Skip to content

Latest commit

 

History

History

core

README.md

segment_analytics

Warning This project is currently only available in Beta phase and is covered by Segment's First Access & Beta Preview Terms. We encourage you to try out this new library. Please provide feedback via Github issues/PRs, and feel free to submit pull requests.

The hassle-free way to add Segment analytics to your Flutter1 app.

Supports the following platforms:

  • Android
  • iOS
  • MacOS
  • Web

Some destination plugins might not support all platform functionality. Refer to their own Platform SDKs for more details.

Table of Contents

Installation

Run:

flutter pub add segment_analytics

import 'package:segment_analytics/client.dart';

Upgrading from Pilot

Since pilot phase we have renamed the package of this library from analytics to segment_analytics. Some changes have to be applied after upgrading to the v1 package:

In your pubspec.yaml remove the analytics package and use segment_analytics instead.

-   analytics:
-     git:
-       url: https://github.com/segmentio/analytics_flutter
-       ref: main
-       path: packages/core
+   segment_analytics: ^1.0.1

In your dart files change the imports from package:segment_analytics to package:segment_analytics:

- import 'package:segment_analytics/client.dart';
+ import 'package:segment_analytics/client.dart';

Permissions

Android In your app's `AndroidManifest.xml` add the below line between the `` tags. 
<uses-permission android:name="android.permission.ACCESS_NETWORK_STATE" />

Usage

Setting up the client

The package exposes a method called createClient which we can use to create the Segment Analytics client. This central client manages all our tracking events. It is recommended you add this as a property on your main app's state class.

const writeKey = 'SEGMENT_API_KEY';
final analytics = createClient(Configuration(writeKey));

You must pass at least the writeKey. Additional configuration options are listed below:

Client Options

Name Default Description
writeKey (REQUIRED) '' Your Segment API key.
debug false When set to false, it will not generate any info logs.
collectDeviceId false Set to true to automatically collect the device ID from the DRM API on Android devices.
flushPolicies count=30,time=20s List of flush policies controlling when to send batches of events to the plugins
apiHost "api.segment.io/v1" Used to specify the regional Segment event endpoint
cdnHost "cdn-settings.segment.com/v1" Used to specify the regional Segment settings endpoint
errorHandler null Custom error handler. By default logs errors to the standard flutter logger
trackApplicationLifecycleEvents false Enable automatic tracking for app lifecycle events: application installed, opened, updated, backgrounded)
trackDeeplinks false Enable automatic tracking for when the user opens the app via a deep link. *NOTE: when sending this flag, the sdk plugin_appsflyer will ignore onAppOpenAttribution
autoAddSegmentDestination true Set to false to skip adding the SegmentDestination plugin
defaultIntegrationSettings null Plugin settings that will be used if the request to get the settings from Segment fails.
maxBatchSize true 100 Maximum number of events to send to the API at once.
appStateStream null Set to override the stream of application foreground or background events.
requestFactory true Set to override the factory to generate HTTP requests. Type: RequestFactory
storageJson true Enable or disable automatic the generation JSON files for the serialization library

Client methods

Track

The track method is how you record any actions your users perform, along with any properties that describe the action.

Method signature:

Future track(String event: string, {Map<String, dynamic>? properties});

Example usage:

analytics.track("View Product", properties: {
  "productId": 123,
  "productName": "Striped trousers"
});

Screen

The screen call lets you record whenever a user sees a screen in your mobile app, along with any properties about the screen.

Method signature:

Future screen(String name: string, {Map<String, dynamic>? properties});

Example usage:

analytics.screen("ScreenName", properties: {
  "productSlug": "example-product-123",
});

For setting up automatic screen tracking, see the instructions below.

Identify

The identify call lets you tie a user to their actions and record traits about them. This includes a unique user ID and any optional traits you know about them like their email, name, etc. The traits option can include any information you might want to tie to the user, but when using any of the reserved user traits, you should make sure to only use them for their intended meaning. All reserved traits are strongly typed by the UserTraits class. When using traits not listsed as a reserved user trait, these will go under the custom property.

