Rails: Where do I put my API methods? - ruby-on-rails

I'm very new to Rails, and I'm a little overwhelmed where I do simple things like create an API call. I've set up a route at /reports which has this controller:
class ReportsController < ApplicationController
#client = # Api-accessing gem
#all_reports = []
def self.request_report
begin
puts "Step 1:"
step1 = #client.request_report(opts = {"max_count" => 1})
step1_result = step1.parse
puts "Done!"
puts step1_result
rescue Excon::Errors::ServiceUnavailable => e
puts "Didn't work"
logger.warn e.response.message
retry
end
end # End request_report
request_report
end
This correctly calls the external API when I first load the /reports route, but when I refresh the page the code isn't re-run.
Perhaps I'm misunderstanding what controllers are used for? Am I meant to be putting this code somewhere else? Or is there a caching issue?

The only public API of controller are the actions which respond to a HTTP request. In your case get "/reports" => "reports#request_report" is a route which corresponds to the action request_report.
However actions are instance methods, not class methods:
class ReportsController
def request_report # self.request_report would make this a class method!
# #todo get reports from somewhere and
# return some sort of response.
end
# any method call here happens when the class is evaluated.
end
You are declaring the action as a class method and then calling it when the ReportsController class is evaluated. Sorry to say but just about everything about your controller is wrong.
The Rails convention would be to call the action index.
Controllers in Rails should only be instantiated by the router (or your test framework). So they are definatly the wrong place to put resuable bits and bobs. If you ever see someone doing ReportsController.new.foo or ReportsController.foo - fire them on the spot.
So where do you put external API calls?
If its a pretty trivial one-off you can place it in private method in your controller.
Some place API calls on the model layer - however that is debatable since ActiveRecord models already are supercharged to the gills with powers and responsibilities.
One solution that has worked well for me is Service Objects. They are easy to test and have a clear single responsibility.
class RequestReportService
def initalize(client)
#client = client
end
def call(opts = {})
begin
return #client.request_report(opts.merge("max_count" => 1))
rescue Excon::Errors::ServiceUnavailable => e
nil
end
end
end
class ReportsController
def index
#reports = RequestReportService.new(#client).call
end
end

To add to #max's excellent answer, you need to appreciate that Rails is based on a stateless protocol (HTTP)...
each request message can [only] be understood in isolation.
This means that if you want to create a set of controller actions, you have to appreciate that each call is going to create a new instance of your classes etc. This, coupled with the idea of a RESTful set of actions, should give you a basis from which to build your functionality.
--
#config/routes
scope constraints: { subdomain: "api" } do
resources :reports #-> http://api.url.com/reports
end
#app/controllers/reports_controller.rb
class ReportsController < ApplicationController
respond_to :json #-> requires "responders" gem
def index #-> instance method
#reports = Report.all
respond_with #reports #-> all reports
end
def show
#report = Report.find params[:id]
respond_with #report
end
end
I'll leave the service object stuff as I have no experience with it.
--
If you're pulling from an external API, you have several considerations:
Calls ideally need to be asynchronous (unless you use multi-threading)
Calls need to be made in the instance method
Your current pattern calls the API on the class, which is why you can't refresh it:
class ReportsController < ApplicationController
#client = # Api-accessing gem
#client is only invoked (I don't know why it works, as it should be a class variable) with the class.
So if you send a new request (which creates an instance of ReportsController), #client is going to be declared that one time.
To get it working correctly, #client needs to be defined with each instance method:
class ReportsController < ApplicationController
def index
#client = # Api-accessing gem
This way, each time you invoke ReportsController#index, a new API call will be made. Might seem trivial, but the data scope is massive.
Finally, you need to read up about MVC (Model View Controller):
This will show you how controllers are meant to be used in Rails applications etc.

Well I actually never seen anyone code like this in a rails controller. Rails is a mvp framework. Controller are use to negotiate between your model and the views. First of all, if you routed correctly to your controller like
get "/reports" => "request_report#reports"
your controller should have a method like the following
def request_report
#client = Client.find(params[:id])
end
And then the controller will render and display the view in your app/views/reports/request_report.html.erb with access to the #client variable you just search from your database.
I am not sure why you are calling the block request_report at the bottom of the page, it just doesn't make sense in a controller. And you certainly don't really need to write self in front of a controller method.
def self.request_report
your code
end
As for where to put your api controller, usually for an api controller, we can create new folders under controllers, so the structure will be like
app/controllers/api/v1/your_api_controller.rb
Then in your_api_controller.rb you will need to add namespace infront of your controller like this.
class Api::V1::ReportsController < ActionController::Base
end
It is the same with your routes, you will add namespace in your route.rb
namespace :api do
namespace :v1 do
get "/reports" => "request_report#reports"
end
end

