Background: So I have roughly (Ruby on Rails app)
class A
def calculate_input_datetimes
# do stuff to calculate datetimes - then for each one identified
process_datetimes(my_datetime_start, my_datetime_end)
end
def process_datetimes(input_datetime_start, input_datetime_end)
# do stuff
end
end
So:
I want to test that calculate_input_datetimes algorithms are working
and calculating the correct datetimes to pass to process_datetimes
I know I can STUB out process_datetimes so that it's code won't be
involved in the test
QUESTION: How can I setup the rspec test however so I can specifically
test that the correct datestimes were attempted to be passed over to
process_datetimes, So for a given spec test that process_datetimes was
called three (3) times say with the following parameters passed:
2012-03-03T00:00:00+00:00, 2012-03-09T23:59:59+00:00
2012-03-10T00:00:00+00:00, 2012-03-16T23:59:59+00:00
2012-03-17T00:00:00+00:00, 2012-03-23T23:59:59+00:00
thanks
Sounds like you want should_receive and specifying what arguments are expected using with, for example
a.should_receive(:process_datetimes).with(date1,date2)
a.should_receive(:process_datetimes).with(date3,date4)
a.calculate_input_datetimes
There are more examples in the docs, for example you can use .ordered if the order of these calls is important
Related
I'm trying to test my rails application which using Stripe APIs, So I started with models, I'm using Rspec, The model which i want to test is called bank_account.rb inside it there is a function called (create_bank_account) with argument (bank_token) its pseudocode is something like this:
def create_bank_account(bank_token)
# make a Stripe request and save it in local variable
# save needed data in my bank_account table in my DB
end
when i started to test this function, I found that there is an API call inside it, Which is not good, I need my test not to depend on Internet, So after searching I found 'StripeMock` gem, It is useful and i started to use it with Rspec, but I found my self writing a test like this:
it 'with valid bank_token` do
# create a double for bank_account
# using StripeMock to get a faked response for creating
# new bank_account
# expect the doube to receive create_bank_account
# function and response with saving the data inside the DB
end
but after writing this I noticed that I didn't actually run create_bank_account function i faked it, So my questions are:
1- How can i test function that includes API request but run the function it self not faking it?
2- I read a lot about when we use doubles and stubs and what i understood is when a function is not completed, but if the functions is already implemented should i use doubles to avoid something like functions that call APIs?
First and foremost:
Do not create a double for bank_account.
Do not mock/stub bank_account.create_bank_account.
If you do either of these things, in a test that is supposed to be testing behaviour of BankAccount#create_bank_account, then your test is worthless.
(To prove this point, try writing broken code in the method. Your tests should obviously fail. But if you're mocking the method, everything will remain passing!!)
One way or another, you should only be mocking the stripe request, i.e. the behaviour at the boundary between your application and the internet.
I cannot provide a working code sample without a little more information, but broadly speaking you could refactor your code from this:
def create_bank_account(bank_token)
# make a Stripe request and save it in local variable
# save needed data in my bank_account table in my DB
end
To this:
def create_bank_account(bank_token)
stripe_request = make_stripe_request(bank_token)
# save needed data in my bank_account table in my DB
end
private
def make_stripe_request(bank_token)
# ...
end
...And then in your test, you can use StripeMock to only fake the response of BankAccount#make_stripe_request.
If the code is not so easy to refactor(?!), then stubbing the Stripe library directly like this might not be practical. An alternative approach you can always take is use a library like webmock to simply intercept all HTTP calls.
Using Mocha on Rails 4.2.
I'm testing a method that it should make a call to another method with the correct parameters. These parameters are ActiveRecord objects that it calls up from the database. Here is the key line in my test:
UserMailer.expects(:prompt_champion).with(users(:emma), [[language, 31.days.ago]]).once
Both users(:emma) and language are ActiveRecord objects.
Even though the correct call is made, the test fails because the parameters don't match the expectations. I think this might be because it's a different Ruby object each time a record is pulled up from the database.
I think one way around it is to see what method is being used in my code to pull up the records and stub that method to return mocks, but I don't want to do this because a whole bunch of Records are retrieved then filtered down to get to the right one, mocking all those records would make the test way too complex.
Is there a better way of doing this?
You could use block form of allow/expect.
expect(UserMailer).to receive(:prompt_champion) do |user, date|
expect(user.name).to eq "Emma"
expect(date).to eq 31.days.ago # or whatever
end
Sergio gave the best answer and I accepted it. I discovered the answer independently and found out along the way that I needed to return a mock from the ActionMailer method to make everything work properly.
