I have two classes; customer and reservation. And my project is consist of only ruby code, not rails project.
Class reservation reads bulk json file line by line which includes customer hash.
From that hash, I create customer object.
Here's the code within the reservation class.
def parse_json
File.open(#filename, "r" ).each do |line|
#customers << Customer.new(JSON.parse(line))
end
return #customers
end
And in customer.rb I have the following;
validates_presence_of :hash
validate :hash_should_include_all_fields
attr_reader :name, :userid, :latitude, :longitude, :distance
def hash_should_include_all_fields
puts "I'm here #{hash}"
if hash.assert_valid_keys('user_id', 'name', 'latitude', 'longitude')
puts "Valid"
else
puts "Not valid"
end
end
However as I create customer objects, hash_should_include_all_fields method is not called.
What do I miss in here, that would be great if you can help.
Thanks
You're calling new which will build a Customer object but do nothing more.
Calling create! or create will also run validations and write the Object to the database if it's an active_record model.
If you don't want to write it to the databas you can call valid? on the instance created with new
Related
I'm using ActiveRecord 4.1.8 in a Ruby (not Rails) application. I have a table and a corresponding model that looks like the following:
create_table 'people', :force => true do |t|
t.string 'name'
end
class Person < ActiveRecord::Base
def name=(name)
puts "Attribute setter for name called with #{name}"
write_attribute(:name, name)
end
end
When I create a new instance of Person, I see the Attribute setter for name called with... written to STDOUT. However, when I reload the model instance, I do not see the message written to STDOUT.
p = Person.create(name: 'foobar')
--> Attribute setter for name called with foobar
p.reload
--> <nothing>
The model is getting persisted to the database, so this makes me think name= isn't used when data is loaded into a model from the database. I need to modify certain data attributes when they're read in from the database, so does anyone know what other method I need to override?
From the active_record/persistence.rb source:
def reload(options = nil)
clear_aggregation_cache
clear_association_cache
fresh_object =
if options && options[:lock]
self.class.unscoped { self.class.lock(options[:lock]).find(id) }
else
self.class.unscoped { self.class.find(id) }
end
#attributes = fresh_object.instance_variable_get('#attributes')
#new_record = false
self
end
It just replaces the attributes hash directly. Seems like the easiest way to handle this is to override reload and patch things up after its called.
I wrote a very simple User class. The instance variable email has a reader accessor and my own writer accessor that validates the email address with a regex.
class User
attr_reader :email
def email=(value)
if (value =~ /^[a-z\d\-\_\+\.]+#([a-z\d\-]+\.)+[a-z]+$/)
#email = value
else
# bonus question: is ArgumentError the right error type to use here?
raise ArgumentError, "#{value} is not a valid email address."
end
end
end
I wrote the following test:
require 'test/unit'
require_relative '../lib/user'
class TC_UserTest < Test::Unit::TestCase
def setup
#user = User.new()
end
def test_email
# using the writer accessor
#user.email = 'user#example.com'
# bypassing the writer accessor. evil.
#user.email[4] = '#'
assert_equal('user#example.com', #user.email)
end
end
By using the reference given to me by the reader accessor, I am able to manipulate the email instance variable without going through the writer accessor.
The same principe would apply to any data type that allows manipulation without outright assigning a new value with =
Am I being overzealous? I just want to write robust code. Is there a way to ensure that my email address can only be set using the writer accessor?
I'm new to the language and I'm trying to get a feel for the best practices.
An option to make the test pass (and protect the #email variable) is to expose a duplicate.
def email
#email.dup
end
To do what you're trying to do, my advice is to move the regexp into its own validation method.
Better still, don't write an email regexp unless you really want to do it right.
Use a gem instead: https://github.com/SixArm/sixarm_ruby_email_address_validation
After you set the email, freeze it with http://ruby-doc.org/core-1.9.3/Object.html#method-i-freeze
Bonus answer: yes, ArgumentError is the right error type in general. If you're using Rails, consider using the Rails validation methods.
You can freeze value in writer, that way you'll be able to assign new one via writer, but already assigned would be immutable:
class User
attr_reader :email
def email=(value)
if (value =~ /^[a-z\d\-\_\+\.]+#([a-z\d\-]+\.)+[a-z]+$/)
# make email immutable:
#email = value.freeze
else
# bonus question: is ArgumentError the right error type to use here?
raise ArgumentError, "#{value} is not a valid email address."
end
end
end
I'm having an issue with passing the generated JSON notation of my object to my Sinatra application. The problem I have is twofold:
I have 2 classes that are mapped to a database using the Sequel gem. When they generate JSON it is ok and properly implemented.
I have a custom class called registration that maps one of the classes with an additional field. The goal is to generate JSON out of this and pass that JSON to the application using cucumber (test purpose)
The application code responsible for handling the request has the following function defined:
post '/users' do
begin
hash = JSON.parse(self.request.body.read)
registration = Registration.new.from_json(#request.body.read)
registration.user.country = Database::Alaplaya.get_country_by_iso_code(registration.user.country.iso_code)
return 400 unless(registration.is_valid?)
id = Database::Alaplaya.create_user(registration.user)
# If the registration failed in our system, return a page 400.
return 400 if id < 1
end
problem 1: I cannot use the params hash. It exists but is just an empty hash. Why?
problem 2: I cannot deserialize the JSON generated by the class itself. Why?
