中文版正在翻译中，敬请期待

点击查看Github SDK源码与ReadMe

Introduction

This is the JavaScript client side SDK for the feature management platform feature-flags.co. We will document all the methods available in this SDK, and detail how they work.

Be aware, this is a client side SDK, it is intended for use in a single-user context, which can be mobile, desktop or embeded applications. This SDK can only be ran in a browser environment, it is not suitable for NodeJs applications, server side SDKs are available in our other repos.

This SDK has two main works:

• Makes feature flags avaible to the client side code
• Sends feature flags usage, click, pageview and custom events for the insights and A/B/n testing.

Data synchonization

We use websocket to make the local data synchronized with the server, and then persist in localStorage. Whenever there is any changes to a feature flag, the changes would be pushed to the SDK, the average synchronization time is less than 100 ms. Be aware the websocket connection can be interrupted by any error or internet interruption, but it would be restored automatically right after the problem is gone.

Offline mode support

As all data is stored locally in the localStorage, in the following situations, the SDK would still work when there is temporarily no internet connection:

• it has already recieved the data from previous conections
• the Ffc.bootstrap(featureFlags) method is called with all necessary feature flags

In the mean time, the SDK would try to reconnect to the server by an incremental interval, this makes sure that the websocket would be restored when the internet connection is back.

Evaluation of a feature flag

After initialization, the SDK has all the feature flags locally and it does not need to request the remote server for any feature flag evaluation. All evaluation is done locally and synchronously, the average evaluation time is about 1 ms.

Getting started

Install

npm

npm install ffc-js-client-side-sdk

yarn

yarn add ffc-js-client-side-sdk

To import the SDK:

// Using ES2015 imports
import Ffc from 'ffc-js-client-side-sdk';

// Using TypeScript imports
import Ffc from 'ffc-js-client-side-sdk';

// Using react imports
import Ffc from 'ffc-js-client-side-sdk';

If using typescipt and seeing the following error:

Cannot find module 'ffc-js-client-sdk/esm'. Did you mean to set the 'moduleResolution' option to 'node', or to add aliases to the 'paths' option?

just add this in your tsconfig.json file

  "compilerOptions": {
    "moduleResolution": "node"
  },

Initializing the SDK

Before initializing the SDK, you need to get the client-side env secret of your environment from our SaaS platform.

const option = {
    secret: 'your env secret',
    user: {
      userName: 'the user's user name',
      id: 'the user's unique identifier'
    }
};

Ffc.init(option);

The complete list of the available parameters in option:

• secret: the client side secret of your environment. mandatory (NB. this becomes optional if enableDataSync equals false)
• anonymous: true if you want to use a anonymous user, which is the case before user login to your APP. If that is your case, the user can be set later with the identify method after the user has logged in. The default value is false. not mandatory
• bootstrap: init the SDK with feature flags, this will trigger the ready event immediately instead of requesting from the remote. not mandatory
• enableDataSync: false if you do not want to sync data with remote server, in this case feature flags must be set to bootstrap option or be passed to the method bootstrap. The default value is true. not mandatory
• devModePassword: if setted, the developer mode can only be activated by calling the method activateDevMode with password on Ffc . not mandatory
• api: the API url of the server, set it only if you are self hosting the back-end. not mandatory
• appType: the app type, the default value is javascript, not mandatory
• user: the user connected to your APP, can be ignored if anonymous equals to true.

  • userName: the user name. mandatory

  • id: the unique identifier. mandatory

  • email: can be useful when you configure your feature flag rules. not mandatory

  • country: can be useful when you configure your feature flag rules. not mandatory

  • customizedProperties: any customized properties you want to send to the back end. It is extremely powerful when you define targeting rules or segments. not mandatory

    • it must have the following format:

     [{
       "name": "the name of the property",
       "value": "the value of the property"
     }]
    

Initialization delay

Initializing the client makes a remote request to featureflag.co, so it may take 100 milliseconds or more before the SDK emits the ready event. If you require feature flag values before rendering the page, we recommend bootstrapping the client. If you bootstrap the client, it will emit the ready event immediately.

Get the varation value of a feature flag

Two methods to get the variation of a feature flag

// Use this method for all cases
var flagValue = Ffc.variation("YOUR_FEATURE_KEY", 'the default value');

// Syntactic sugar of the variation method, but this method return a boolean value instead of string. Use this method if the options are strings of true or false
var the defaultValue = true; // or false
var flagValue = Ffc.boolVariation("YOUR_FEATURE_KEY", defaultValue);

Developer mode

Developer mode is a powerful tool we created allowing developers to manipulate the feature flags locally instead of modifying them on feature-flags.co. This will not change the remote values.