Method signature:

Future identify({String? userId, UserTraits? userTraits});

Example usage:

analytics.identify(userId: "testUserId", userTraits: UserTraits(
  username: "MisterWhiskers",
  email: "[email protected]",
  custom: {
    "plan": "premium"
  }
);

Group

The group API call is how you associate an individual user with a group—be it a company, organization, account, project, team or whatever other crazy name you came up with for the same concept! This includes a unique group ID and any optional group traits you know about them like the company name industry, number of employees, etc. The traits option can include any information you might want to tie to the group, but when using any of the reserved group traits, you should make sure to only use them for their intended meaning. All reserved traits are strongly typed by the GroupTraits class. When using traits not listsed as a reserved user trait, these will go under the custom property.

Method signature:

Future group(String groupId, {GroupTraits? groupTraits});

Example usage:

analytics.group("some-company", groupTraits: GroupTraits(
  name: 'Segment',
  custom: {
    "region": "UK"
  }
);

Alias

The alias method is used to merge two user identities, effectively connecting two sets of user data as one. This is an advanced method, but it is required to manage user identities successfully in some of our destinations.

Method signature:

Future alias(String newUserId);

Example usage:

analytics.alias("user-123");

Reset

The reset method clears the internal state of the library for the current user and group. This is useful for apps where users can log in and out with different identities over time.

Note: Each time you call reset, a new AnonymousId is generated automatically.

Method signature:

void reset();

Example usage:

analytics.reset();

Flush

By default, the analytics will be sent to the API after 30 seconds or when 20 items have accumulated, whatever happens sooner, and whenever the app resumes if the user has closed the app with some events unsent. These values can be modified by the flushAt and flushInterval config options. You can also trigger a flush event manually.

Method signature:

Future flush();

Example usage:

analytics.flush();

(Advanced) Cleanup

You probably don't need this!

In case you need to reinitialize the client, that is, you've called createClient more than once for the same client in your application lifecycle, use this method on the old client to clear any subscriptions and timers first.

var analytics = createClient(Configuration(writeKey));

analytics.cleanup();

analytics = createClient(Configuration(writeKey));

If you don't do this, the old client instance would still exist and retain the timers, making all your events fire twice.

Ideally, you shouldn't need this though, and the Segment client should be initialized only once in the application lifecycle.

Automatic screen tracking

Sending a screen() event with each navigation action will get tiresome quick, so you'll probably want to track navigation globally. To set this up, you'll need to add the analytics navigator observer to your app's navigator observers. For example, if you're using the MaterialApp class, you would add the following:

return MaterialApp(navigatorObservers: [
  ScreenObserver()
]);

Plugins + Timeline architecture

You have complete control over how the events are processed before being uploaded to the Segment API.

In order to customise what happens after an event is created, you can create and place various Plugins along the processing pipeline that an event goes through. This pipeline is referred to as a Timeline.

Plugin Types

Plugin Type Description
before Executed before event processing begins.
enrichment Executed as the first level of event processing.
destination Executed as events begin to pass off to destinations.
after Executed after all event processing is completed. This can be used to perform cleanup operations, etc.
utility Executed only when called manually, such as Logging.

Plugins can have their own native code (such as the iOS-only analytics_plugin_idfa) or wrap an underlying library (such as analytics_plugin_firebase which uses firebase_core and firebase_analytics under the hood)

Destination Plugins

Segment is included as a DestinationPlugin out of the box. You can add as many other DestinationPlugins as you like, and upload events and data to them in addition to Segment.

Or if you prefer, you can pass autoAddSegmentDestination = false in the options when setting up your client. This prevents the SegmentDestination plugin from being added automatically for you.

