Where should I create an after create hook - ruby-on-rails

Question

I have the following models:
class Page < ApplicationRecord
has_one :architecture
end
class Architecture < ApplicationRecord
belongs_to :page
end
And after a new page is saved I need to capture it's architecture (number of paragraphs por example). I would like to know what is the proper way to do that. I'm not sure if I should leave that responsible for the Page model:
class Page < ApplicationRecord
has_one :architecture
after_create :scrape_architecture
private
def scrape_architecture
data = call_something_to_capture_architecture(url)
create_architecture(data)
end
end
class Architecture < ApplicationRecord
belongs_to :page
end
or if it should be the responsibility of the Architecture model:
class Page < ApplicationRecord
has_one :architecture
after_create :create_architecture
end
class Architecture < ApplicationRecord
belongs_to :page
before_create :scrape_page
private
def scrape_page
data = call_something_to_capture_architecture(page.url)
create(data)
end
end
Which is actually incorrectly because before_create runs after the validation – causing MySQL errors duo to non null constraints
Thank you.

Answer 1

I would just create a job or service object that handles scraping.
class PageScrapingJob < ApplicationJob
queue_as :default
def perform(page)
data = call_something_to_capture_architecture(page.url)
architecture = page.create_actitecture(data)
# ...
end
end
You would then call the service/job in your controller after saving the page:
class PagesController < ApplicationController
def create
#page = Page.new(page_params)
if #page.save
PageScrapingJob.perform_now(#page)
redirect_to #page
else
render :new
end
end
end
This gives you a perfect control of exactly when this is fired and avoids placing even more responsibilities onto your models. Even though your models may contain little code they have a huge amount of responsibilities such as validations, I18n, form binding, dirty tracking etc that are provided by ActiveModel and ActiveRecord. The list really goes on and on.
This instead creates a discrete object that does only one job (and hopefully does it well) and that can be tested in isolation from the controller.

Answer 2

For such things you could use a service pattern
class PageScrapper
Result = Struct.new(:success?, :data)
def initialize(url)
#url = url
end
def call
result = process(#url)
...
if result.success? # pseudo methods
Result.new(true, result)
else
Result.new(false, nil)
end
end
end
class Fetcher
...
def call
scrapper = PageScrapper.new(url)
result = scrapper.call
if scrapper.success?
page = Page.build(parsed_result_if_needed(result)
page.architecture.build(what_you_need)
page.save # here you need to add error handling if save fails
else
# error handling
end
end
There are a lot of resources about why callbacks are bad.
Here is one from Marcin Grzywaczewski but you can also google it "why callbacks are bad ruby on rails".
By using service you are liberating models from having too much business logic and they do not need to know about other parts of your system.