I have a Rails app with an Order and a Refund model. Order has_many :refunds. All well and good. I'm trying to write a functional test for refund logic in a controller. Here's what I have right now:
test "should not process partial paypal refund for partially refunded order when refund total plus refund amount is greater than order total" do
set_super_admin_login_credentials
o = Order.new
o.stubs({:id => 1234567, :source => "PayPal", :total => 39.95, :user => users(:dave)})
Order.stubs(:find).with(1234567).returns(o)
get :refund, {:order_id => 1234567}
assert_equal o, assigns(:order)
o.refunds.build.stubs({:amount => 1.0})
o.refunds.build.stubs({:amount => 30.00})
assert_raise do
post :refund, {:order_id => 1234567, :refund_amount => 10.00}
end
end
And in the controller, the refund method looks like this:
def refund
#order = Order.find(params[:order_id])
return if request.get?
amount = params[:refund_amount].to_f
raise "Cannot refund order for more than total" if (#order.refunds.sum(&:amount) + amount)
# Do refund stuff
end
Some notes:
I'm basing the o.refunds.build bit on Ryan Bates' Railscast. If this is not right or no longer relevant, that's helpful information.
I've seen a lot of conflicting information about how to actually do the sum method, some with the & and some without. In script/console, the & blows up but without it, I get an actual sum. In my controller, however, if I switch from &:amount to :amount, I get this message: NoMethodError: undefined method+' for :amount:Symbol`
I feel like there's some conceptual information missing rather than a bug somewhere, so I'll appreciate some pointers.
Finally figured out the issue. I was stubbing an empty association as [] rather than leaving it nil for Rails to handle on some other methods. So, when I would change one, the other would fail. Word to the wise: Enumerable#sum and ActiveRecord::Associations::AssociationCollection#sum take entirely different parameters. :)
So, by changing the stubs to leave off :refunds => [] and using a string for the field name in sum I got things back to normal. So, here's the functional version of the above code:
test "should not process partial paypal refund for partially refunded order when refund total plus refund amount is greater than order total" do
set_super_admin_login_credentials
o = Order.new
o.stubs({:id => 1234567, :source => "PayPal", :total => 39.95, :user => users(:dave)})
Order.stubs(:find).with(1234567).returns(o)
get :refund, {:order_id => 1234567}
assert_equal o, assigns(:order)
o.refunds.build.stubs({:amount => 1.0})
o.refunds.build.stubs({:amount => 30.00})
assert_raise do
post :refund, {:order_id => 1234567, :refund_amount => 10.00}
end
end
def refund
#order = Order.find(params[:order_id])
return if request.get?
amount = params[:refund_amount].to_f
raise "Cannot refund order for more than total" if (#order.refunds.sum('amount') + amount)
# Do refund stuff
end
Related
I want to create subscription plans in stripe programatically, and I want to be able to run this many times so if the plans exist it should just ignore it.
I noticed if I try and retrieve a plan it throws an exception if it doesn't exist:
plan1 = Strie::Plan.retrieve("abcd123")
>>Stripe::InvalidRequestError: No such plan: abcd123
I create a plan using:
Stripe::Plan.create(.....)
I have a plans model that has all my plans, so I ideally want to do this:
Plan.all.each do |plan|
# create stripe plan here if it doesn't exist
end
What is the best way to handle this exception if the plan exists in stripe already?
Take a look at https://stripe.com/docs/api#error_handling
If you attempt to create a plan with an id that already exists, the request will fail and Stripe will throw an invalid request error. You should be able to wrap your plan creation call to account for errors. A barebones example:
require "stripe"
Stripe.api_key = "sk_test_xxxyyyzzz"
MyPlans.each do |plan|
# try to create a plan
begin
my_plan = Stripe::Plan.create(
:amount => plan.amount,
:interval => "month",
:name => plan.name,
:currency => "usd",
:id => plan.id
)
puts my_plan
# catch any invalid request errors
rescue Stripe::InvalidRequestError => e
puts e.json_body[:error]
end
end
I have a weird issue with the PayPal express checkout. When I pass in an options[:item][:quantity], I get an error code from PayPal that the transaction in invalid.