To activate the developer mode, the activateDevMode method should be called as following, the password parameter is

// This will activate developer mode, you should be able to see an icon on bottom right of the screen. 
// PASSWORD should be the same as the value passed to option or empty if devModePassword is not set
Ffc.activateDevMode('PASSWORD'); 

// or
// this method is equivalent to Ffc.activateDevMode('PASSWORD')
window.activateFfcDevMode('PASSWORD'); 

To open the developer mode editor or quit developer mode, use the following code:

// The method will open the developer mode editor, or you can just click on the developer mode icon
Ffc.openDevModeEditor(); 

// call this method to quit developer mode
Ffc.quitDevMode();

// or
// this is equivalent to Ffc.quitDevMode()
window.quitFfcDevMode();

bootstrap

If you already have the feature flags available, two ways to pass them to the SDK instead of requesting from the remote.

• By the init method

  // define the option with the bootstrap parameter
  const option = {
    ...
    bootstrap = [{ // the array should contain all your feature flags
      id: string, // the feature flag key
      variation: string,
      sendToExperiment: boolean,
      timestamp: number,
      variationOptions: [{
        id: number,
        value: string
      }]
    }]
    ...
  }

  Ffc.init(option);

• By the bootstrap method

const featureflags = [{ // the array should contain all your feature flags
  id: string, // the feature flag key
  variation: string,
  sendToExperiment: boolean,
  timestamp: number,
  variationOptions: [{
    id: number,
    value: string
  }]
}]

Ffc.bootstrap(featureflags);

If you want to disable the synchronization with remote server, set enableDataSync to false in option. In this case, bootstrap option must be set or bootstrap method must be called with feature flags.

To find out when the client is ready, you can use one of two mechanisms: events or promises.

The client object can emit JavaScript events. It emits a ready event when it receives initial flag values from feature-flags.co. You can listen for this event to determine when the client is ready to evaluate flags.

Ffc.on('ready', (data) => {
  // data has the following structure [ {id: 'featureFlagKey', variation: 'variation value'} ]
  var flagValue = Ffc.variation("YOUR_FEATURE_KEY", 'the default value');
});

Or, you can use a promise instead of an event. The SDK has a method that return a promise for initialization: waitUntilReady(). The behavior of waitUntilReady() is equivalent to the ready event. The promise resolves when the client receives its initial flag data. As with all promises, you can either use .then() to provide a callback, or use await if you are writing asynchronous code.

Ffc.waitUntilReady().then((data) => {
  // data has the following structure [ {id: 'featureFlagKey', variation: 'variation value'} ]
  // initialization succeeded, flag values are now available
});
// or, with await:
const featureFlags = await Ffc.waitUntilReady();
// initialization succeeded, flag values are now available

The SDK only decides initialization has failed if it receives an error response indicating that the environment ID is invalid. If it has trouble connecting to feature-flags.co, it will keep retrying until it succeeds.

Set the user after initialization

If the user parameter cannot be passed by the init method, the following method can be used to set the user after initialization.

  Ffc.identify(user);

Set the user to anonymous user

We can manully call the method logout, which will switch the current user back to anonymous user if exists already or create a new anonymous user.

  Ffc.logout(user);

Subscribe to the changes of feature flag(s)

To get notified when a feature flag is changed, we offer two methods

• subscribe to the changes of any feature flag(s)

Ffc.on('ff_update', (changes) => {
  // changes has this structure [{id: 'the feature_flag_key', oldValue: '', newValue: ''}]
  ...
});

• subscribe to the changes of a specific feature flag

// replace feature_flag_key with your feature flag key
Ffc.on('ff_update:feature_flag_key', (change) => {
  // change has this structure {id: 'the feature_flag_key', oldValue: '', newValue: ''}
  const myFeature = Ffc.variation('feature_flag_key', 'default value');
  ...
});

Experiments (A/B/n Testing)

We support automatic experiments for pageviews and clicks, you just need to set your experiment on our SaaS platform, then you should be able to see the result in near real time after the experiment is started.

In case you need more control over the experiment data sent to our server, we offer a method to send custom event.

Ffc.sendCustomEvent([{
    eventName: 'your event name',
    numericValue: 1
}])

numericValue is not mandatory, the default value is 1.

Make sure sendCustomEvent is called after the related feature flag is called by simply calling Ffc.variation('featureFlagKeyName', 'default value'), otherwise, the custom event won't be included into the experiment result.