General Notes

Very nice tutorial for rails 2.1 caching.

API references: Fragments, Sweepers, the Store

Post how to move the page cache in a folder with some server examples

Observing Controllers with Sweepers

Here is described how you can use the Sweepers to observe controller actions. I totaly agree that the documenation is very bad on the sweepers and it is like a rule book of MTG than usefull rdoc.

Here is what I have found on the naming of the callbacks.

Checking the source code in sweeping.rb shows that the callbacks that are used for the controller are constructed in this way:

controller_callback_method_name = "#{timing}_#{controller.controller_name.underscore}"
action_callback_method_name = "#{controller_callback_method_name}_#{controller.action_name}"

This means that the method which will be called in the sweeper will be named

• after/before_user – for the controller UserController
• after/before_user_login – for the action login in the controller UserController

Here is an example.

class Dagens::UserController < Dagens::BaseController
  cache_sweeper DagensAgentSweeper, :only => [:do_login, :do_delete_account, :do_logout]
  def do_login
  end
  ....
end

class DagensAgentSweeper < ActionController::Caching::Sweeper
  observe DagensAgent

  def after_user_do_login
    puts "after do login \n \n \n \n \n \n \n "
    expire_cache_for(@logged_user)
  end
end

Skipping callbacks from an observer

I know that it is not good idea, but some times there is no time/way to do skip it. I have found a plugin and this post here. I have choose to use the second one in my project. Here is the example

class DagensAgentSweeper < ActionController::Caching::Sweeper
  observe DagensAgent

  def self.skip_callback(callback, &block)
    method = instance_method(callback)
    remove_method(callback) if respond_to?(callback)
    define_method(callback) { true }
    result = yield
    remove_method(callback)
    define_method(callback, method)
    result
  end
........
end

And later in the controller….model…

      DagensAgentSweeper.skip_callback(:after_update) do
        @agent.shown_times += 1
        @agent.save
      end

timed_fragment_cache

Some patches on the timed_fragment_cache plugin.

Note how Jolyon Says suggest to expire the cache when you have time zone : when_fragment_expired ‘tags_home_page’, Time.now + 10.minutes do

I have notice that timed_fragment_cache (or the patch) have a problem with expiring the cache from the controller. In the documenation it is stated that it is ok to expire the cache only in the controller with

when_fragment_expired  "_last_update",  Time.now + 1.minutes  do

and there is no need to expire it in the view, but you MUST expire it in the view.

cache "_last_update" do

must be

cache "_last_update", Time.now + 1.minutes  do

Rails action to send binary data

I have one old application where the attachment is store as an id in the filesystem. The customer wants to have the attachment with the right name. So I have made this small action which accepts the id of the requested object and send the file data.

  def send_bios
    redirect_to :action => 'bios' if params[:id].blank?
    @artist = Artist.find(params[:id]);
    contents = File.open(@artist.bios_path,"rb") {|io| io.read}
    send_data contents, :filename => "#{@artist.name}.pdf", :type => "application/pdf", :disposition => 'attachment'
  end

Here is how to make random images generation for your website.

First you must put your images in a folder /public/images/random_images or what ever.

Second you put images there and mark the filenames with a prefix, which image to which section will be visible on.

index_bottom_1.jpg -> for the index page on the bottom :)

index_bottom_2.jpg -> for the index page on the bottom

featured_art_1.jpg -> some panel.

featured_art_2.jpg -> some panel.

and so on…

Paste this method in your base controller. This method will be called to make list for a specific image folder and “filter” the images that are supposed to be showed in the different regions in @variables.

  before_filter :init_random_images
  protected # :) be carefull for this if you copy paste
  def init_random_images
    # collect some static images
    web_root = "images/imagestorage/random_images"

    all_images = Dir["#{RAILS_ROOT}/public/#{web_root}/*.jpg"]

    @random_head_images = []
    @random_indexbtm = []
    @random_featured = []
    @random_sales = []
    @random_services = []
    @random_framing = []
    @random_artistcollection = []
    all_images.each do |f|
      if File.stat(f).file?
        @random_head_images << ("/#{web_root}/" + File.basename(f)) if f =~ /head_images_/
        @random_indexbtm << ("/#{web_root}/" + File.basename(f)) if f =~ /indexbtm_/
        @random_featured << ("/#{web_root}/" + File.basename(f)) if f =~ /featured_/
        @random_sales << ("/#{web_root}/" + File.basename(f)) if f =~ /sales_/
        @random_services << ("/#{web_root}/" + File.basename(f)) if f =~ /services_/
        @random_framing << ("/#{web_root}/" + File.basename(f)) if f =~ /framing_/
        @random_artistcollection << ("/#{web_root}/" + File.basename(f)) if f =~ /artistcollection_/
      end
    end

Then paste this as a helper or in the base controller, to use it in your views. The method will give you from the images array, count number of images.

  helper_method :get_random_image
  def get_random_image(images, count)
    requested_images = []
    images_copy = images.clone
    [images_copy.size, count].min.times do
      index = rand(images_copy.size)
      requested_images << images_copy[index]
      images_copy.delete_at(index)
    end
    requested_images.flatten.compact
  end

here is an example how to use it.

<div id="static_logo" style="background:url(<%= get_random_image(@random_sales, 1) %>) no-repeat scroll left top;">
 

Future optimization

Put some caching, not to make

Dir["#{RAILS_ROOT}/public/#{web_root}/*.jpg"]
 

on every request.

10x Bl8cki to the nice code optimizations :)