Why are the results loading from the db, instead of cache? - ruby-on-rails

When I run:
IpPermission.blacklist.pluck(:ip)
I get the results:
=> ["127.0.0.11012q", "50.36.46.48"]
If I manually change any of the rows in the ip_permissions table, and run IpPermission.blacklist.pluck(:ip) again. It displays the updated results. Shouldn't it load the results from cache and not directly from the db?
My model looks like:
class IpPermission < ActiveRecord::Base
validates_presence_of :ip, :note, :category
validates_uniqueness_of :ip, :scope => [:category]
validates :category, :inclusion => { :in => ['whitelist', 'blacklist'] }
def self.whitelist
Rails.cache.fetch('whitelist', :expires_in => 1.month) { self.where(:category => 'whitelist') }
end
def self.blacklist
Rails.cache.fetch('blacklist', :expires_in => 1.month) { self.where(:category => 'blacklist') }
end
end

I think this should solve our problem:
Change this
def self.whitelist
Rails.cache.fetch('whitelist', :expires_in => 1.month) { self.where(:category => 'whitelist') }
end
def self.blacklist
Rails.cache.fetch('blacklist', :expires_in => 1.month) { self.where(:category => 'blacklist') }
end
to this
def self.whitelist
Rails.cache.fetch('whitelist', :expires_in => 1.month) { self.where(:category => 'whitelist').all }
end
def self.blacklist
Rails.cache.fetch('blacklist', :expires_in => 1.month) { self.where(:category => 'blacklist').all }
end
The where does not make the actual query, if you put .all then you make the query, so you are not storing the data but you just storing the query

Related

ThinkingSphinx excerpt issue with associations with special chars

I'm having an issue with special characters (apostrophe, namely), only when present is a nested association, however.
I have a 'Vendor' model and an 'Event' model, where a Vendor has_many Events. Here are the index files:
vendor_index:
ThinkingSphinx::Index.define :vendor, :with => :active_record do
indexes :name
indexes city
set_property :min_prefix_len => 2
set_property :enable_star => true
end
event_index:
ThinkingSphinx::Index.define :event, :with => :active_record do
indexes title
indexes subtitle
indexes venue_name
indexes vendor.name, :as => :vendor_name
indexes vendor.city, :as => :vendor_city
indexes genre.name, :as => :genre_name
where "workflow_state = 'published'"
set_property :min_prefix_len => 2
set_property :enable_star => true
end
I'm using an ExcerptorPane, like so, in my search#index action :
class SearchController < ApplicationController
helper_method :format_autocomplete
def index
#events = Event.search params[:search], {:star => true , :per_page => 5, :page => params[:events_page]}
#events.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
#vendors = Vendor.search params[:search], { :star => true , :per_page => 5, :page => params[:vendors_page]}
#vendors.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
#users = User.search params[:search], { :star => true , :per_page => 5, :page => params[:users_page]}
#users.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
end
# methods used for ajax-y pagination
def vendor_results
#vendors = Vendor.search params[:search], { :star => true , :per_page => 5, :page => params[:vendors_page]}
#vendors.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
respond_to do |format|
format.js
end
end
def user_results
#users = User.search params[:search], { :star => true , :per_page => 5, :page => params[:users_page]}
#users.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
respond_to do |format|
format.js
end
end
def event_results
#events = Event.search params[:search], { :star => true , :per_page => 5, :page => params[:events_page]}
#events.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
respond_to do |format|
format.js
end
end
def get_terms
results = ThinkingSphinx.search(params[:search], {:star => true})
results.context[:panes] << ThinkingSphinx::Panes::ExcerptsPane
results_json = format_autocomplete(results)
respond_to do |format|
format.js { render :json => results_json }
end
end
private
def format_autocomplete(r)
bucket = [];
r.each do |result|
puts result.class
if result.class.name == "Event"
title = result.excerpts.title
name = result.excerpts.vendor_name
bucket << {
:label => title,
:value => title,
:category => "Events",
:subtitle => result.excerpts.subtitle,
:url => event_url(result),
:vendor_name => name,
:vendor_city => result.excerpts.vendor_city,
:genre_name => result.excerpts.genre_name,
:venue_name => result.excerpts.venue_name
}
elsif result.class.name == "Vendor"
name = result.excerpts.name
bucket << {
:label => name,
:value => name,
:category => "Vendors",
:subtitle => result.excerpts.city,
:url => vendor_url(result)
}
elsif result.class.name == "User"
name = result.excerpts.name
bucket << {
:label => name,
:value => name,
:category => "Users",
:subtitle => result.excerpts.city,
:url => user_url(result)
}
end
end
bucket
end
end
I have also included a charset_table and ignore_chars in my thinking_sphinx.yml file.
Now, when I search for a Vendor with an apostrophe in their name, everything goes fine if the Vendor has no events. If a Vendor has events though, I get an error trying to render the event's vendor_name: (the full vendor name is "VIFF's Vancity Theatre, and the search query is 'viff')
sphinxql: syntax error, unexpected IDENT, expecting ')' near 's Vancity Theatre', 'event_core', '*viff*', '<span class="match">' AS before_match, '</span>' AS after_match, ' … ' AS chunk_separator)'
raised at this line in my view:
<p><%= link_to ( raw event.excerpts.vendor_name ), vendor_path(event.vendor) %></p>
I've been searching for a while, but can't find anything of help...Any ideas as to what might be causing this?
Thanks!
UPDATE:
It gets weirder ... with vendor name "Viff's Vancity Theatre", (and all events and vendors have city = 'Vancouver' ) if I search "van" or "vanc", everything renders fine, with "Vancity" marked as a match. However if I search "vanco" it breaks again. This happens when I am performing a search on specific models. When I perform a global search however (for autocomplete), I get the opposite behaviour - 'vanco' will work, but anything shorter throws back the same error. I've updated the code above with the full search_controller.rb.
I just received a pull request for Riddle that may contain a fix for this. Try adding the following to your Gemfile:
gem 'riddle', '~> 1.5.6',
:git => 'git://github.com/pat/riddle.git',
:branch => 'master',
:ref => '50d410cda6'

