required	The feature key you are targeting
return	boolean		true if the feature's value is equal to 'off' provided, otherwise false

featureflow.evaluate(featureKey).value()

Returns the value of a feature for the given user.

Params	Type	Default	Description
featureKey*	string	Required	The feature key you are targeting
return	string		The value of the feature, or the default feature value from config.defaultFeatures[featureKey] if present, or 'off'

featureflow.goal(goalKey)

Sends a goal event, along with the current evaluated features to featureflow.io. Use with experiments in the admin console.

Params	Type	Default	Description
goalKey*	string		The key of the goal you want to target.

featureflow.updateUser(user)

Updates the current user of the instance and reevaluates all feature features using the new user. Returns a Promise that resolves when the update is complete.

Params	Type	Default	Description
user	FeatureflowUser	...	See the user object below. The id property is required
return	Promise<Features>		Promise that resolves with the updated features

Fires a Featureflow.events.LOADED event when the features have been evaluated.

Example:

// JavaScript
await featureflow.updateUser(newUser);

// TypeScript
import type { FeatureflowUser, Features } from 'featureflow-client';

const newUser: FeatureflowUser = {
  id: 'user456',
  attributes: { tier: 'premium' }
};

const features: Features = await featureflow.updateUser(newUser);

featureflow.getFeatures()

Returns the current evaluated features as flat key-value map

Params	Type	Default	Description
return	EvaluatedFeatures		The current features object (map of feature keys to variant values)

Example:

import type { EvaluatedFeatures } from 'featureflow-client';

const features: EvaluatedFeatures = featureflow.getFeatures();
// features = { 'feature-1': 'on', 'feature-2': 'off', ... }

featureflow.getUser()

Returns the current user

Params	Type	Default	Description
return	FeatureflowUser		The current user

Example:

import type { FeatureflowUser } from 'featureflow-client';

const user: FeatureflowUser = featureflow.getUser();
console.log(user.id); // 'user123'
console.log(user.attributes?.tier); // 'gold'

featureflow.on(event, callback, [bindContext])

Listen to events when the featureflow instance is updated

Params	Type	Default	Description
event*	string	Required	The name of the event to subscribe to. See Events section below for available events.
callback*	EventCallback	Required	The function to call when the event is emitted.
bindContext	unknown	undefined	The context to bind the event callback to (useful for binding this in class methods).

TypeScript example:

import { events, type EvaluatedFeatures } from 'featureflow-client';

featureflow.on(events.INIT, (features: EvaluatedFeatures) => {
  console.log('Features initialized:', features);
});

// With bindContext for class methods
class MyComponent {
  handleInit(features: EvaluatedFeatures) {
    console.log('Features:', features);
  }
}

const component = new MyComponent();
featureflow.on(events.INIT, component.handleInit, component);

featureflow.off(event, [callback])

Unsubscribe from events when the featureflow instance is updated

Params	Type	Default	Description
event*	string	Required	The name of the event to unsubscribe from.
callback	EventCallback	undefined	The callback to remove. If not provided, removes all listeners for the event.

TypeScript example:

import { events, type EvaluatedFeatures } from 'featureflow-client';

const handler = (features: EvaluatedFeatures) => {
  console.log('Features:', features);
};

featureflow.on(events.INIT, handler);
// Later...
featureflow.off(events.INIT, handler);

// Or remove all listeners for an event
featureflow.off(events.INIT);

featureflow.getAnonymousId()

Returns the anonymous user id assigned for the user in localStorage.

Params	Type	Default	Description
return	string		The string of the anonymous user id in localStorage.

featureflow.hasReceivedInitialResponse()

Returns true if an initial response has been returned from the server, regardless of the status code.

Params	Type	Default	Description
return	boolean	false	true if the initial request to featureflow has completed

featureflow.resetAnonymousId()

Resets the anonymous user id for the user stored in localStorage. This will not re-evaluate the features, you must still call updateUser() to evaluate the latest features variants.

Params	Type	Default	Description
return	string		The string of the new anonymous user id.

Object Types

user (FeatureflowUser)

Property	Type	Default	Description
id	string	Required	Uniquely identifies the current user. Also used to calculate split variants. If not provided when calling init(), a random string prefixed with 'anonymous:' will be automatically generated. This will set a cookie that can be used to link the anonymous user with your server's Featureflow SDK.
attributes	object	undefined	Flat key-value object containing extra meta about the current user. Used to serve different features for specifically targeted attributes. Values can be strings, numbers, or arrays of strings/numbers.

config (Config)

All properties are optional. If not provided, sensible defaults will be used.

Property	Type	Default	Description
baseUrl	string	'https://app.featureflow.io'	The base URL for the Featureflow API
eventsUrl	string	'https://events.featureflow.io'	The URL for sending events
useCookies	boolean	true	Set to false if you do not want to use cookies (you will have to pass the result of featureflow.getAnonymousId() to any future requests if you wish for the server to match the client anonymous key)
defaultFeatures	object	{}	A flat key-value object representing the default variants a feature should be set to if there is an interrupted connection and no cached value. e.g. if you set config.defaultFeatures to {'my-feature': 'on'}, featureflow.evaluate('my-feature').isOn() will return true when there is an interrupted connection to Featureflow and no locally cached feature features.
offline	boolean	false	Set to true to run in offline mode, this is for testing purposes. Featureflow will not attempt any calls and will use the defaultFeatures values only
delayInit	boolean	false	Set to true to delay initialization. You must call featureflow.initialise() manually
uniqueEvals	boolean	true	Set to false to allow duplicate evaluation events to be sent

Events

Featureflow.events.LOADED_FROM_CACHE

Fired when features have been loaded from localstorage. Triggered by both Featureflow.init(...) and featureflow.updateUser.

Callback is fired with one parameter with the value of all evaluated features.

Featureflow.events.INIT

Fired when features have been evaluated and loaded. Triggered by both Featureflow.init(...) and featureflow.updateUser.

Callback is fired with one parameter with the value of all evaluated features.

Featureflow.events.ERROR

Fired when an error has occurred evaluating the feature features for the given user.

Callback is fired with one parameter with the error message.

License

Apache-2.0