Related

Required certain params on certain actions

I have a requirement to need to validate presence of some params in certain situations. Here is the example of that :
In my user controller, for update action, I am required to validate the presence of these params. Same deal for car controller, update action as well, you could see recurring theme here. Params are additional_info.
My base controller provides additional_info_params which pulls the right data from the request.
Here is what I tried so far. I created a AR controller concern and included it in the controller, here is some code:
module ClassMethods
def require_additional_info_for(*methods)
binding.pry
return unless methods.include?(action_name)
if additional_info_params.empty?
head 400
end
end
end
My idea was to be able to define methods that require these params on the top of controller file, just like before_action from rails or skip_authorization_check from cancan. Like so:
MyController < BaseController
include Concerns::AdditionalInformation
require_additional_info_for :update
def update
...
end
end
This code above however does not work as I intended, mainly because this fires on the request class without much knowledge about the request (where I need to derive action name from via action_name).
So how can I do something like this?
Yes, you can, but i suggest you to use the before_action callback!
In a 'abstract' controller, register your method like this:
class SameController < ApplicationController
...
protected
def require_additional_params
render status: :unprocessable_entity if additional_info_params.empty?
end
end
After this, all the controllers who will use this methods, must extends SameController, and runs before_action passing the above method for the wanted actions, for example:
class UserController < SameController
before_action :require_additional_params, only: [:action1, :action2]
end
Note: You can put the require_additional_params in a module and include in your controller, or just put it in the ApplicationController
You might also look at making these regular strong params in the respective controller. It looks something like this:
def update_params
params.require(:car).permit(:engine, :wheels, :rims).tap do |car_params|
car_params.require(:engine)
end
end
This would expect a top-level :car key params (which it strips), and require an :engine param, but allow the other 2 (:wheels and :rims). If :engine isn't present, it will raise a ActionController::ParameterMissing (just like if :cars was missing)
This is straight from the action controller strong params docs (last example at bottom)
I'll sometimes throw these into separate private methods on the respective controller, so there would also possibly be a create_params method with different requirements. I prefer this method over using a custom method as a before_action.

Encapsulating controller logic in rails

I have 2 controllers in rails with different authentications schemes,
but they do almost the same.
What is the best way in rails to encapsulate
the logic of a controller in another class or helper?
Sample:
def ControllerA < BasicAuthController
def create
blablacode
end
end
def ControllerB < TokenAuthController
def create
blablacode
end
end
Whats the proper way to do this? create a model with the code?
Create a helper? other?
The simplest thing is to make a module and then include it into the other controllers:
module ControllerMixin
def create
blablacode
end
end
The remaining question, though, is where to put this code such that it is works with Rails autoloader, since it needs to be loaded before the controllers. One way to do it would be to write the module to a file in the lib/ directory, then add that to the autoload paths (see auto-loading-lib-files-in-rails-4
Why don't you enable both schemes for a single controller? Especially if the only difference is Authentication. You could have two app/controllers/concerns to encapsulate both authentication methods and include Auth1 and include Auth2 for a single controller who is only responsible for whatever resource it manages.
Otherwise, services are the best approach to encapsulate controller logic.
Create a folder called services in your app folder and write PORO classes here. Say you have a few places in your app where you want to pay for stuff via make Stripe.
# app/services/stripe_service.rb
module StripeService
def customer(args)
...
end
def pay(amount, customer)
...
end
def reverse(stripe_txn_id)
...
end
end
# controller
StripeService.customer(data)
=> <#Stripe::Customer>
Or if you only need to do one thing.
# app/services/some_thing.rb
module SomeThing
def call
# do stuff
end
end
# controller
SomeThing.call
=> # w/e
If you need an object with multiple reponsibilities you could create a class instead.
class ReportingService
def initialize(args)
...
end
def query
...
end
def data
...
end
def to_json
...
end
end
https://blog.engineyard.com/2014/keeping-your-rails-controllers-dry-with-services
I do it something like this:
#app/services/my_app/services/authentication.rb
class MyApp::Services::Authentication
class < self
def call(params={})
new(params).call
end
end # Class Methods
#==============================================================================================
# Instance Methods
#==============================================================================================
def initialize(params)
#params = params
end
def call
... do a lot of clever stuff
... end by returning true or false
end
private
def params() #params end
end
Then:
class FooController < ApplicationController
before_action :authenticate
def authenticate
redirect_to 'some_path' unless MyApp::Services::Authenticate.call(with: 'some_params')
end
end
Short answer, i choose to create a Helper.
From all the suggestions in the answers
Create a Module:
Seems correct but it didnt feel right to have logic outside
the app directory. This wasnt an external module or library but
something very related to the logic of my app.
Integrate diferents authentications in one controller:
Was a good suggestion but i have to change all the logic of my app.
Create a Helpers:
It seems to me the better solution, i had the code on a helper, and
is inside the app directory, very near from the other logic.

