Rails 3: Find parent of polymorphic model in controller? - ruby-on-rails-3

I'm trying to find an elegant (standard) way to pass the parent of a polymorphic model on to the view. For example:
class Picture < ActiveRecord::Base
belongs_to :imageable, :polymorphic => true
end
class Employee < ActiveRecord::Base
has_many :pictures, :as => :imageable
end
class Product < ActiveRecord::Base
has_many :pictures, :as => :imageable
end
The following way (find_imageable) works, but it seems "hackish".
#PictureController (updated to include full listing)
class PictureController < ApplicationController
#/employees/:id/picture/new
#/products/:id/picture/new
def new
#picture = imageable.pictures.new
respond_with [imageable, #picture]
end
private
def imageable
#imageable ||= find_imageable
end
def find_imageable
params.each do |name, value|
if name =~ /(.+)_id$/
return $1.classify.constantize.find(value)
end
end
nil
end
end
Is there a better way?
EDIT
I'm doing a new action. The path takes the form of parent_model/:id/picture/new and params include the parent id (employee_id or product_id).

I'm not sure exactly what you're trying to do but if you're trying to find the object that 'owns' the picture you should be able to use the imageable_type field to get the class name. You don't even need a helper method for this, just
def show
#picture = Picture.find(params[:id])
#parent = #picture.imagable
#=> so on and so forth
end
Update
For an index action you could do
def index
#pictures = Picture.includes(:imagable).all
end
That will instantiate all 'imagables' for you.
Update II: The Wrath of Poly
For your new method you could just pass the id to your constructor, but if you want to instantiate the parent you could get it from the url like
def parent
#parent ||= %w(employee product).find {|p| request.path.split('/').include? p }
end
def parent_class
parent.classify.constantize
end
def imageable
#imageable ||= parent_class.find(params["#{parent}_id"])
end
You could of course define a constant in your controller that contained the possible parents and use that instead of listing them in the method explicitly. Using the request path object feels a little more 'Rails-y' to me.

I just ran into this same problem.
The way I 'sort of' solved it is defining a find_parent method in each model with polymorphic associations.
class Polymorphic1 < ActiveRecord::Base
belongs_to :parent1, :polymorphic => true
def find_parent
self.parent1
end
end
class Polymorphic2 < ActiveRecord::Base
belongs_to :parent2, :polymorphic => true
def find_parent
self.parent2
end
end
Unfortunately, I can not think of a better way. Hope this helps a bit for you.

This is the way I did it for multiple nested resources, where the last param is the polymorphic model we are dealing with: (only slightly different from your own)
def find_noteable
#possibilities = []
params.each do |name, value|
if name =~ /(.+)_id$/
#possibilities.push $1.classify.constantize.find(value)
end
end
return #possibilities.last
end
Then in the view, something like this:
<% # Don't think this was needed: #possibilities << picture %>
<%= link_to polymorphic_path(#possibilities.map {|p| p}) do %>
The reason for returning the last of that array is to allow finding the child/poly records in question i.e. #employee.pictures or #product.pictures

Related

Load associations to one level while conditionally sideloading associations in Active model serializers

AMS version 0.8.3,
I created a base_serializer.rb like this and extended the same.
class BaseSerializer < ActiveModel::Serializer
def include_associations!
if #options[:embed]
embed = #options[:embed].split(',').map{|item| item.strip.to_sym}
embed.each do |assoc|
include! assoc if _associations.keys.include?(assoc)
end
end
end
end
class EventSerializer < BaseSerializer
attributes :id, :name
has_many :organizers, serializer: OrganizerSerializer
has_many :participants, serializer: ParticipantSerializer
end
class OrganizerSerializer < BaseSerializer
attributes :id, :name
has_many :related, serializer: RelatedSerializer
end
class ParticipantSerializer < BaseSerializer
attributes :id, :name
has_many :related, serializer: RelatedSerializer
end
class RelatedSerializer < BaseSerializer
attributes :id, :name
has_many :something, serializer: SomethingSerializer
end
and the index method in EventsController is written as
# GET /events?embed=organizers,participants
def index
#events = Event.all
render json: #events, embed: params[:embed]
end
With this I can get the :id and :name of events, organizers and participants. But, I want the attributes of related association as well. I don't need details of something serializer. I want to go till this level for each association. How can I achieve that?
I ended up doing this to achieve the same.
class BaseSerializer < ActiveModel::Serializer
def include_associations!
#options[:embed_level] ||= 2
return unless #options.key?(:embed) && #options[:embed_level] != 0
embed = #options[:embed].split(',').map{|item| item.strip.to_sym}
embed.each do |assoc|
next unless _associations.key?(assoc)
assoc_serializer = serializer_for(assoc)
embed = #options[:embed]
embed_level = #options[:embed_level]
#options[:embed_level] = #options[:embed_level] - 1
#options[:embed] = assoc_serializer._associations.keys.join(",")
include! assoc
#options[:embed_level] = embed_level
end
end
def serializer_for(assoc)
serializer = _associations[assoc].options[:serializer]
return serializer if serializer
assoc.to_s.classify.concat("Serializer").constantize
end
end
Ref: Github Issue Link
Special Thanks to Yohan Robert!!!

