This is a rather simple gem for working the the Juicer API for resizing images. It's currently pretty rough around the edges (no tests! ack!) but it works. :)

The Job class provides two class-level methods for you to interact with the API. First, the submit method takes an API key, a source URL, and a array of output hashes. This is POSTed to the API as JSON and a hash is returned to you with the results of the API call. You'll get back a job id, the source URL and the outputs you defined, and an indication of the status ("queued" or "failed").

Here's an example request to create a job:

UploadJuicer::Job.submit("your api key",
    "http://farm3.static.flickr.com/2084/2222523486_5e1894e314.jpg",
    [{"size" => "100x100>"}])

=> {"id" => "a string", "outputs" => [{"size" => "100x100>"}], "status" => "queued"}

You can specify your own destination URLs as S3 locations:

UploadJuicer::Job.submit("your api key",
    "http://farm3.static.flickr.com/2084/2222523486_5e1894e314.jpg",
    [{"size" => "100x100>", "url" => "s3://mybucket/path/to/destfile.jpg"}])

=> {"id" => "a string", "outputs" => [{"size" => "100x100>", "url" => "s3://mybucket/path/to/destfile.jpg"}], "status" => "queued"}

When you specify output URLs, those URLs show up in the hash returned when queuing the job. For this to work, you'll need to grant access to Juicer to write to your bucket. You can find Juicer's S3 ID in the documentation at the Juicer site.

Though size is the only required key in the outputs hashes, you can add other keys (e.g., a label, or your own ids) to the hashes and they will be passed back to you. This could be useful if you aren't specifying output URLs and you want to associate some data with the URLs you'll get back.

You can use the id that was returned to you as the second argument (after your API key) to the info method. The return hash would look exactly like the hash you got back from submit, but hopefully the status will now be "finished" rather than "queued" (or "failed").

The intended usage is to have your users upload files directly to S3 via swfupload, create an UploadJuicer::Upload record via ajax once the upload is complete, and then associate that UploadJuicer::Upload record and call the Juicer API when the record your using is working with (e.g., a Project or a Contact) is created or updated.

A Rails Engine and a generator is included with this gem to make it easy to integrate Juicer with your Rails application. The engine provides an UploadJuicer::Upload model, an UploadJuicer::Uploads controller, and a helper to use in your app. The generator creates a migration, creates a configuration file, and copies swfupload in place for you to use in your forms.

To use the generator, either specify your API key that you got from the Juicer site with the --api-key option, or you specify that your API key will be loaded in your Heroku environment (if you are using Juicer as a Heroku add-on) with the --heroku option. After running the generator you'll need to run the migration and edit the config/upload_juicer.yml file to add your S3 credentials and specify which bucket you want to use for your uploads.

For swfupload to be able to upload to your bucket, you'll need to upload a crossdomain.xml to the top-level of your bucket. A sample crossdomain.xml file is placed in your public directory by the generator.

Once you are all set up, here's an example of how you'd use the gem in a Rails view and model:

<%= form_for @person do |p| %>
  <%= p.hidden_field :image_key, :id => :image_key %>
  <%= p.hidden_field :image_id, :id => :image_id %>

  <p>
    <%= p.label :name %>
    <%= p.text_field :name %>
  </p>

  <div id="file_container"></div>
  <p id="upload_placeholder" class="uploadify"></p>

  <p><%= p.submit %></p>
<% end %>

<% content_for :head do %>
  <%= stylesheet_link_tag('swfupload.css') %>
<% end %>

<% content_for :foot do %>
  <%= javascript_include_tag('swfupload.js', 'uploader.js') %>
  <script type="text/javascript">
    $('#upload_placeholder').uploadify(<%= swfupload_params %>);
  </script>
<% end %>

This view just has two fields, the name and the file that is being uploaded. the reference javascripts are copied into your public directory by the generator. Once the file is selected, the upload begins (and the form is disabled). Once the file is finished uploading, an ajax request is made to the UploadJuicer::Uploads controller to create an UploadJuicer::Upload record, the form is enabled, and the image_key and image_id fields are populated with the JSON response from the UploadJuicer::Uploads controller with info about the UploadJuicer::Upload model that was just created. This info is used when the form is submitted to associated the UploadJuicer::Upload with the Person. The Uploads helper provides the swfupload_params method, which does all the request signing, etc. that S3 requires.

By the way, for Rails 2 change that <%= form_for to <% form_for.

class Person < ActiveRecord::Base
  has_one :image, :class_name => 'UploadJuicer::Upload', :as => :uploadable

  after_save :process_image

  attr_accessor :image_key, :image_id

  def process_image
    return if @image_key.blank? || @image_id.blank?
    if self.image = UploadJuicer::Upload.first(:conditions => { :key => @image_key, :id => @image_id, :uploadable_id => nil })
      image.juice_upload(:avatar => { :size => '40x40>' }, :thumb => { :size => '100x100>' })
    end
  end
end

This model simply associates the UploadJuicer::Upload that was created via the AJAX call, then calls the juice_upload method with a hash of labels and size info to pass on to the Juicer API. The UploadJuicer::Upload model creates partitioned paths for the files in your S3 bucket like so:

Original: http://s3.amazonaws.com/your_bucket/34j/e8r/9fu/file_name.jpg
Avatar:   http://s3.amazonaws.com/your_bucket/34j/e8r/9fu/avatar/file_name.jpg
Thumb:    http://s3.amazonaws.com/your_bucket/34j/e8r/9fu/thumb/file_name.jpg

<h1><%= @person.name %></h1>
<p>Avatar: <%= image_tag(@person.image.url(:avatar)) %></p>
<p>Thumb: <%= image_tag(@person.image.url(:thumb)) %></p>
<p>Full image: <%= image_tag(@person.image.url) %></p>

This view shows how to get the publicly-readable S3 URLs from the UploadJuicer::Upload record associated with the Person.

Just for completeness' sake, here's a minimal People controller, and a sample layout that includes jquery and the two yields the new view needs:

class PeopleController < ApplicationController
  before_filter :build_person, :only => [ :new, :create ]
  helper :uploads

  def create
    @person.save
    redirect_to :action => "show", :id => @person
  end

  def show
    @person = Person.find(params[:id])
  end

  protected

    def build_person
      @person = Person.new(params[:person])
    end
end

<html>
  <head>
    <%= yield :head %>
  </head>
  <body>
    <%= yield %>
    <%= javascript_include_tag 'http://ajax.googleapis.com/ajax/libs/jquery/1.4.2/jquery.min.js' %>
    <%= yield :foot %>
  </body>
</html>