Using attribute option with custom primary key will break fetching

[ellnix]

Description
If you provide a custom primary key method and also specify attributes other than the real primary key of the model in the meilisearch block, the documents in the meilisearch instance will not contain the model's real primary key and therefore cannot map to database records.

Expected behavior
MeiliSearch still includes ID in its documents and fetching does not fail.

Reproduction script

require 'bundler/inline'

gemfile do
  gem "rails", "~> 7.1.3"
  gem "sqlite3", "~> 1.4"
  gem "meilisearch-rails", "~> 0.11.1"
end

MeiliSearch::Rails.configuration = {
  meilisearch_url: ENV.fetch('MEILISEARCH_HOST', 'http://localhost:7700'),
  meilisearch_api_key: ENV.fetch('MEILISEARCH_API_KEY', 'masterKey')
}

MeiliSearch::Rails.client.delete_index 'dogs'
MeiliSearch::Rails.client.delete_index 'cats'
sleep 0.1

FileUtils.rm('data.sqlite3') if File.exist?('data.sqlite3')

require 'active_record'

ActiveRecord::Base.establish_connection(
  'adapter' => defined?(JRUBY_VERSION) ? 'jdbcsqlite3' : 'sqlite3',
  'database' => 'data.sqlite3',
  'pool' => 5,
  'timeout' => 5000
)

ActiveRecord::Schema.define do
  create_table(:dogs) { |t| t.string :name }
  create_table(:cats) { |t| t.string :name }
end

# With custom primary key
class Dog < ActiveRecord::Base
  include MeiliSearch::Rails

  meilisearch index_uid: 'dogs', primary_key: :ms_id do
    attribute :name
  end

  def ms_id
    "dog_#{id}"
  end
end

# Without custom primary key
class Cat < ActiveRecord::Base
  include MeiliSearch::Rails

  meilisearch index_uid: 'cats' do
    attribute :name
  end
end

Dog.create name: 'spikey'
Cat.create name: 'fluffy'

sleep 0.5

p Dog.search 'spikey' # Will return nothing
# []

p Dog.index.search('spikey')['hits'] # Documents without IDs
# [{"ms_id"=>"dog_2", "name"=>"spikey"}]

p Cat.search 'fluffy' # Returns a cat correctly
# [#<Cat id: 1, name: "fluffy">]

p Cat.index.search('fluffy')['hits'] # Documents still have IDs
# [{"id"=>"1", "name"=>"fluffy"}]

[hack3rvaillant]

I think I encountered that issue and the task is failing because of that.

[hack3rvaillant]

NVM is was because I created the index from the Cloud dashboard without being able to specify the primary key