Ruby on Rails - Simplifying similar methods that access different variables

I'm working on a fairly simple site that allows users to choose recipe ingredients, their quantities and then shows them nutritional info based on their recipe and a large database.
Right now, I feel like I'm repeating myself a bit. I want to be able to make this "DRY" by having one method each in the Recipe and Recipe_Ingredient model that will do the same thing only accept the right parameter, which will be the type of nutrient.
Here is the relevant code in my view that currently calls two different methods (and will call more when extended to the other nutrients):
<ul>Calories <%= #recipe.total_calories %></ul>
<ul>Fat (grams) <%= #recipe.total_fat %></ul>
In my recipe model, I have methods that iterate over each of the ingredients in the recipe:
def total_calories
recipe_ingredients.to_a.sum { |i| i.total_calories }
end
def total_fat
recipe_ingredients.to_a.sum { |i| i.total_fat }
end
In the block, we call two separate methods that actually calculate the nutrients for each individual recipe ingredient:
def total_calories
ingredient.calories*ingredient.weight1*quantity/100
end
def total_fat
ingredient.fat*ingredient.weight1*quantity/100
end
This last piece is where we reference the database of ingredients. For context, here are the relationships:
class RecipeIngredient < ActiveRecord::Base
belongs_to :ingredient
belongs_to :recipe
class Recipe < ActiveRecord::Base
has_many :recipe_ingredients
Thanks in advance for any help.
Lev
The send method with a symbol parameter works well for that kind of DRY.
<ul>Calories <%= #recipe.total :calories %></ul>
<ul>Fat (grams) <%= #recipe.total :fat %></ul>
Recipe
def total(type)
recipe_ingredients.to_a.sum { |i| i.total type }
end
RecipeIngredient
def total(type)
ingredient.send(type) * ingredient.weight1 * quantity / 100
end
You could use meta programming to dynamically add the methods. Here is a start, you can get even more DRY than this.
class DynamicTotalMatch
attr_accessor :attribute
def initialize(method_sym)
if method_sym.to_s =~ /^total_of_(.*)$/
#attribute = $1.to_sym
end
end
def match?
#attribute != nil
end
end
Recipe
class Recipe
def self.method_missing(method_sym, *arguments, &block)
match = DynamicTotalMatch.new(method_sym)
if match.match?
define_dynamic_total(method_sym, match.attribute)
send(method_sym, arguments.first)
else
super
end
end
def self.respond_to?(method_sym, include_private = false)
if DynamicTotalMatch.new(method_sym).match?
true
else
super
end
end
protected
def self.define_dynamic_total(method, attribute)
class_eval <<-RUBY
def self.#{method}(#{attribute})
recipe_ingredients.to_a.sum { |i| i.send(attribute)
end
RUBY
end
end
RecipeIngredient
class RecipeIngredient
def self.method_missing(method_sym, *arguments, &block)
match = DynamicTotalMatch.new(method_sym)
if match.match?
define_dynamic_total(method_sym, match.attribute)
send(method_sym, arguments.first)
else
super
end
end
def self.respond_to?(method_sym, include_private = false)
if DynamicTotalMatch.new(method_sym).match?
true
else
super
end
end
protected
def self.define_dynamic_total(method, attribute)
class_eval <<-RUBY
def self.#{method}(#{attribute})
ingredient.send(attribute) * ingredient.weight1 * quantity / 100
end
RUBY
end
end
Example was copied from ActiveRecord and this page: http://technicalpickles.com/posts/using-method_missing-and-respond_to-to-create-dynamic-methods/

best way to save dependent objects in a has_and_belongs_to_many relation?

