I am doing TDD/BDD in Ruby on Rails 3 with Rspec (2.11.0) and FactoryGirl (4.0.0). I have a factory for a Category model:
FactoryGirl.define "Category" do
factory :category do
name "Foo"
end
end
If I drop, create then migrate the database in the test enviroment I get this error:
rake aborted!
Could not find table 'categories'
This problem occurs because FactoryGirl expects the tables to already exist (for some odd reason). If I remove the spec folder from my rails app and do db:migrate, it works. Also if I mark factory-girl-rails from my Gemfile as :require => false it also works (then I have to comment that require in order to run rspec).
I found some information about this problem here: https://github.com/thoughtbot/factory_girl/issues/88
Is there something wrong that I'm doing? How can I "pass by" the FactoryGirl stage in the db:migration task?
I think you need to have factory girl definition like that in Gemfile:
gem 'factory_girl_rails', :require => false
And then you just require it in your spec_helper.rb like that:
require 'factory_girl_rails'
This is the way I'm always using this gem. You don't need to require it in other places than spec_helper.rb. Your current desired approach is just wrong.
A simple fix to this issue is to delay evaluation of any models in your factories by wrapping them in blocks. So, instead of this:
factory :cake do
name "Delicious Cake"
frosting Frosting.new(:flavor => 'chocolate')
filling Filling.new(:flavor => 'red velvet')
end
Do this (notice the curly braces):
factory :cake do
name "Delicious Cake in a box"
frosting { Frosting.new(:flavor => 'chocolate') }
filling { Filling.new(:flavor => 'red velvet') }
end
If you have a lot of factories this may not be feasible, but it is rather straightforward. See also here.
Information from: http://guides.rubyonrails.org/testing.html
When you do end up destroying your testing database (and it will happen, trust me),
you can rebuild it from scratch according to the specs defined in the development
database. You can do this by running rake db:test:prepare.
The rake db:migrate above runs any pending migrations on the development environment
and updates db/schema.rb. The rake db:test:load recreates the test database from the
current db/schema.rb. On subsequent attempts, it is a good idea to first run db:test:prepare, as it first checks for pending migrations and warns you appropriately.
rake db:test:clone Recreate the test database from the current environment’s database schema
rake db:test:clone_structure Recreate the test database from the development structure
rake db:test:load Recreate the test database from the current schema.rb
rake db:test:prepare Check for pending migrations and load the test schema
rake db:test:purge Empty the test database.
You shouldn't need to do any of that.. I think the issue is that your argument to FactoryGirl.define..
try this.
FactoryGirl.define do
factory :category do
name "Foo"
end
end
That should work fine, and does not screw up my migrations or load.. Today, I had to fix an issue where I was referencing a model constant from my factory directly and had to put it in a block to fix things.
FactoryGirl.define do
factory :category do
# this causes unknown table isseus
# state Category::Active
# this does not.
state { Category::Active }
end
end
Related
I got a rake task which invokes other rake tasks, so my development data can be easily reset.
the first rake task (lib/tasks/populate.rake)
# Rake task to populate development database with test data
# Run it with "rake db:populate"
namespace :db do
desc 'Erase and fill database'
task populate: :environment do
...
Rake::Task['test_data:create_company_plans'].invoke
Rake::Task['test_data:create_companies'].invoke
Rake::Task['test_data:create_users'].invoke
...
end
end
the second rake task (lib/tasks/populate_sub_scripts/create_company_plans.rake)
namespace :test_data do
desc 'Create Company Plans'
task create_company_plans: :environment do
Company::ProfilePlan.create!(name: 'Basic', trial_period_days: 30, price_monthly_cents: 4000)
Company::ProfilePlan.create!(name: 'Professional', trial_period_days: 30, price_monthly_cents: 27_500)
Company::ProfilePlan.create!(name: 'Enterprise', trial_period_days: 30, price_monthly_cents: 78_500)
end
end
when I run bin/rake db:populate then i get this error
rake aborted! LoadError: Unable to autoload constant
Company::ProfilePlan, expected
/home/.../app/models/company/profile_plan.rb to define it
but when I run the second rake task independently it works well.
The model (path: /home/.../app/models/company/profile_plan.rb)
class Company::ProfilePlan < ActiveRecord::Base
# == Constants ============================================================
# == Attributes ===========================================================
# == Extensions ===========================================================
monetize :price_monthly_cents
# == Relationships ========================================================
has_many :profile_subscriptions
# == Validations ==========================================================
# == Scopes ===============================================================
# == Callbacks ============================================================
# == Class Methods ========================================================
# == Instance Methods =====================================================
end
Rails 5.0.1
Ruby 2.4.0
The App was just upgraded from 4.2 to 5
It works when I require the whole path:
require "#{Rails.root}/app/models/company/profile_plan.rb"
But this seems strange to me, because in the error message rails has the correct path to the Model. Does someone know why I have to require the file when invoked from another rake task?
