I am assigned to write some ruby code that will work with the following (segment of a) rspec test:
before do
#book = Book.new
end
describe 'title' do
it 'should capitalize the first letter' do
#book.title = "inferno"
#book.title.should == "Inferno"
end
This is the solution, but I don't understand it:
class Book
attr_reader :title
def title=(new_title)
words = new_title.split(" ")
words = [words[0].capitalize] +
words[1..-1].map do |word|
little_words = %w{a an and the in of}
if little_words.include? word
word
else
word.capitalize
end
end
#title = words.join(" ")
end
end
I think I am correct to deduce that #book.title = "inferno" will run the title method and eventually create a new value for the #title variable at the bottom. I know that this causes #book.title to update to "Inferno" (capitalized), but I'm not sure why. Is this a case of def title being some sort of variable method, and #title being it's final value? That's my best guess at this point.
EDIT in case it's not clear, what I'm not understanding is why setting #book.title ='inferno' causes #book.title to update to "Inferno".
When you have setter and getter methods in Ruby:
attr_writer :something
attr_reader :something
From my little understanding of this, these methods are equivalent to
def something=(value)
#something = value
end
def something
#something
end
Respectively.
Or in one statement, it could be:
attr_accessor :something
Anyway, what you are doing is to write the setter method yourself, capitalising each word of the string passed as an argument.
Your understanding is almost correct. Here is a simple example
class Chapter
attr_reader :title
def title=(new_title)
#title = new_title.reverse
end
end
#c = Chapter.new
#c.title = "ybuR"
#c.title #=> Ruby
Related
I have a difficult problem to solve.
I would like to replicate, the title_with_if_statement methods results, but using the syntax used in the title method. However, I do not know what to put in the enforce_policy_to_level() method to accomplish this.
I can do 'title_less_readable', but I find the title method much more readable since the policy is right at the top, and since I will be doing this for hundreds of methods, I want to keep it DRY and readable.
Any advice?
Is this even possible? It's basically a before_action but uses the method's own results.
class Car
def title
enforce_policy_to_level(1)
# Code that returns the "Car Title" string
"Car Title"
end
def title_less_readable
# Code that returns "Car Title" string
content = "Car Title"
return secure_text(content) unless authorized_to_level?(1)
content
end
def title_with_if_statement
# Code that returns the "Car Title" string
content = "Car Title"
if authorized_to_level?(1)
content
else
secure_text(content)
end
end
private
def enforce_policy_to_level(level)
# Return from title method with the secured result of title method
# if it's not authorized. In this example it would be secure_text('Car Title')
# which output return '---------'
# If if authorized, just continue with how title method would normally return
end
def authorized_to_level?(level)
current_user_level = 1 # Dynamic from user record
current_user_level >= level
end
def secure_text(text)
'X' * text.length
end
end
Why not just something like this?
def title
enforce_policy_to_level(1) do
"Car Title"
end
end
private
def enforce_policy_to_level(level, &block)
if authorized_to_level?(level)
secure_txt yield
else
yield
end
end
Consider separating your concerns: I'd approach this by creating a separate function that handles the authorization.
module AuthorizedText
module_function
def secure_text(text, authorized)
return authorized ? text : 'X' * text.length
end
end
Call this in Car like this:
AuthorizedText.secure_text("Car Title", authorized_to_level(1))
You can shorthand this with a convenience method:
def secure(level, text)
AuthorizedText.secure_text(text, authorized_to_level(level))
end
so that your car_title method would look like
def car_title
secure(1, "Car Title")
end
You could generalize a method wrapper by defining a class method that would let you do something like
def car_title
"Car Title"
end
secure :car_title
But I'd call this over-engineered and ultimately less readable to anyone who isn't you (including you 8 months from now, likely!). I'd avoid trying to be fancy like this, as fun as it may be.
Also, if current_user and current_user_level make sense outside the context of Car and you may want to reuse this elsewhere, you could go further by moving the auth checking methods into AuthorizedText, if that's appropriate:
module AuthorizedText
module_function
def secure_text(text, user, level = 0)
return authorized?(user, level) ? text : 'X' * text.length
end
def authorized?(user, level)
# for example; I don't know what your user object looks like
user.level >= level
end
end
In my app that I am building to learn Rails and Ruby, I have below iteration/loop which is not functioning as it should.
What am I trying to achieve?