Rails Testing controller methods with instance variables set in before filters

Ok, so my main issue is I have implemented Mailboxer into our project to handle messaging and I am trying to write tests for it. However, I keep stumbling over and over again. I have attempted several different stub/mocks but have made no progress.
We have a conversations_controller.rb that relies on before_filters for setting all the instance variables necessary for doing each action. Then in the controller actions, the instance variables are referenced directly to do any sort of action or to return specific data.
Here is an example of our index action which returns all conversations in the "box" that is specified in the before_filter, of the mailbox also specified in another before_filter:
class ConversationsController < ::ApplicationController
before_filter :get_user_mailbox, only: [:index, :new_message, :show_message, :mark_as_read, :mark_as_unread, :create_message, :reply_message, :update, :destroy_message, :untrash]
before_filter :get_box
def index
if #box.eql? "inbox"
#conversations = #mailbox.inbox
elsif #box.eql? "sentbox"
#conversations = #mailbox.sentbox
else
#conversations = #mailbox.trash
end
end
And before filters:
private
def get_user_mailbox
#user = User.where(:user_name => user.user_name.downcase).where(:email => user.email.downcase).first_or_create
#mailbox = #user.mailbox if #user
end
def get_box
if params[:box].blank? or !["inbox","sentbox","trash"].include?params[:box]
params[:box] = 'inbox'
end
#box = params[:box]
end
So I guess I have 2 questions in one. First, how to I get my tests to generate the correct data #mailbox, #user, and #box that is needed for the index action. Next, how do I pass the fake parameter to set #box to different "inbox/sentbox/trash". I have tried controller.index({box: "inbox"}) but always get "wrong arguments 1 for 0" messages.
I have tried the following in various different ways, but always get nil:class errors which means that my instance variables are definitely not being set properly.
describe "GET 'index' returns correct mailbox box" do
before :each do
#user = User.where(:user_name => 'test').where(:email => 'test#test.com').first_or_create
#mailbox = #user.mailbox
end
it "#index returns inbox when box = 'inbox'" do
mock_model User
User.stub_chain(:where, :where).and_return(#user)
controller.index.should == #mailbox.inbox
end
end
Filters and callbacks are hard to test and debug. Avoid them when possible.
In this case I don't think your before_filter is necessary, thus no need to test it. The better home for the methods is model.
Check my refacoring:
class User < ActiveRecord::Base
delegate :inbox, :sentbox, :trash, to: :mailbox
end
class ConversationsController < ::ApplicationController
def index
#conversations = current_user.send get_box
end
private
def get_box
# your code
end
end
That's all. Should be enough.
You can then test regularly.
First of all, read the oficial documentation for rails testing: using data for testing and passing parameters to controlers is explained there.
To generate data for your tests you can:
fill your test database with some mailbox and users using rails fixtures or use something like factory girl
use mock objects to fake data. I personally use the mocha gem but there are others
I tend to use a combination of both, prefering mock objects when possible and falling back to factory girl when mocking needs too much code.