inner methods on attributes in respond_with

I'm rendering a serialized JSON response with the following code:
respond_with(#posts, :only => [:id, :content, :created_at],
:include => { :user => { :only => [:id, :name] },
:comments => { :only => [:content, :created_at] }})
the response is parsed in JAVA code so I want to convert the created_at value to a format that I can use, how do i run a method on each created_at value (even .to_i for example) ?
I would do it defining a new method inside comments model e.g.
class Comment < ActiveRecode::Base
...
def created_at_to_i
created_at.to_i
end
...
end
and while rendering
respond_with(#posts, :only => [:id, :content, :created_at],
:include => { :user => { :only => [:id, :name] },
:comments => { :only => [:content, :created_at_to_i] }})
Define method in each model like :
def formated_created_at
self.created_at.strftime(FORMAT_AS_YOU_LIKE)
end
At the time of rendering use it like :
respond_with(#posts, :only => [:id, :content, :formated_created_at
,
:include => { :user => { :only => [:id, :name] }, :comments => { :only => [:content, :formated_created_at
] }})
EDITED :
Instead of using the :only in respond_with , you can create your own hash and pass it to respond_with .
Like :
post_hash= { :post => [{:id => 1, :content => "abc", :created_at => self.formated_created_at,:user => {:id => 1, :name => 'Vik' }, :comments => { :content => "comment text", :created_at => self.comment.formated_created_at}] }
respond_with(post_hash)
And I think u can format the created_at at the time of displaying , via javascript, jQuery .

Rails 3.1 deep nesting with RABL

I'm using the RABL gem to render JSON user data for users of comments which are children of annotations which are children of images. I'd like to do something similar to:
object #image
node :annotations do
#image.annotations.map { |a| {
:id => a.id,
:x1 => a.x1,
:x2 => a.x2,
:y1 => a.y1,
:y2 => a.y2,
node :comments do
#image.annotations.comments.map { |c| {
:body => c.body,
:user_id => c.user_id,
:name => User.find(c.user_id).name,
:user_icon => user_icon(User.find(c.user_id), 'square', 30)
}}
end
}}
end
I know this isn't valid in RABL, I also tried using child instead of node, but couldn't access the user data that way. How should I go about doing this and whats the proper syntax to make this happen?
I got this answer from #calvin-l. The trick was to just map the a.comments then grab the data from each comment that way:
node :annotations do
#image.annotations.map { |a| {
:id => a.id,
:x1 => a.x1,
:x2 => a.x2,
:y1 => a.y1,
:y2 => a.y2,
:comments => a.comments.map { |c| {
:body => c.body,
:created_at => c.created_at,
:user => {
:id => c.user.id,
:facebook_id => c.user.facebook_id,
:name => c.user.name,
:username => c.user.username
}
}}
}}
end

Acts_as_paranoid and validation

Using rails 3 and acts_as_paranoid. I want to make sure that every task has a comment.
How can I bypass this one validation (check_if_notes) if mark_completed_and_msg is called?
EDIT - Is this the right way?
task.rb
has_one :comment, :as => :commentable
attr_accessor :force_task
# original was - before_update, :check_if_notes
validate :check_if_notes, :on => :update, :unless => proc { |a| a.force_task }
def mark_completed_and_msg(user_id, msg)
Comment.create!(:commentable_type => self.class, :commentable_id => self.id, :content => msg )
self.update_attributes(:completed_by => user_id, :deleted_at => Time.now, :force_task => true)
end
def check_if_notes
if self.comment.blank? || self.comment.content.blank? || self.comment.content.scan(/[\w-]+/).size <= 2
saved = false
self.comment.errors[:content] << "You must have a comment and more than 3 words."
raise ActiveRecord::Rollback
end
end
I think you meant using == in your Proc.
:unless => proc { |a| a[:force_task] == true }
You can also use
:unless => proc { |a| a.force_task? }

Adding an rails activerecord association within a loop

I want to add a has_many through association to a activerecord model class for each symbol in an array. for example
PeopleOrganisation::ROLES.each do |role|
has_many role.to_s.pluralize.to_sym, :through => :people_organisations, :source => :person,
:conditions => "people_organisations.role = '#{role.to_s}'" do
def << (object)
PeopleOrganisation.send(:with_scope, :create => {:role => **role**}) { self.concat object }
end
end
end
everything works fine except for the reference to the role variable inside the method def. This is because the method def is not a closure. Is there a way of achieving what I want?
Try this:
PeopleOrganisation::ROLES.each do |role|
has_many(role.to_s.pluralize.to_sym,
:through => :people_organisations, :source => :person,
:conditions => ["people_organisations.role = ?", role]
) do
define_method("<<") do |object|
PeopleOrganisation.send(:with_scope, :create => {:role => role}) {
self.concat object
}
end
end
end
Instead of defining method using def you can try define_method method:
PeopleOrganisation::ROLES.each do |role|
has_many role.to_s.pluralize.to_sym, :through => :people_organisations, :source => :person,
:conditions => "people_organisations.role = '#{role.to_s}'" do
define_method(:<<) do |object|
PeopleOrganisation.send(:with_scope, :create => {:role => role}) { self.concat object }
end
end
end

Resources