Thank you very much
Well, it seems that rake doesn't eager load, so when you call the create_company_plans.rake alone it loads the referred objects, however when you invoke it from another rake, it doesn't know you will need them and so they are not loaded.
You can take a look at this other QA which was similar to yours.
I think maybe you don't need to require the whole path, just:
require 'models/company/profile_plan'
From what I understand, you can probably overcome the problem by reenable ing and then revoke ing the task as given below. Pardon me if this doesn't work.
['test_data:create_company_plans', 'test_data:create_companies'].each do |task|
Rake::Task[task].reenable
Rake::Task[task].invoke
end
There is more info on this stackoverflow question how-to-run-rake-tasks-from-within-rake-tasks .
If you are using multiple databases (e.g. to shard tables across databases) for a given environment is it possible to use the built in Rails rake tasks to manipulate the databases beyond the primary one for said environment?
e.g. if I use specific connection details for a set of models, can I use rake db:create to create said database?
If so, how?
I've actually overwritten the built-in Rails rake tasks in order to manipulate a second database. The neat thing is that it totally works with the original rake tasks as well.
This is what I did... credit to Nikolay Sturm, as I based my rake task off of his blog.
I added new database settings in my database.yml, as detailed in the blog I linked.
You also create a base model that has these database settings:
class BaseModel < ActiveRecord::Base
establish_connection "database_connection_name"
self.abstract_class = true #you want to make this an abstract class because you don't have a table for this
end
You then have any model that connects to this secondary database inherit from your BaseModel.
In your rake task, you set it up as follows (keep in mind that this is for a Postgres connection.)
namespace :db do
desc 'create database'
task create: :environment do
connection_config = BaseModel.connection_config #rename BaseModel to name of your model
ActiveRecord::Base.establish_connection connection_config.merge('database => 'postgres') #you don't need this bit if you're not using Postgres
ActiveRecord::Base.connection.create_database connection_config[:database], connection_config
ActiveRecord::Base.establish_connection connection_config
end
desc 'migrate database'
task migrate: :environment do
ActiveRecord::Base.establish_connection BaseModel.connection_config
ActiveRecord::Migrator.migrate('db/migrate/')
end
desc 'drop database'
task drop: :environment do
connection_config = BaseModel.connection_config
ActiveRecord::Base.establish_connection connection_config.merge('database' => 'postgres')
ActiveRecord::Base.connection.drop_database connection_config[:database]
end
Now, if you do a rake -T in your project, you should be able to see the your descriptions for the rake tasks instead of the default. When you run them, it will run the db:create, db:migrate, and db:drop for both your default and your secondary database. Hope this helps!
There are comments in the rails codebase that indicate that the test database should be reset between runs
rake -T
rake test:all # Run tests quickly by merging all types and not resetting db
rake test:all:db # Run tests quickly, but also reset db
config/database.yml
# Warning: The database defined as "test" will be erased and
# re-generated from your development database when you run "rake".
# Do not set this db to the same as development or production.
test:
This doesn''t seem to be the case for me.
I'm using factory girl generate test models, here is an example factory
FactoryGirl.define do
factory :podcast do
sequence(:title) { |n| "Podcast #{n}" }
sequence(:feed_url) { |n| "http://podcast.com/#{n}" }
end
end
The podcast should have a unique feed_url so I validate it's uniqueness in the model.
class Podcast < ActiveRecord::Base
validates :feed_url, uniqueness: true, presence: true
end
In test_helper.rb I lint all factories
ENV["RAILS_ENV"] ||= "test"
require File.expand_path('../../config/environment', __FILE__)
require 'rails/test_help'
require 'minitest/autorun'
FactoryGirl.lint
My test creates a podcast, builds another with the same name, then asserts that the second
is invalid.
require 'test_helper'
describe Podcast do
describe '#feed_url' do
it 'must be unique' do
podcast = create(:podcast)
new_podcast = build(:podcast, feed_url: podcast.name)
assert_invalid podcast, :feed_url, 'has already been taken'
end
end
end
The first time I run the tests it executes without errors and the tests all pass.
The second time I run the tests the Factory Girl lint fails because podcast feed_url has already been taken.
Why isn't the test database being rebuilt between runs?
We have a more involved FactoryGirl set up that prepares our database with some canonical items, but I think you could probably put this code directly in your test_helper.rb to assure the database is emptied:
# Destroy all models because they do not get destroyed automatically
(ActiveRecord::Base.connection.tables - %w{schema_migrations}).each do |table_name|
ActiveRecord::Base.connection.execute "TRUNCATE TABLE #{table_name};"
end
Alternatively, run rake db:test:prepare before every run.
There is a gem too that you can use, but I don't have any experience with it: http://rubygems.org/gems/database_cleaner.