I am trying to find the business partner (within only the active once (uses a scope)) where the value of the field business_partner.bank_account is contained in the field self_extracted_data and then set the business partner found as self.sender (self here is a Document).
So once a match is found, I want to end the loop. A case exists where no match is found and sender = nil so a user needs to set it manually.
What happens now, is that on which ever record of the object I save (it is called as a callback before_save), it uses the last identified business partner as sender and the method does not execute again.
Current code:
def set_sender
BusinessPartner.active.where.not(id: self.receiver_id).each do |business_partner|
bp_bank_account = business_partner.bank_account.gsub(/\s+/, '')
rgx = /(?<!\w)(#{Regexp.escape(bp_bank_account)})?(?!\w)/
if self.extracted_data.gsub(/\s+/, '') =~ rgx
self.sender = business_partner
else
self.sender = nil
end
end
end
Thanks for helping me understand how to do this kind of case.
p.s. have the pickaxe book here yet this is so much that some help / guidance would be great. The regex works.
Using feedback from #moveson, this code works:
def match_with_extracted_data?(rgx_to_match)
extracted_data.gsub(/\s+/, '') =~ rgx_to_match
end
def set_sender
self.sender_id = matching_business_partner.try(:id) #unless self.sender.id.present? # Returns nil if no matching_business_partner exists
end
def matching_business_partner
BusinessPartner.active.excluding_receiver(receiver_id).find { |business_partner| sender_matches?(business_partner) }
end
def sender_matches?(business_partner)
rgx_registrations = /(#{Regexp.escape(business_partner.bank_account.gsub(/\s+/, ''))})|(#{Regexp.escape(business_partner.registration.gsub(/\s+/, ''))})|(#{Regexp.escape(business_partner.vat_id.gsub(/\s+/, ''))})/
match_with_extracted_data?(rgx_registrations)
end
In Ruby you generally want to avoid loops and #each and long, procedural methods in favor of Enumerable iterators like #map, #find, and #select, and short, descriptive methods that each do a single job. Without knowing more about your project I can't be sure exactly what will work, but I think you want something like this:
# /models/document.rb
class Document < ActiveRecord::Base
def set_sender
self.sender = matching_business_partner.try(:id) || BusinessPartner.active.default.id
end
def matching_business_partners
other_business_partners.select { |business_partner| account_matches?(business_partner) }
end
def matching_business_partner
matching_business_partners.first
end
def other_business_partners
BusinessPartner.excluding_receiver_id(receiver_id)
end
def account_matches?(business_partner)
rgx = /(?<!\w)(#{Regexp.escape(business_partner.stripped_bank_account)})?(?!\w)/
data_matches_bank_account?(rgx)
end
def data_matches_bank_account?(rgx)
extracted_data.gsub(/\s+/, '') =~ rgx
end
end
# /models/business_partner.rb
class BusinessPartner < ActiveRecord::Base
scope :excluding_receiver_id, -> (receiver_id) { where.not(id: receiver_id) }
def stripped_bank_account
bank_account.gsub(/\s+/, '')
end
end
Note that I am assigning an integer id, rather than an ActiveRecord object, to self.sender. I think that's what you want.
I didn't try to mess with the database relations here, but it does seem like Document could include a belongs_to :business_partner, which would give you the benefit of Rails methods to help you find one from the other.
EDIT: Added Document#matching_business_partners method and changed Document#set_sender method to return nil if no matching_business_partner exists.
EDIT: Added BusinessPartner.active.default.id as the return value if no matching_business_partner exists.
What's the best Ruby/Rails way to allow users to use decimals or commas when entering a number into a form? In other words, I would like the user be able to enter 2,000.99 and not get 2.00 in my database.
Is there a best practice for this?
Does gsub work with floats or bigintegers? Or does rails automatically cut the number off at the , when entering floats or ints into a form? I tried using self.price.gsub(",", "") but get "undefined method `gsub' for 8:Fixnum" where 8 is whatever number I entered in the form.
I had a similar problem trying to use localized content inside forms. Localizing output is relatively simple using ActionView::Helpers::NumberHelper built-in methods, but parsing localized input it is not supported by ActiveRecord.