RoR: instances variables within controller methods

My question is about controller methods (possibly included from an outside class) that work with instance variables. I frequently use a before_filter in controllers to set up certain variables, e.g.:
class DocumentController < ApplicationController
before_filter :fetch_document
def action
#document.do_something
end
private
def fetch_document
#document = Document.find(params[:id])
end
end
I've been working on a project in which a few controllers will share some functionality, say, document editing. My first thought was to extract the relevant methods, and get them from application_controller.rb or a separate module. But then I noticed I was writing code that looks like this:
def fetch_document
#document = Document.find(params[:id])
end
def do_something_to_document
#document.do_something
end
This sets off warning bells: do_something_to_document is essentially assuming the existence of #document, rather than taking it as an argument. Is this, in your sage opinions, a bad coding practice? Or am I being paranoid?
Assuming it is an issue, I see two general approaches to deal with it:
Check for the instance var and bail unless it's set:
def do_something_to_document
raise "no doc!" unless #document
[...]
end
Call the action with the instance var as an argument:
def do_something_to_document(document)
[...]
end
2 looks better, because it hides the context of the calling object. But do_something_to_doc will only be called by controllers that have already set up #document, and taking #document as a method argument incurs the overhead of object creation. (Right?) 1 seems hackish, but should cover all of the cases.
I'm inclined to go with 1 (assuming I'm right about the performance issue), even though seeing a list of methods referencing mysterious instance vars gives me hives. Thoughts? Let me know if I can be more clear. (And of course, if this is answered somewhere I didn't see it, just point me in the right direction...)
Thanks,
-Erik
If you really need document in different controllers, I'd do something like this:
class ApplicationController < ActionController::Base
private
def document
#document ||= Document.find(params[:document_id])
end
end
class FooController < ApplicationController
before_filter :ensure_document, :only => [:foo]
def foo
document.do_something
end
private
# TODO: not sure if controller_name/action_name still exists
def ensure_document
raise "#{controller_name}##{action_name} needs a document" unless document
end
end
As #variable are session/instance variable you will get a Nil exception in do_something_to_document method.
The first code is fine, because before_filter will always load your #document.
I suggest you to write something like that
def fetch_document(doc_id)
#document ||= Document.find(doc_id)
end
def do_something_to_document
my_doc = fetch_document(params[:id])
end
where do_something_to_document is in the controller (if not, dont use params[:id], even if you know you can access this global, use another explicit parameter). The ||= thing, will asssure that you call the base only once by request.

Rails: call another controller action from a controller

I need to call the create action in controller A, from controller B.
The reason is that I need to redirect differently when I'm calling from controller B.
Can it be done in Rails?
To use one controller from another, do this:
def action_that_calls_one_from_another_controller
controller_you_want = ControllerYouWant.new
controller_you_want.request = request
controller_you_want.response = response
controller_you_want.action_you_want
end
You can use a redirect to that action :
redirect_to your_controller_action_url
More on : Rails Guide
To just render the new action :
redirect_to your_controller_action_url and return
The logic you present is not MVC, then not Rails, compatible.
A controller renders a view or redirect
A method executes code
From these considerations, I advise you to create methods in your controller and call them from your action.
Example:
def index
get_variable
end
private
def get_variable
#var = Var.all
end
That said you can do exactly the same through different controllers and summon a method from controller A while you are in controller B.
Vocabulary is extremely important that's why I insist much.
You can use url_for to get the URL for a controller and action and then use redirect_to to go to that URL.
redirect_to url_for(:controller => :controller_name, :action => :action_name)
This is bad practice to call another controller action.
You should
duplicate this action in your controller B, or
wrap it as a model method, that will be shared to all controllers, or
you can extend this action in controller A.
My opinion:
First approach is not DRY but it is still better than calling for another action.
Second approach is good and flexible.
Third approach is what I used to do often. So I'll show little example.
def create
#my_obj = MyModel.new(params[:my_model])
if #my_obj.save
redirect_to params[:redirect_to] || some_default_path
end
end
So you can send to this action redirect_to param, which can be any path you want.
Perhaps the logic could be extracted into a helper? helpers are available to all classes and don't transfer control. You could check within it, perhaps for controller name, to see how it was called.
Composition to the rescue!
Given the reason, rather than invoking actions across controllers one should design controllers to seperate shared and custom parts of the code. This will help to avoid both - code duplication and breaking MVC pattern.
Although that can be done in a number of ways, using concerns (composition) is a good practice.
# controllers/a_controller.rb
class AController < ApplicationController
include Createable
private def redirect_url
'one/url'
end
end
# controllers/b_controller.rb
class BController < ApplicationController
include Createable
private def redirect_url
'another/url'
end
end
# controllers/concerns/createable.rb
module Createable
def create
do_usefull_things
redirect_to redirect_url
end
end
Hope that helps.
You can call another action inside a action as follows:
redirect_to action: 'action_name'
class MyController < ApplicationController
def action1
redirect_to action: 'action2'
end
def action2
end
end
Separate these functions from controllers and put them into model file. Then include the model file in your controller.

Resources