The reason the database is not resetting is that you are running your tests outside of the database transaction that rails provides. The ActiveSupport::TestCase class is the basis for all rails tests. ActiveRecord adds a per-test database transaction to this class. This transaction will reset the database after each test. But you aren't running your tests with ActiveSupport::TestCase, you are running your tests with Minitest::Spec which isn't configured to run the transaction.
The simplest solution is to add minitest-rails to your Gemfile, and change the require in your test_helper.rb file from minitest/autorun to minitest/rails. If you would rather add your own support for Minitest's spec DSL you can use this article as a starting point.
Do you have another factory that might be creating a podcast via an association?
FactoryGirl linting builds each factory and checks it's validity, and if another factory has a podcast as an association, it'll create a podcast record.
FactoryGirl recommends clearing the database after running the linting. They use database_cleaner in their example:
https://github.com/thoughtbot/factory_girl/blob/2bf15e45305ac03315cf2ac153db523d3ce89ce1/GETTING_STARTED.md#linting-factories
If you're using 'Rspec' to be your unit test framework. After the installation of gem 'rspec-rails', you will got one configuration file called: spec/rails_helper.rb and within it you will find one configuration which looks like this:
# If you're not using ActiveRecord, or you'd prefer not to run each of your
# examples within a transaction, remove the following line or assign false
# instead of true.
config.use_transactional_fixtures = true
this means that if it is true, then your each teat case will be running in a separate transaction.
Trying to use FactoryGirl to seed my development db with some data. Following this tutorial so my seeds.rb file looks like this:
require 'factory_girl'
Dir[Rails.root.join("spec/factories/*.rb")].each {|f| require f}
100.times do
FactoryGirl.create :idea
end
When running rake db:seed it complains:
rake aborted!
Factory already registered: idea
Why is it a bad thing that the factory is registered? I'm trying to use it, not register it (whatever that means...). Any idea what I'm doing wrong?
Most likely your factories are already being loaded. Try removing the line that does each of the requires and see if that corrects the issue.
I'm attempting to use the new standard way of loading seed data in Rails 2.3.4+, the db:seed rake task.
I'm loading constant data, which is required for my application to really function correctly.
What's the best way to get the db:seed task to run before the tests, so the data is pre-populated?
The db:seed rake task primarily just loads the db/seeds.rb script. Therefore just execute that file to load the data.
load "#{Rails.root}/db/seeds.rb"
# or
Rails.application.load_seed
Where to place that depends on what testing framework you are using and whether you want it to be loaded before every test or just once at the beginning. You could put it in a setup call or in a test_helper.rb file.
I'd say it should be
namespace :db do
namespace :test do
task :prepare => :environment do
Rake::Task["db:seed"].invoke
end
end
end
Because db:test:load is not executed if you have config.active_record.schema_format = :sql (db:test:clone_structure is)
Putting something like this in lib/tasks/test_seed.rake should invoke the seed task after db:test:load:
namespace :db do
namespace :test do
task :load => :environment do
Rake::Task["db:seed"].invoke
end
end
end
I believe Steve's comment above should be the correct answer. You can use Rails.application.load_seed to load seed data into your test envoironment. However, when and how often this data is loaded depends on a few things:
Using Minitest
There is no convenient way to run this file once before all tests (see this Github issue). You'll need to load the data once before each test, likely in the setup method of your test files:
# test/models/my_model_test.rb
class LevelTest < ActiveSupport::TestCase
def setup
Rails.application.load_seed
end
# tests here...
end
Using RSpec
Use RSpec's before(:all) method to load seed data for all test for this model:
describe MyModel do
before(:all) do
Rails.application.load_seed
end
describe "my model..." do
# your tests here
end
Hope this helps.
Building on Matt's answer, if taking that sort of route, I recommend calling Rails.application.load_seed in a before(:suite) block in rspec_helper.rb rather than in a before(:all) block in any file. That way the seeding code is invoked only once for the entire test suite rather than once for each group of tests.
rspec_helper.rb:
RSpec.configure do |config|
...
config.before(:suite) do
Rails.application.load_seed
end
...
end
We're invoking db:seed as a part of db:test:prepare, with:
Rake::Task["db:seed"].invoke
That way, the seed data is loaded once for the entire test run, and not once per test class.
Adding Rake::Task["db:seed"].invoke to the db:test:prepare rake task did not work for me. If I prepared the database with rake db:test:prepare, and then entered the console within the test environment, all my seeds were there. However, the seeds did not persist between my tests.
Adding load "#{Rails.root}/db/seeds.rb" to my setup method worked fine, though.
I would love to get these seeds to load automatically and persist, but I haven't found a way to do that yet!
For those using seedbank, it changes how seeds are loaded, so you probably can't/don't want to use the load ... solution provided here.
And just putting Rake::Task['db:seed'].invoke into test_helper resulted in:
Don't know how to build task 'db:seed' (RuntimeError)
But when we added load_tasks before that, it worked:
MyApp::Application.load_tasks
Rake::Task['db:seed'].invoke