This is my solution, please, tell me if I'm doing anything wrong. It seems to me too simple to be the right solution. Thanks! :)
First of all, let's add a method to String.
class String
def to_delocalized_decimal
delimiter = I18n::t('number.format.delimiter')
separator = I18n::t('number.format.separator')
self.gsub(/[#{delimiter}#{separator}]/, delimiter => '', separator => '.')
end
end
Then let's add a class method to ActiveRecord::Base
class ActiveRecord::Base
def self.attr_localized(*fields)
fields.each do |field|
define_method("#{field}=") do |value|
self[field] = value.is_a?(String) ? value.to_delocalized_decimal : value
end
end
end
end
Finally, let's declare what fields should have an input localized.
class Article < ActiveRecord::Base
attr_localized :price
end
Now, in your form you can enter "1.936,27" and ActiveRecord will not raise errors on invalid number, because it becomes 1936.27.
Here's some code I copied from Greg Brown (author of Ruby Best Practices) a few years back. In your model, you identify which items are "humanized".
class LineItem < ActiveRecord::Base
humanized_integer_accessor :quantity
humanized_money_accessor :price
end
In your view templates, you need to reference the humanized fields:
= form_for #line_item do |f|
Price:
= f.text_field :price_humanized
This is driven by the following:
class ActiveRecord::Base
def self.humanized_integer_accessor(*fields)
fields.each do |f|
define_method("#{f}_humanized") do
val = read_attribute(f)
val ? val.to_i.with_commas : nil
end
define_method("#{f}_humanized=") do |e|
write_attribute(f,e.to_s.delete(","))
end
end
end
def self.humanized_float_accessor(*fields)
fields.each do |f|
define_method("#{f}_humanized") do
val = read_attribute(f)
val ? val.to_f.with_commas : nil
end
define_method("#{f}_humanized=") do |e|
write_attribute(f,e.to_s.delete(","))
end
end
end
def self.humanized_money_accessor(*fields)
fields.each do |f|
define_method("#{f}_humanized") do
val = read_attribute(f)
val ? ("$" + val.to_f.with_commas) : nil
end
define_method("#{f}_humanized=") do |e|
write_attribute(f,e.to_s.delete(",$"))
end
end
end
end
You can try stripping out the commas before_validation or before_save
Oops, you want to do that on the text field before it gets converted. You can use a virtual attribute:
def price=(price)
price = price.gsub(",", "")
self[:price] = price # or perhaps price.to_f
end
Take a look at the i18n_alchemy gem for date & number parsing and localization.
I18nAlchemy aims to handle date, time and number parsing, based on current I18n locale format. The main idea is to have ORMs, such as ActiveRecord for now, to automatically accept dates/numbers given in the current locale format, and return these values localized as well.
I have written following code in my project. This solved all of my problems.
config/initializers/decimal_with_comma.rb
# frozen_string_literal: true
module ActiveRecord
module Type
class Decimal
private
alias_method :cast_value_without_comma_separator, :cast_value
def cast_value(value)
value = value.gsub(',', '') if value.is_a?(::String)
cast_value_without_comma_separator(value)
end
end
class Float
private
alias_method :cast_value_without_comma_separator, :cast_value
def cast_value(value)
value = value.gsub(',', '') if value.is_a?(::String)
cast_value_without_comma_separator(value)
end
end
class Integer
private
alias_method :cast_value_without_comma_separator, :cast_value
def cast_value(value)
value = value.gsub(',', '') if value.is_a?(::String)
cast_value_without_comma_separator(value)
end
end
end
end
module ActiveModel
module Validations
class NumericalityValidator
protected
def parse_raw_value_as_a_number(raw_value)
raw_value = raw_value.gsub(',', '') if raw_value.is_a?(::String)
Kernel.Float(raw_value) if raw_value !~ /\A0[xX]/
end
end
end
end
I was unable to implement the earlier def price=(price) virtual attribute suggestion because the method seems to call itself recursively.
I ended up removing the comma from the attributes hash, since as you suspect ActiveRecord seems to truncate input with commas that gets slotted into DECIMAL fields.
In my model:
before_validation :remove_comma
def remove_comma
#attributes["current_balance"].gsub!(',', '') # current_balance here corresponds to the text field input in the form view
logger.debug "WAS COMMA REMOVED? ==> #{self.current_balance}"
end
Here's something simple that makes sure that number input is read correctly. The output will still be with a point instead of a comma. That's not beautiful, but at least not critical in some cases.