I think it best to post here my complete test here for the sake of any other hapless adventurer to come this way. I'm using Minitest-Spec.
it 'prompts champions when there have been no edits for over a month' do
language.updated_at = 31.days.ago
language.champion = users(:emma)
language.save
mail = mock()
mail.stubs(:deliver_now).returns(true)
UserMailer.expects(:prompt_champion).with do |user, languages|
_(user.id).must_equal language.champion_id
_(languages.first.first.id).must_equal language.id
end.once.returns(mail)
Language.prompt_champions
end
You could use an RSpec custom matcher and compare expected values in that function.
It said here https://www.relishapp.com/rspec/rspec-core/v/3-5/docs/helper-methods/let-and-let what variable defined by let is changing across examples.
I've made the same simple test as in the docs but with the AR model:
RSpec.describe Contact, type: :model do
let(:contact) { FactoryGirl.create(:contact) }
it "cached in the same example" do
a = contact
b = contact
expect(a).to eq(b)
expect(Contact.count).to eq(1)
end
it "not cached across examples" do
a = contact
expect(Contact.count).to eq(2)
end
end
First example passed, but second failed (expected 2, got 1). So contacts table is empty again before second example, inspite of docs.
I was using let and was sure it have the same value in each it block, and my test prove it. So suppose I misunderstand docs. Please explain.
P.S. I use DatabaseCleaner
P.P.S I turn it off. Nothing changed.
EDIT
I turned off DatabaseCleaner and transational fixtures and test pass.
As I can understand (new to programming), let is evaluated once for each it block. If I have three examples each calling on contact variable, my test db will grow to three records at the end (I've tested and so it does).
And for right test behevior I should use DatabaseCleaner.
P.S. I use DatabaseCleaner
That's why your database is empty in the second example. Has nothing to do with let.
The behaviour you have shown is the correct behaviour. No example should be dependant on another example in setting up the correct environment! If you did rely on caching then you are just asking for trouble later down the line.
The example in that document is just trying to prove a point about caching using global variables - it's a completely different scenario to unit testing a Rails application - it is not good practice to be reliant on previous examples to having set something up.
Lets, for example, assume you then write 10 other tests that follow on from this, all of which rely on the fact that the previous examples have created objects. Then at some point in the future you delete one of those examples ... BOOM! every test after that will suddenly fail.
Each test should be able to be tested in isolation from any other test!
How do I test a method available only to an ActiveRecord relation proxy class in rspec? Like for example sum which would look something like #collection.sum(:attribute)
Here is what I'm trying to do:
#invoice = stub_model(Invoice)
#line_item = stub_model(LineItem, {quantity: 1, cost: 10.00, invoice: #invoice})
#invoice.stub(:line_items).and_return([#line_item])
#invoice.line_items.sum(:cost).should eq(10)
This doesn't work because #invoice.line_items returns a regular array that doesn't define sum in the same way as an ActiveRecord::Relation object does.
Any help is greatly appreciated.
I'm not sure which Rails you are on so I'll use Rails 4.0.x for this example; the principle still holds for Rails 3.x.
TL;DR: You don't want to take this route.
Consider not stubbing model specs
Consider adding domain specific APIs
You are rapidly heading down the road of over mocking/stubbing. I have been down this road, it does not lead to fun. Part of all of this comes down to violating the Law of Demeter. Part of it comes down to using the Rails APIs instead of creating your own domain APIs.
When you request an relation collection from an ActiveRecord model it does not return an Array as you are aware. In Rails 4.0.x, with a has_many association, the class which is returned is: ActiveRecord::Associations::CollectionProxy::ActiveRecord_Associations_CollectionProxy_Model.
Issue #1: Stubbing the wrong return value
Here your return type is an Array. While the actual return type is the ActiveRecord_Associations_CollectionProxy_Model. In stub/mock land, this isn't necessarily a bad thing. However, if you intend to use other calls on the object returned by the stub they need to match the same API contracts. Otherwise, you're not stubbing the same behavior.
In this case, the sum method defined on the AR association proxy actually executes SQL when it runs. The sum method defined on Array is patched in via Active Support. The Array#sum behavior is fundamentally different:
def sum(identity = 0, &block)
if block_given?
map(&block).sum(identity)
else
inject { |sum, element| sum + element } || identity
end
end
As you can see, it sums the elements, not the sum of the requested attribute.
Issue #2: Asserting on your stub'd object
The other main problem you have, is you are attempting to spec that you're stub returns what you stubbed. This doesn't make sense. The point of a stub is to return a canned answer. It's not to assert on how it behaves.
What you wrote isn't fundamentally different from:
invoice = stub_model(Invoice)
line_item = stub_model(LineItem, {quantity: 1, cost: 10.00, invoice: invoice})
invoice.stub(:line_items).and_return([line_item])
invoice.line_items.should eq([line_item])
Unless this is supposed to be a sanity check, it adds no real value to your specs.