Adding Plugins

You can add a plugin at any time through the add() method.

import 'package:segment_analytics/client.dart';
import 'package:segment_analytics/event.dart';
import 'package:segment_analytics/state.dart';
import 'package:segment_analytics_plugin_advertising_id/plugin_advertising_id.dart';
import 'package:segment_analytics_plugin_idfa/plugin_idfa.dart';
import 'package:segment_analytics_plugin_firebase/plugin_firebase.dart'
    show FirebaseDestination;

const writeKey = 'SEGMENT_API_KEY';

class _MyAppState extends State<MyApp> {
  final analytics = createClient(Configuration(writeKey));

  @override
  void initState() {
    super.initState();
    initPlatformState();

    analytics
        .addPlugin(FirebaseDestination(DefaultFirebaseOptions.currentPlatform));
    analytics.addPlugin(PluginAdvertisingId());
    analytics.addPlugin(PluginIdfa());
  }
}

Writing your own Plugins

Plugins are implemented by extending one of the provided plugin classes. The available plugin classes are:-

  • Plugin
  • EventPlugin
  • DestinationPlugin
  • UtilityPlugin
  • PlatformPlugin

Any plugins must be an extension of one of these classes.

You can them customise the functionality by overriding different methods on the base class. For example, here is a simple Logger plugin:

import 'dart:convert';

import 'package:segment_analytics/analytics.dart';
import 'package:segment_analytics/event.dart';
import 'package:segment_analytics/plugin.dart';
import 'package:segment_analytics/logger.dart';

class EventLogger extends DestinationPlugin {
  var logKind = LogFilterKind.debug;

  EventLogger() : super("event_logger");

  @override
  void configure(Analytics analytics) {
    pAnalytics = analytics;
  }

  @override
  Future<RawEvent?>? execute(RawEvent event) async {
    log("${event.type.toString().toUpperCase()} event${event is TrackEvent ? " (${event.event})" : ''} saved: \n${jsonEncode(event.toJson())}",
        kind: logKind);
    return event;
  }
}

As it overrides the execute() method, this Logger will call log for every event going through the Timeline.

Supported Plugins

Refer to the following table for Plugins you can use to meet your tracking needs:

Plugin Package
Adjust analytics_plugin_adjust
AppsFlyer analytics_plugin_appsflyer
Firebase analytics_plugin_firebase
IDFA analytics_plugin_idfa
Android Advertising ID analytics_plugin_advertising-id

Controlling Upload With Flush Policies

To more granurily control when events are uploaded you can use FlushPolicies

A Flush Policy defines the strategy for deciding when to flush, this can be on an interval, on a certain time of day, after receiving a certain number of events or even after receiving a particular event. This gives you even more flexibility on when to send event to Segment.

To make use of flush policies you can set them in the configuration of the client:

import 'package:segment_analytics/flush_policies/count_flush_policy.dart';
import 'package:segment_analytics/flush_policies/timer_flush_policy.dart';

final analytics = createClient(Configuration(/*...*/, flushPolicies: [
  CountFlushPolicy(10),
  TimerFlushPolicy(100000)
]));

You can set several policies at a time. Whenever any of them decides it is time for a flush it will trigger an upload of the events. The rest get reset so that their logic restarts after every flush.

That means only the first policy to reach shouldFlush gets to trigger a flush at a time. In the example above either the event count gets to 5 or the timer reaches 500ms, whatever comes first will trigger a flush.

We have several standard FlushPolicies:

  • CountFlushPolicy triggers whenever a certain number of events is reached
  • TimerFlushPolicy triggers on an interval of milliseconds
  • StartupFlushPolicy triggers on client startup only

Adding or removing policies

One of the main advatanges of FlushPolicies is that you can add and remove policies on the fly. This is very powerful when you want to reduce or increase the amount of flushes.