The registration class looks like this:
require 'json'
class Registration
attr_accessor :user, :project_id
def to_json(*a)
{
'json_class' => self.class.name,
'data' => [#user.to_json(*a), #project_id]
}.to_json(*a)
end
def self.json_create(o)
new(*o['data'])
end
# Creates a new instance of the class using the information provided in the
# hash. If a field is missing in the hash, nil will be assigned to that field
# instead.
def initialize(params = {})
#user = params[:user]
#project_id = params[:project_id]
end
# Returns a string representing the entire Registration.
def inspect
"#{#user.inspect} - #{#user.country.inspect} - #{#project_id}"
end
# Returns a boolean valid representing whether the Registration instance is
# considered valid for the API or not. True if the instance is considered
# valid; otherwise false.
def is_valid?
return false if #user.nil? || #project_id.nil?
return false if !#user.is_a?(User) || !#project_id.is_a?(Fixnum)
return false if !#user.is_valid?
true
end
end
I had to implement the methods to generate the JSON output correctly. When I run this in console I get the following output generated:
irb(main):004:0> r = Registration.new(:user => u, :project_id => 1)
=> new_login - nil - 1
irb(main):005:0> r.to_json
=> "{\"json_class\":\"Registration\",\"data\":[\"{\\\"json_class\\\":\\\"User\\\
",\\\"login\\\":\\\"new_login\\\"}\",1]}"
Which looks like valid JSON to me. However when I POST this to the application server and try to parse this, JSON complains that at least 2 octets are needed and refuses to deserialize the object.
If you're using Sequel as your ORM, try something like this:
In your model:
class Registration < Sequel::Model
many_to_one :user
many_to_one :project
plugin :json_serializer
end
The server:
before do
#data = JSON.parse(request.body.read) rescue {}
end
post '/users' do
#registration = Registration.new #data
if #registration.valid?
#registration.save
#registration.to_json #return a JSON representation of the resource
else
status 422 #proper status code for invalid input
#registration.errors.to_json
end
end
I think you may be overcomplicating your registration process. If the HTTP action is POST /users then why not create a user? Seems like creating a registration is overly complex. Unless your user already exists, in which case POST /users would be incorrect. If what you're really intending to do is add a user to to a project, then you should PUT /projects/:project_id/users/:user_id and the action would look something like this:
class User < Sequel::Model
many_to_many :projects
end
class Project < Sequel::Model
many_to_many :users
end
#make sure your db schema has a table called users_projects or projects_users
put '/projects/:project_id/users/:user_id' do
#find the project
#project = Project.find params[:project_id]
raise Sinatra::NotFound unless #project
#find the user
#user = Project.find params[:project_id]
raise Sinatra::NotFound unless #user
#add user to project's users collection
#project.add_user #user
#send a new representation of the parent resource back to the client
#i like to include the child resources as well
#json might look something like this
#{ 'name' : 'a project name', 'users' : ['/users/:user_id', '/users/:another_user_id'] }
#project.to_json
end
I'm writing a few helpers to DRY up my tests. I pictured something like:
class ActiveSupport::TestCase
def self.test_presence_validation_of model, attribute
test "should not save #{model.to_s} with null #{attribute.to_s}", <<-"EOM"
#{model.to_s} = Factory.build #{model.to_sym}, #{attribute.to_sym} => nil
assert !#{model.to_s}.save, '#{model.to_s.capitalize} with null #{attribute.to_s} saved to the Database'
EOM
# Another one for blank attribute.
end
end
So that this:
class MemberTest < ActiveSupport::TestCase
test_presence_validation_of :member, :name
end
Executes exactly this at MemberTest class scope:
test 'should not save member with null name' do
member = Factory.build :member, :name => nil
assert !member.save, 'Member with null name saved to the Database'
end
Is it possible to do it this way (with a few adaptations, of course; I doubt my "picture" works), or do I have to use class_eval?
Have you seen Shoulda? It's great for testing common Rails functionality such as validations, relationships etc. https://github.com/thoughtbot/shoulda-matchers
In this case, it seems class_eval is necessary since I want to interpolate variable names into actual code.
Illustrated here.
Consider the following parent/child relationship where Parent is 1..n with Kids (only the relevant stuff here)...
class Parent < ActiveRecord::Base
# !EDIT! - was missing this require originally -- was the root cause!
require "Kid"
has_many :kids, :dependent => :destroy, :validate => true
accepts_nested_attributes_for :kids
validates_associated :kids
end
class Kid < ActiveRecord::Base
belongs_to :parent
# for simplicity, assume a single field: #item
validates_presence_of :item, :message => "is expected"
end
The validates_presence_of methods on the Kid model works as expected on validation failure, generating a final string of Item is expected per the custom message attribute supplied.
But if try validates_with, instead...
class Kid < ActiveRecord::Base
belongs_to :parent
validates_with TrivialValidator
end
class TrivialValidator
def validate
if record.item != "good"
record.errors[:base] << "Bad item!"
end
end
end
...Rails returns a NameError - uninitialized constant Parent::Kid error following not only an attempt to create (initial persist) user data, but also when even attempting to build the initial form. Relevant bits from the controller:
def new
#parent = Parent.new
#parent.kids.new # NameError, validates_* methods called within
end
def create
#parent = Parent.new(params[:parent])
#parent.save # NameError, validates_* methods called within
end
The error suggests that somewhere during model name (and perhaps field name?) resolution for error message construction, something has run afoul. But why would it happen for some validates_* methods and not others?
Anybody else hit a wall with this? Is there some ceremony needed here that I've left out in order to make this work, particularly regarding model names?
After a few hours away, and returning fresh -- Was missing require "Kid" in Parent class. Will edit.