#controllers/orders_controller.rb
def express
options = {
:ip => request.remote_ip,
:return_url => new_order_url,
:cancel_return_url => products_url,
:items => current_cart.line_items_hash
}
response = EXPRESS_GATEWAY.setup_purchase(current_cart.build_order.price_in_cents,
options)
redirect_to EXPRESS_GATEWAY.redirect_url_for(response.token)
end
# models/cart.rb
def line_items_hash
self.line_items.map do |line_item|
{
:name => Product.find(line_item.product_id).name
# :quantity => line_item.quantity,
:description => line_item.product.description,
:amount => line_item.gross_price_in_cents}
end
end
So this works fine. The problem is that the correct quantity is not shown on the PayPal order confirmation page. However, if I uncomment the quantity variable in the line_items_hash function the whole thing breaks and I get "invalid transaction from paypal". Any ideas?
Silly old me. Paypal keeps invalidating my transaction because I'm passing bad information. I was setting the amount to line_item total which is already the total of quantity * product.price. So the information that Paypal was receiving is quantity * (quantity * product.price).
Silly little mistake but I managed to catch it in the end.
# models/cart.rb
def line_items_hash
self.line_items.map do |line_item|
{
:name => Product.find(line_item.product_id).name
:quantity => line_item.quantity,
:description => line_item.product.description,
:amount => line_item.price.in_cents}
end
end
I would like this before filter to run every time the page is loaded (for now) to check if an item is over 7 days old or not and if so, run some actions on it to update its attributes.
I have before_filter :update_it in the application controller. update_it is defined below that in the same controller as:
def update_it
#books = Book.all
#books.each do |book|
book.update_queue
end
end
Then update_queue is defined in the book model. Here's everything in the model that pertains to this:
scope :my_books, lambda {|user_id|
{:conditions => {:user_id => user_id}}
}
scope :reading_books, lambda {
{:conditions => {:reading => 1}}
}
scope :latest_first, lambda {
{:order => "created_at DESC"}
}
def move_from_queue_to_reading
self.update_attributes(:queued => false, :reading => 1);
end
def move_from_reading_to_list
self.update_attributes(:reading => 0);
end
def update_queue
days_gone = (Date.today - Date.parse(Book.where(:reading => 1).last.created_at.to_s)).to_i
# If been 7 days since last 'currently reading' book created
if days_gone >= 7
# If there's a queued book, move it to 'currently reading'
if Book.my_books(user_id).where(:queued => true)
new_book = Book.my_books(user_id).latest_first.where(:queued => true).last
new_book.move_from_queue_to_reading
currently_reading = Book.my_books(user_id).reading_books.last
currently_reading.move_from_reading_to_list
# Otherwise, create a new one
else
Book.my_books(user_id).create(:title => "Sample book", :reading => 1)
end
end
end
My relationship is that a book belongs_to a user and a user has_many books. I'm showing these books in the view through the user show view, not that it matters though.
So the errors I keep getting are that move_from_queue_to_reading and move_from_reading_to_list are undefined methods. How can this be? I'm clearly defining them and then calling them below. I really am at a loss and would greatly appreciate some insight into what I'm doing wrong. I'm a beginner here, so any structured criticism would be great :)
EDIT
The exact error message I get and stack trace is as follows:
NoMethodError in UsersController#show
undefined method `move_from_queue_to_reading' for nil:NilClass
app/models/book.rb:41:in `update_queue'
app/controllers/application_controller.rb:22:in `block in update_it'
app/controllers/application_controller.rb:21:in `each'
app/controllers/application_controller.rb:21:in `update_it'
I suspect that the collection returned is an empty array (which is still 'truthy' when tested). So calling .last is returning nil to the new_book and currently_reading local variables. Try changing:
if Book.my_books(user_id).where(:queued => true)
to:
if Book.my_books(user_id).where(:queued => true).exists?
Additionally, you are modifying the scope when finding currently_reading. This can potentially cause the query to again return no results. Change:
currently_reading.move_from_reading_to_list
to:
currently_reading.move_from_reading_to_list if currently_reading
This is more of a style question, I'm wondering what other people do.
Let's say I have a field in my database called "status" for a blog post. And I want it to have several possible values, like "draft", "awaiting review", and "posted", just as an example.
Obviously we don't want to "hard code" in these magic values each time, that wouldn't be DRY.
So what I sometimes do is something like this:
class Post
STATUS = {
:draft => "draft",
:awaiting_review => "awaiting review",
:posted => "posted"
}
...
end
Then I can write code referring to it later as STATUS[:draft] or Post::STATUS[:draft] etc.
This works ok, but there are a few things I don't like about it.
If you have a typo and call something like STATUS[:something_that_does_not_exist] it won't throw an error, it just returns nil, and may end up setting this in the database, etc before you ever notice a bug
It doesn't look clean or ruby-ish to write stuff like if some_var == Post::STATUS[:draft] ...