It requires one method call in the controller where you want to enable the comma delimiter. Maybe not perfect in terms of MVC but pretty simple, e.g.:
class ProductsController < ApplicationController
def create
# correct the comma separation:
allow_comma(params[:product][:gross_price])
#product = Product.new(params[:product])
if #product.save
redirect_to #product, :notice => 'Product was successfully created.'
else
render :action => "new"
end
end
end
The idea is to modify the parameter string, e.g.:
class ApplicationController < ActionController::Base
def allow_comma(number_string)
number_string.sub!(".", "").sub!(",", ".")
end
end
You can try this:
def price=(val)
val = val.gsub(',', '')
super
end
i have created a LocalizedString custom data type for storing / displaying translations using mongo_mapper.
This works for one field but as soon as i introduce another field they get written over each and display only one value for both fields. The to_mongo and from_mongo seem to be not workings properly. Please can any one help with this ? her is the code :
class LocalizedString
attr_accessor :translations
def self.from_mongo(value)
puts self.inspect
#translations ||= if value.is_a?(Hash)
value
elsif value.nil?
{}
else
{ I18n.locale.to_s => value }
end
#translations[I18n.locale.to_s]
end
def self.to_mongo(value)
puts self.inspect
if value.is_a?(Hash)
#translations = value
else
#translations[I18n.locale.to_s] = value
end
#translations
end
end
Thank alot
Rick
The problem is that from within your [to|from]_mongo methods, #translations refers to a class variable, not the instance variable you expect. So what's happening is that each time from_mongo is called, it overwrites the value.
A fixed version would be something like this:
class LocalizedString
attr_accessor :translations
def initialize( translations = {} )
#translations = translations
end
def self.from_mongo(value)
if value.is_a?(Hash)
LocalizedString.new(value)
elsif value.nil?
LocalizedString.new()
else
LocalizedString.new( { I18n.locale.to_s => value })
end
end
def self.to_mongo(value)
value.translations if value.present?
end
end
I found that jared's response didn't work for me -- I would get that translations was not found when using LocalizedString in an EmbeddedDocument.
I would get a similar problem on rick's solution where translations was nil when using embedded documents. To get a working solution, I took Rick's solution, changed the translation variable to be an instance variable so it wouldn't be overwritten for each new field that used LocalizedString, and then added a check to make sure translations wasn't nil (and create a new Hash if it was).
Of all the LocalizedString solutions floating around, this is the first time I've been able to get it working on EmbeddedDocuments and without the overwritting problem -- there still may be other issues! :)
class LocalizedString
attr_accessor :translations
def self.from_mongo(value)
puts self.inspect
translations ||= if value.is_a?(Hash)
value
elsif value.nil?
{}
else
{ I18n.locale.to_s => value }
end
translations[I18n.locale.to_s]
end
def self.to_mongo(value)
puts self.inspect
if value.is_a?(Hash)
translations = value
else
if translations.nil?
translations = Hash.new()
end
translations[I18n.locale.to_s] = value
end
translations
end
end
I found this post: which was very helpful. He extended HashWithIndifferentAccess to work as a LocalizedString. The only thing I didn't like about it was having to explicly specify the locale when setting it each time -- I wanted it to work more like a string. of course, you can't overload the = operator (at least I don't think you can) so I used <<, and added a to_s method that would output the string of the current locale....
class LocalizedString < HashWithIndifferentAccess
def self.from_mongo(value)
LocalizedString.new(value || {})
end
def available_locales
symbolize_keys.keys
end
def to_s
self[I18n.locale]
end
def in_current_locale=(value)
self[I18n.locale] = value
end
def << (value)
self[I18n.locale] = value
end
end
and then I have a class like:
class SimpleModel
include MongoMapper::Document
key :test, LocalizedString
end
and can do things like
I18n.locale = :en
a = SimpleModel.new
a.test << "English"
I18n.locale = :de
a.test << "German"
puts a.test # access the translation for the current locale
I18n.locale = :en
puts a.test # access the translation for the current locale
puts a.test[:de] # access a translation explicitly
puts a.test[:en]
puts a.test.inspect
and get
German
English
German
English
{"en"=>"English", "de"=>"German"}
so there we go -- this one actually seems to work for me. Comments welcome, and hope this helps someone!
