Amplitude Ruby Library

pipeline status coverage report

Overview

This is a library designed to help developers more simply interact with the Amplitude analytics service. It communicates with Amplitude over their HTTP API.

Installation

Via the command-line:

gem install amplitude-rb

Via a Gemfile:

gem 'amplitude-rb', '~> x.y.z'

Configuration

Amplitude requires an API key to accompany all HTTP API requests. This can be configured through any of the following means (in lookup order):

  1. Passing it to an event-tracking call in the options hash via :api_key
  2. Through Amplitude.configure or by setting Amplitude.config.api_key
  3. Environment variable, AMPLITUDE_API_KEY

Examples:

# 1. Options hash, overrides everything else
Amplitude.track(event, api_key: 'abc')

# 2. Using Amplitude.configure in an initializer
Amplitude.configure do |config|
  config.api_key = 'abc'
end

# 3. Setting it as an environment variable
export AMPLITUDE_API_KEY=abc

Additional Configuration Values

time_formatter

Amplitude requires that the time key is given as milliseconds since the Unix epoch. The default formatter assumes that a number or a Time (or date) object is given and performs (time.to_f * 1_000).to_i.

event_prop_formatter

Amplitude allows for an optional set of user-defined key-value pairs relating to the event being tracked. The default formatter assumes that the given event properties responds to to_h and uses that result.

user_prop_formatter

Amplitude allows for an optional set of user-defined key-value pairs relating to the user being tracked. The default formatter assumes that the given user properties responds to to_h and uses that result.

job_queue

If ActiveJob is available then this value will be the name of the queue that the job should run in. Defaults to :default.

iid_generator

It is good practice to include the insert_id key when tracking events. This should be a proc or lambda (or anything that responds to call) that returns a unique identifier that can be used for the insert_id field. By default this will use securerandom to generate a random UUID. If you do not wish to use insert_id except when explicitly included in the event hash, set this value to nil.

Usage

Simply ensure an API key is set (or passed as an option) before calling any event-tracking functions. Then pass a hash, or Amplitude event object, to the tracking function.

Amplitude.config.api_key = 'abc'
Amplitude.track(event_hash)

# or

Amplitude.track(event_hash, api_key: 'abc')

Asynchronous Worker

If ActiveJob is available, a job workerwill be included and the Amplitude.queue function will be usable. It takes the exact same parameters as Amplitude.track.

Note: while Amplitude.track returns a result object that you may inspect for errors, Amplitude.queue will raise an error if the tracking call fails. This is because many background queues will add jobs to a retry queue if an uncaught error bubbles up, plus there's no good way to inspect the result of a tracking call when it is performed asynchronously.