I dunno, something tells me there is a better way, but just wanted to see what other people do. Thanks!
You can use Hash.new and give it a block argument which is called if a key is unknown.
class Post
STATUS = Hash.new{ |hash, key| raise( "Key #{ key } is unknown" )}.update(
:draft => "draft",
:awaiting_review => "awaiting review",
:posted => "posted" )
end
It's a bit messy but it works.
irb(main):007:0> Post::STATUS[ :draft ]
=> "draft"
irb(main):008:0> Post::STATUS[ :bogus ]
RuntimeError: Key bogus is unknown
from (irb):2
from (irb):8:in `call'
from (irb):8:in `default'
from (irb):8:in `[]'
from (irb):8
This is a common problem. Consider something like this:
class Post < ActiveRecord::Base
validates_inclusion_of :status, :in => [:draft, :awaiting_review, :posted]
def status
read_attribute(:status).to_sym
end
def status= (value)
write_attribute(:status, value.to_s)
end
end
You can use a third-party ActiveRecord plugin called symbolize to make this even easier:
class Post < ActiveRecord::Base
symbolize :status
end
You could use a class method to raise an exception on a missing key:
class Post
def self.status(key)
statuses = {
:draft => "draft",
:awaiting_review => "awaiting review",
:posted => "posted"
}
raise StatusError unless statuses.has_key?(key)
statuses[key]
end
end
class StatusError < StandardError; end
Potentially, you could also use this method to store the statuses as integers in the database by changing your strings to integers (in the hash), converting your column types, and adding a getter and a setter.
I do it like this:
class Post
DRAFT = "draft"
AWAITING_REPLY = "awaiting reply"
POSTED = "posted"
STATUSES = [DRAFT, AWAITING_REPLY, POSTED]
validates_inclusion_of :status, :in => STATUSES
...
end
This way you get errors if you misspell one. If I have multiple sets of constants, I might do something like DRAFT_STATUS to distinguish.
Take a look at the attribute_mapper gem.
There's a related article that shows how you can handle the problem declaratively, like this (borrowed from the article):
class Post < ActiveRecord::Base
include AttributeMapper
map_attribute :status, :to => {
:draft => 1,
:reviewed => 2,
:published => 3
}
end
...which looks rather stylish.
Even though this is an old post, for somebody stumbling across this, you can use the fetch method on Hash, which raises an error (when no default is passed) if the given key is not found.
STATUS = {
:draft => "draft",
:awaiting_review => "awaiting review",
:posted => "posted"
}
STATUS.fetch(:draft) #=> "draft"
STATUS.fetch(:invalid_key) #=> KeyError: key not found: invalid_key
I have a user model in which I have a method for seeing if the user has earned a "badge"
def check_if_badges_earned(user)
if user.recipes.count > 10
award_badge(1)
end
If they have earned a badge, the the award_badge method runs and gives the user the associated badge. Can I do something like this?
def check_if_badges_earned(user)
if user.recipes.count > 10
flash.now[:notice] = "you got a badge!"
award_badge(1)
end
Bonus Question! (lame, I know)
Where would the best place for me to keep all of these "conditions" for which my users could earn badges, similar to stackoverflows badges I suppose. I mean in terms of architecture, I already have badge and badgings models.
How can I organize the conditions in which they are earned? some of them are vary complex, like the user has logged in 100 times without commenting once. etc. so there doesn’t seem to be a simple place to put this sort of logic since it spans pretty much every model.
I'm sorry for you but the flash hash is not accessible in models, it gets created when the request is handled in your controller. You still can use implement your method storing the badge infos (flash message included) in a badge object that belongs to your users:
class Badge
# columns:
# t.string :name
# seed datas:
# Badge.create(:name => "Recipeador", :description => "Posted 10 recipes")
# Badge.create(:name => "Answering Machine", :description => "Answered 1k questions")
end
class User
#...
has_many :badges
def earn_badges
awards = []
awards << earn(Badge.find(:conditions => { :name => "Recipeador" })) if user.recipes.count > 10
awards << earn(Badge.find(:conditions => { :name => "Answering Machine" })) if user.answers.valids.count > 1000 # an example
# I would also change the finds with some id (constant) for speedup
awards
end
end
then:
class YourController
def your_action
#user = User.find(# the way you like)...
flash[:notice] = "You earned these badges: "+ #user.earn_badges.map(:&name).join(", ")
#...
end
end