Hi I am new to rails and I would like to know what is the best way who save dependent objects in an HBTM relation.
Specifically, I have two classes Post and Tag
class Post < ActiveRecord::Base
has_and_belongs_to_many :tags
end
class Tag < ActiveRecord::Base
has_and_belongs_to_many :posts
end
I have a migration to create the joining table
class AddPostsTagsJoinTable < ActiveRecord::Migration
def self.up
create_table :posts_tags, :id => false do |t|
t.integer :post_id
t.integer :tag_id
end
end
def self.down
drop_table :postss_tags
end
end
All is good up to here
So I have a PostsController from which I handle the creation, updates and deletes for the posts, and I want to encapsulate the Tags so that the creation is via the PostsController... like so:
class PostsController < ApplicationController
#... code removed for brevity
def create
#post = current_user.posts.build(params[:post])
if #post.save
tag_names = params[:post][:tags].strip.split(' ')
tag_names.each do |t|
#see if the tag already exists
tag = Tag.find_by_name(t);
if tag.nil?
#post.tags.create!(:name => t)
else
#post.tags << tag #just create the association
end
end
flash[:success] = "Post created."
redirect_to(user_posts_path(current_user.username))
else
#user = current_user
render 'new'
end
end
end
I am not sure how I should handle the creation of my Tag(s) because if I just call
#post.tags.create!(:name => t)
this will create duplicate records in the Tags table (even when :uniq => true is specified in the model).
So to avoid the duplication I see if a tag is already present and then add it like this
tag = Tag.find_by_name(t);
if tag.nil?
#post.tags.create!(:name => t)
else
#post.tags << tag #just create the association
end
Is this the way it's supposed to be done?
This seems expensive (especially 'cause it's in a loop) so I am wondering if there is another "cleaner" way to do this? (pls forget the DRY'ing up of the action and so on)
Is there a clean way to create my Tags without having to manually check for duplicates?
thank you in advance for your help!
You can save tags attribute of post if automatically by adding accepts_nested_attributes_for to Post model
class Post < ActiveRecord::Base
has_and_belongs_to_many :tags
accepts_nested_attributes_for :tags
end
The next step is to output tags fields inside post form.

Custom return value for new / create ActiveRecord model

So I have a model (Photo), where when I call Photo.new #image => #image / Photo.create :image => #image, I want my model to find an existing photo with the same image hash OR create a new Photo from #image. Assume I can't use Photo.find_or_initialize_by_hash because I have a custom find function which finds close copies of images based on a soft image hash.
My first idea was to do
before_validation :check_duplicates, :on => :create
def check_duplicates
self = self.find_duplicate
end
Unfortunately, I realized you can't just redefine self in a model, so now I think the best approach is doing something along the lines of changing the return value from initialize to the duplicate.
Sort of like this, but it doesn't work (and I've heard horror stories about overriding initialize)
def initialize(*params)
super(*params)
return self.find_duplicate || self
end
From what I gather your model structure looks something like this?
class Photo < ActiveRecord::Base
has_one :image
end
class Image < ActiveRecord::Base
belongs_to :photo
end
If so, you can simply do this:
class Photo < ActiveRecord::Base
has_one :image, :uniq => true
end
Or if :image is just an attribute of Photo your first idea was on track:
class Photo < ActiveRecord::Base
before_create :check_duplicate
private
def check_duplicate
Photo.where(:image => self.image).count == 0 # will be false if Photo is found
end
end
which will cancel the Photo from being created if #check_duplicate returns false (http://api.rubyonrails.org/classes/ActiveRecord/Callbacks.html)
Or simply
class Photo < ActiveRecord::Base
validates_uniqueness_of :image
end

Using after_create

I have a model, Category. And I want to create an new default sub_category when ever the category is created. But I'm not sure how to do it. Here is what I have.
class Category < ActiveRecord::Base
attr_accessible :title, :position
has_many :sub_categories
after_create :make_default_sub
def make_default_sub
#Sub_Categories.new( :title=>' ');
end
end
Why not to use ancestry gem? In the future if you will have more subcategories, it will be easier to manage them.
For example in your case:
class Category < ActiveRecord::Base
attr_accessible :title, :position
has_ancestry
after_create :create_default_subcategory
def make_default_sub
children = self.children.new
children.title = ''
children.position = 1 # or autogenerated
children.save!
end
end
But can you explain, why do you need such a strange default behaviour?
Thanks