Zebra stripes

David Clarke

A Active Model validator for Stripe Ids

In my article on using validators, the example was pretty simple.

In this example, we discuss the real production use of a more complex validator extracted from the billing code in AnswerThePublic. We use Stripe exclusively for billing in our applications.

Every object on Stripe—a transaction, an invoice, a customer—has a unique and similarly-patterned identifier. We often need to store these ids locally in our application to interact with Stripe so that they can interact with Stripe and we can collect payments from customers.

We want to make sure we don’t store incorrectly-formatted Stripe ids in our application’s database or we might introduce errors in our eventual API calls.

Instead of…

…using a regex in validates when setting a format for the stripe_id:

class Subscription < ActiveRecord
  validates :stripe_id, format: { with: /\Asub\_[A-Za-z0-9]{8,}\z/ }


…a validator class to ensure the format of a Stripe id

# app/validators/stripe_validator.rb
class StripeValidator < ActiveModel::EachValidator
  VALID_PREFIXES = %w[ch cus in plan sub txr].freeze

  def validate_each(record, attribute, value)
    record.errors[attribute] << (options[:message] || 'is not a valid Stripe ID') unless regex(options[:prefix]).match?(value)

  def check_validity!
    raise ArgumentError, 'Requires :prefix' unless options.include?(:prefix)
    raise ArgumentError, 'Invalid :prefix supplied' unless VALID_PREFIXES.include?(options[:prefix].to_s)


  def regex(value)

…then use it in models like so:

class Subscription < ActiveRecord
  validates :stripe_id, stripe: { prefix: :sub }

class Invoice < ActiveRecord
  validates :stripe_id, stripe: { prefix: :in }


I wanted to provide a useful, concrete example where extracting a validator (in this case for a popular and well-used service) can help keep your code clear and concise.

This validator can be reused across multiple models in our own domain that have records on Stripe accessible via their API, reducing potential errors, and ensuring our application data matches up to our data on Stripe.

The extra structure improves the clarity of your code when you, or another developer, come back to it. There’s now one place in our app to validate the format of all our Stripe identifiers.

Why not?

You could use a specific regex for each model’s validation, but you’d lose the extra expressiveness in the model from naming and extracting the validator.

“This field must be a stripe id with the prefix ‘in’” is clearer to read and understand than “This field must match a confusing regular expression”.

Sign up to get a nugget of Ruby knowledge every couple of weeks or so.

Last updated on September 22nd, 2019