Suggestions
I'm not sure what type of spec you are writing here. If this is a more traditional unit test or an acceptance test, then I probably wouldn't stub anything. There isn't necessarily anything wrong with hitting a database at times, especially when the thing you are testing is how you interact with it; which is really what you are doing here.
Another thing you can do is start to use this to create your own specific domain model APIs. All this really means is defining interfaces on objects that make sense for your domain, which may or may not be backed by a DB or other resource.
For example, take your invoice.line_items.sum(:cost).should eq(10), this is clearly testing the Rails AR API. In domain terms it means nothing really. However, invoice.subtotal probably means a lot more to your domain:
# app/models/invoice.rb
class Invoice < ActiveRecord::Base
def subtotal
line_items.sum(:cost)
end
end
# spec/models/invoice_spec.rb
# These are unit specs on the model, which directly works with the DB
# it probably doesn't make sense to stub things here
describe Invoice do
specify "the subtotal is the sum of all line item cost" do
invoice = create(:invoice)
3.times do |i|
cost = (i + 1) * 2
invoice.line_items.create(cost: cost)
end
expect(invoice.subtotal).to eq 12
end
end
Now later, when you use Invoice in some other part of your code, you can easily stub this if you need to:
# spec/helpers/invoice_helper_spec.rb
describe InvoiceHelper do
context "requesting the formatted subtotal" do
it "returns US dollars to two decimal places" do
invoice = double(Invoice, subtotal: 1012)
assign(:invoice, invoice)
expect(helper.subtotal_in_dollars).to eq "$10.12"
end
end
end
So when it is ok to stub model specs? Well, that's really a judgement call, and will vary from person to person, and code base to code base. However, just because something is in app/models doesn't mean it has to be an ActiveRecord model. In those cases, it's potentially fine to stub domain APIs on collaborators.
EDIT: create vs build
In the example above I used create(:invoice) and invoice.line_items.create(cost: cost). However, if you are concerned about DB slowness, you probably could just as easily use build(:invoice) and invoice.line_items.build(cost: cost).
Be aware that my use of create(:invoice) and build(:invoice) here is in reference to generic "factories", not a reference to a specific gem. You could simply use Model.create and Model.new in their place. Additionally, the line_items.create and line_items.build are provided by AR and have nothing to do with any factory gems.
I have a method that does something like this:
def some_method
chance = rand(4)
if chance == 1 do
# logic here
else
# another logic here
end
end
When I use RSpec to test this method, rand(4) inside it always generates 0. I am not testing rand method of Rails, I am testing my method.
What is a common practice to test my method?
There are two approaches I would consider:
Approach 1:
Use a known value of seed in srand( seed ) in a before :each block:
before :each do
srand(67809)
end
This works across Ruby versions, and gives you control in case you want to cover particular combinations. I use this approach a lot - thinking about it, that's because the code I was testing uses rand() primarily as a data source, and only secondarily (if at all) for branching. Also it gets called a lot, so exerting call-by-call control over returned values would be counter-productive, I would end up shovelling in lots of test data that "looked random", probably generating it in the first place by calling rand()!
You may wish to call your method multiple times in at least one test scenario to ensure you have reasonable coverage of combinations.
Approach 2:
If you have branch points due to values output from rand() and your assertions are of the type "if it chooses X, then Y should happen", then it is also reasonable in the same test suite to mock out rand( n ) with something that returns the values you want to make assertions about:
require 'mocha/setup'
Kernel.expects(:rand).with(4).returns(1)
# Now run your test of specific branch
In essence these are both "white box" test approaches, they both require you to know that your routine uses rand() internally.
A "black box" test is much harder - you would need to assert that behaviour is statistically OK, and you would also need to accept a very wide range of possibilities since valid random behaviour could cause phantom test failures.
I'd extract the random number generation:
def chance
rand(4)
end
def some_method
if chance == 1 do
# logic here
else
# another logic here
end
end
And stub it:
your_instance.stub(:chance) { 1 }
This doesn't tie your test to the implementation details of rand and if you ever decide to use another random number generator, your test doesn't break.
It seems that best idea is to use stub, instead of real rand. This way you would be able to test all values that you are interested in. As rand is defined in Kernel module you should stub it using:
Kernel.stub(:rand).with(anything) { randomized_value }
In particular contexts you can define randomized_value with let method.
I found that just stubbing rand ie. using Kernel.stub(:rand) as answered by Samuil did not initially work. My code to be tested called rand directly e.g
random_number = rand
However, if I changed the code to
random_number = Kernel.rand
then the stubbing worked.
This works in RSpec:
allow_any_instance_of(Object).to receive(:rand).and_return(1)