I have a method call in my view like this
<%= Navigation.with(params) do |menu|
if current_user && current_user.can_verify?
menu.item("Listings", manage_listings_path())
menu.item("Listing changes", needing_change_approval_manage_listings_path())
menu.item("Flagged Items", flagged_manage_listings_path())
menu.item("Transfers", manage_listing_transfers_path())
menu.item("Reviews", manage_listing_reviews_path())
end
if current_user && current_user.admin?
menu.item("Log", manage_verifications_path())
menu.item("Indexer Compensations", manage_compensations_path())
menu.item("Users", manage_users_path())
end
end%>
that splits out the below string
"<li>Listings</li> <li>Listing changes</li> <li>Flagged Items</li> <li>Transfers</li> <li>Reviews</li> <li>Log</li> <li>Indexer Compensations</li> <li>Users</li>"
I just get this string in my page. I wanted them to be menus nicely styled by CSS. I am just getting the above raw text in my page. How do I convert this string to be treated as HTML by the browser.
Please help
Here is the navigation class
class NavigationMenu < ActionView::Base
def initialize(params)
#params = params
end
def item(title, path, options={})
#items ||= Array.new
unless (route = Rails.application.routes.recognize_path(path,:method => options[:method]|| :get))
raise "Unrecognised path #{path}, are you sure it's in routes.rb?"
end
#items << content_tag(:li, link_to(title,path, :class => (#params[:controller] == route[:controller] && #params[:action] == route[:action])? 'active' : nil))
end
def output
return '' if #items.blank?
content_tag(:ul, #items.join("\n"), :id => 'navigation')
end
end
class Navigation
def self.with(params, &block)
menu = NavigationMenu.new(params)
yield menu
menu.output
end
end
You have to add a call to the raw method:
<%= raw ... %>
This is necessary, because in Rails 3 every string is escaped by default, unless you use the raw method.
It's like an inverse of the h method in Rails 2, where every string is unescaped by default, unless you use the h method.
Example:
This code in Rails 2...
<%= h "String which must be escaped" %>
<%= "String which must be output raw %>
... must be this in Rails 3:
<%= "String which must be escaped" %>
<%= raw "String which must be output raw %>
(Although an additional call to h doesn't do any harm in Rails 3)
You need to append .html_safe to the string - this will stop rails from escaping it when it's time to output text. Probably best to put it in the item method that you call repeatedly.
I recently wrote an article regarding XSS protection in Rails 3 when upgrading from Rails 2:
http://developer.uservoice.com/entries/upgrading-to-rails-3-printing-escaped-strings
The idea is to hook code to printing HTML so that we can determine when we are actually printing something we don't want to:
module ActionView
module Helpers
module TextHelper
def simple_format_with_double_escape_reporting(*args)
HtmlDoubleEscapeReporter.assert_sane(simple_format_without_double_escape_reporting(*args))
end
alias_method_chain :simple_format, :double_escape_reporting
end
module TagHelper
private
def content_tag_string_with_double_escape_reporting(*args)
HtmlDoubleEscapeReporter.assert_sane(content_tag_string_without_double_escape_reporting(*args))
end
alias_method_chain :content_tag_string, :double_escape_reporting
end
module UrlHelper
def link_to_with_double_escape_reporting(*args, &block)
HtmlDoubleEscapeReporter.assert_sane(link_to_without_double_escape_reporting(*args, &block))
end
alias_method_chain :link_to, :double_escape_reporting
end
end
end
Method HtmlDoubleEscapeReporter.assert_sane can be written, for example, like this:
class HtmlDoubleEscapeReporter
def self.assert_sane(str)
if (str.match(/<[a-z]/) || str.match(/&(quot|rarr|larr|amp|#)/)) &&
!str.match(/looks something you do not want to print/
send_problem_report('#{str}' looks something you do not want to print")
end
return str
end
end
Here, 'looks something you do not want to print' is used to prevent the possibility of infinite loops. The line send_problem_report('#{str}' looks something you do not want to print") can be replaced with a call to "debugger" (from ruby-debug gem) so that you are able to check the backtrace and see where the problem is coming from.
Here is the new class. At last... I got that bug.
class NavigationMenu < ActionView::Base
def initialize(params)
#params = params
end
def item(title, path, options={})
#items ||= Array.new
unless (route = Rails.application.routes.recognize_path(path,:method => options[:method]|| :get))
raise "Unrecognised path #{path}, are you sure it's in routes.rb?"
end
#items << content_tag(:li, link_to(title,path, :class => (#params[:controller] == route[:controller] && #params[:action] == route[:action])? 'active' : nil))
end
def output
#items = #items.join("\n").html_safe
return '' if #items.blank?
content_tag(:ul, #items, :id => 'navigation')
end
end
class Navigation
def self.with(params, &block)
menu = NavigationMenu.new(params)
yield menu
menu.output
end
end