For example you might want to disable flushes if you detect the user has no network:

if (isConnected) {
  analytics.addFlushPolicy(policiesIfNetworkIsUp);
} else {
  analytics.removeFlushPolicy(policiesIfNetworkIsUp)
}

Creating your own flush policies

You can create a custom FlushPolicy special for your application needs by implementing the FlushPolicy interface. You can also extend the FlushPolicyBase class that already creates and handles the shouldFlush value reset.

A FlushPolicy only needs to implement 1 method:

  • onEvent(RawEvent event): Gets called on every event tracked by your client

and optionally can implement:

  • reset(): Called after a flush is triggered (either by your policy, by another policy or manually)
  • start(): Executed when the flush policy is enabled and added to the client. This is a good place to start background operations, make async calls, configure things before execution

They also have a shouldFlush boolean value. When this is set to true the client will atempt to upload events. Each policy should reset this value to false according to its own logic, although it is pretty common to do it inside the reset method.

import 'package:segment_analytics/event.dart';
import 'package:segment_analytics/flush_policies/flush_policy.dart';

class FlushOnScreenEventsPolicy extends FlushPolicy {

  @override
  onEvent(RawEvent event) {
    // Only flush when a screen even happens
    if (event is ScreenEvent) {
      this.shouldFlush = true;
    }
  }

  @override
  reset() {
    // Superclass will reset the shouldFlush value so that the next screen event triggers a flush again
    // But you can also reset the value whenever, say another event comes in or after a timeout
    super.reset();
  }
}

Custom Logging

By default any logging is done via the standard Flutter logging mechanism. To customise logging, you can build your own logger, which must implement the LogTarget mixin. For example:

import 'package:segment_analytics/logger.dart';

void customDebugLog(String msg) {
  // ...
}

void customWarningLog(String msg) {
  // ...
}

void customErrorLog(String msg) {
  // ...
}

class CustomLogger with LogTarget {
  @override
  void parseLog(LogMessage log) {
    switch (log.kind) {
      case LogFilterKind.debug:
        customDebugLog("Segment: ${log.message}");
        break;
      case LogFilterKind.warning:
        customWarningLog("Segment: ${log.message}");
        break;
      case LogFilterKind.error:
        customErrorLog("Segment: ${log.message}");
        break;
    }
  }
}

// Set the default logger to use the CustomLogger
LogFactory.logger = CustomLogger();

Handling errors

You can handle analytics client errors through the errorHandler option.

The error handler configuration receives a function which will get called whenever an error happens on the analytics client. It will receive an Exception, that will extend one of the errors from errors.dart.

You can use this error handling to trigger different behaviours in the client when a problem occurs. For example if the client gets rate limited you could use the error handler to swap flush policies to be less aggressive:

import 'package:segment_analytics/errors.dart';

//...

final flushPolicies = [CountFlushPolicy(5), TimerFlushPolicy(500)];

void errorHandler(Exception error) {
  if (error is NetworkServerLimited) {
    // Remove all flush policies
    analytics.removeFlushPolicy(analytics.getFlushPolicies());
    // Add less persistent flush policies
    analytics.addFlushPolicy([
      CountFlushPolicy(100),
      TimerFlushPolicy(5000)
    ]);
  }
}

final analytics = createClient(Configuration(writeKey),
  errorHandler: errorHandler,
  flushPolicies: flushPolicies);

Reporting errors from plugins

Plugins can also report errors to the handler by using the .error function of the analytics client, we recommend using the PluginError for consistency, and attaching the innerError with the actual exception that was hit:

import 'package:segment_analytics/errors.dart';

//...

try {
  distinctId = await mixpanel.getDistinctId();
} catch (e) {
  analytics.error(
    PluginError('Error: Mixpanel error calling getDistinctId', e)
  );
}

Example App

See the example app to check a full test app of how to integrate Analytics-Flutter into your own Flutter app.

Contributing

See the contributing guide to learn how to contribute to the repository and the development workflow.

Code of Conduct

Before contributing, please also see our code of conduct.

License

MIT

Footnotes

  1. The Flutter name and logo are trademarks owned by Google. ↩