This Ruby gem lets you move your application logic into into small composable service objects. It is a lightweight framework that helps you keep your models and controllers thin.

• Installation
• Usage
  • Inputs
  • Outputs
  • Defaults
  • Types
  • Allow nil
  • Conditions
  • Result
• Play actors in a sequence
  • Rollback
  • Early success
  • Lambdas
  • Play conditions
• Testing
• Influences
• Development
• Contributing
• License

Add these lines to your application's Gemfile:

# Composable service objects
gem 'service_actor'

Actors are single-purpose actions in your application that represent your business logic. They start with a verb, inherit from Actor and implement a call method.

# app/actors/send_notification.rb
class SendNotification < Actor
  def call
    # …
  end
end

Trigger them in your application with .call:

SendNotification.call # => <Actor::Result…>

When called, actors return a Result. Reading and writing to this result allows actors to accept and return multiple arguments. Let's find out how to do that.

To accept arguments, use input to create a method named after this input:

class GreetUser < Actor
  input :user

  def call
    puts "Hello #{user.name}!"
  end
end

You can now call your actor by providing the correct arguments:

GreetUser.call(user: User.first)

An actor can return multiple arguments. Declare them using output, which adds a setter method to let you modify the result from your actor:

class BuildGreeting < Actor
  output :greeting

  def call
    self.greeting = 'Have a wonderful day!'
  end
end

The result you get from calling an actor will include the outputs you set:

result = BuildGreeting.call
result.greeting # => "Have a wonderful day!"

Inputs can be marked as optional by providing a default:

class BuildGreeting < Actor
  input :name
  input :adjective, default: 'wonderful'
  input :length_of_time, default: -> { ['day', 'week', 'month'].sample }

  output :greeting

  def call
    self.greeting = "Have a #{adjective} #{length_of_time} #{name}!"
  end
end

This lets you call the actor without specifying those keys:

result = BuildGreeting.call(name: 'Jim')
result.greeting # => "Have a wonderful week Jim!"

If an input does not have a default, it will raise a error:

result = BuildGreeting.call
=> Actor::ArgumentError: Input name on BuildGreeting is missing.

You can add simple conditions that the inputs must verify, with the name of your choice under must:

class UpdateAdminUser < Actor
  input :user,
        must: {
          be_an_admin: ->(user) { user.admin? }
        }

  # …
end

In case the input does not match, it will raise an argument error.

Sometimes it can help to have a quick way of making sure we didn't mess up our inputs. For that you can use type with the name of a class or an array of possible classes it must be an instance of.

class UpdateUser < Actor
  input :user, type: 'User'
  input :age, type: %w[Integer Float]

  # …
end

An exception will be raised if the type doesn't match.

By default inputs accept nil values. To raise an error instead:

class UpdateUser < Actor
  input :user, allow_nil: false

  # …
end

All actors return a successful result by default. To stop the execution and mark an actor as having failed, use fail!:

class UpdateUser
  input :user
  input :attributes

  def call
    user.attributes = attributes

    fail!(error: 'Invalid user') unless user.valid?

    # …
  end
end

This will raise an error in your app with the given data added to the result.

To test for the success of your actor instead of raising an exception, use .result instead of .call. This lets you use success? and failure? on the result.

For example in a Rails controller:

# app/controllers/users_controller.rb
class UsersController < ApplicationController
  def create
    result = UpdateUser.result(user: user, attributes: user_attributes)
    if result.success?
      redirect_to result.user
    else
      render :new, notice: result.error
    end
  end
end

Any keys you add to fail! will be added to the result, for example you could do: fail!(error_type: "validation", error_code: "uv52", …).

To help you create actors that are small, single-responsibility actions, an actor can use play to call other actors:

class PlaceOrder < Actor
  play CreateOrder,
       Pay,
       SendOrderConfirmation,
       NotifyAdmins
end

This creates a call method where each actor will be called, taking their arguments from the previous actor's result. In fact, every actor along shares the same result instance to help shape the final result your application needs.

When using play, when an actor calls fail!, the following actors will not be called.

Instead, all the actors that succeeded will have their rollback method called in reverse order. This allows actors a chance to cleanup, for example:

class CreateOrder < Actor
  output :order

  def call
    self.order = Order.create!(…)
  end

  def rollback
    order.destroy
  end
end

Rollback is only called on the previous actors in play and is not called on the failing actor itself. Actors should be kept to a single purpose and not have anything to clean up if they call fail!.

When using play you can use succeed! to stop the execution of the following actors, but still consider the actor to be successful.

You can use inline actions using lambdas. Inside these lambdas, you don't have getters and setters but have access to the shared result:

class Pay < Actor
  play ->(result) { result.payment_provider = "stripe" },
       CreatePayment,
       ->(result) { result.user_to_notify = result.payment.user },
       SendNotification
end

Like in this example, lambdas can be useful for small work or preparing the result for the next actors. If you want to do more work before, or after the whole play, you can also override the call method. For example:

class Pay < Actor
  # …

  def call
    Time.with_timezone('Paris') do
      super
    end
  end
end

Actors in a play can be called conditionally:

class PlaceOrder < Actor
  play CreateOrder,
       Pay
  play NotifyAdmins, if: ->(ctx) { ctx.order.amount > 42 }
end

In your application, add automated testing to your actors as you would do to any other part of your applications.

You will find that cutting your business logic into single purpose actors makes your application much simpler to test.

This gem is heavily influenced by Interactor ♥. However there are a few key differences which make actor unique:

• Does not hide errors when an actor fails inside another actor.
• Requires you to document all arguments with input and output.
• Defaults to raising errors on failures: actor uses call and result instead of call! and call. This way, the default is to raise an error and failures are not hidden away because you forgot to use !.
• Allows defaults, type checking, requirements and conditions on inputs.
• Delegates methods on the context: foo vs context.foo, self.foo = vs context.foo = , fail!vscontext.fail!`.
• Shorter setup syntax: inherit from < Actor vs having to include Interactor and include Interactor::Organizer.
• Organizers allow lambdas, being called multiple times, and having conditions.
• Allows triggering an early success with succeed!.
• No before, after and around hooks, prefer using play with lambdas or overriding call.

Thank you to @nicoolas25, @AnneSottise & @williampollet for the early thoughts and feedback on this gem.

After checking out the repo, run bin/setup to install dependencies. Then, run rake to run the tests and linting. You can also run bin/console for an interactive prompt.

To release a new version, update the version number in version.rb, and then run rake release, which will create a git tag for the version, push git commits and tags, and push the gem to rubygems.org.

Bug reports and pull requests are welcome on GitHub.

This project is intended to be a safe, welcoming space for collaboration, and everyone interacting in the project’s codebase and issue tracker is expected to adhere to the Contributor Covenant code of conduct.

The gem is available as open source under the terms of the MIT License.