shutterbug

Shutterbug

Overview

Shutterbug has two parts: a browser javascript library for taking html snapshots, and a server side utility for turning those html snapshots into images. This repository consists of JavaScript library.

Server Side Utility

A rack utility using phantomjs that will create and save images (pngs) from parts of your html's documents current dom. These images become available as public png resources in the rack application. Currently shutterbug supports HTML, SVG and Canvas elements. Here is a sample config.ru file:

use Shutterbug::Rackapp do |config|
  conf.resource_dir       = "/Users/npaessel/tmp"
  config.uri_prefix       = "http://shutterbug.herokuapp.com"
  config.path_prefix      = "/shutterbug"
  config.phantom_bin_path = "/app/vendor/phantomjs/bin/phantomjs"
end

Configuration options default to reasonable defaults.

Shutterbug is distributed as a Ruby Gem. The rack service delivers a javascript library to the browser that can send HTML fragments back to the Rack service. The Rack service generates images from these fragments using PhantomJS. In the following image “getDomSnapshot()” triggers a request to the Shutterbug service. The response from the POST request contains an image tag, that points to a newly created image on the server.

Requirements & Dependencies

• Ruby 1.9x or greater is required to run the Rack application.
• PhantomJS is requird to run the Rack application.

Installation

Add this line to your application's Gemfile:

gem 'shutterbug'

And then execute:

$ bundle

Or install it yourself as:

$ gem install shutterbug

License

Shutterbug is Copyright 2012 © by the Concord Consortium and is distributed under the MIT License.

Deploying on Heroku

To deploy on heroku, you are going to want to modify your stack following these instructions.

Your app should have a config.ru that looks something like this:

require 'shutterbug'
require 'rack/cors'

use Rack::Cors do
  allow do
    origins '*'
    resource '/shutterbug/*'', :headers => :any, :methods => [:get, :post, :options]
  end
end

# Without a complete set of S3 credentials, Shutterbug
# Places images in a temporary directory where
# you will LOSE your images...
use Shutterbug::Rackapp do |config|
  config.uri_prefix = "http://<your app name>.herokuapp.com/"
  config.path_prefix = "/shutterbug"
  config.phantom_bin_path = "/app/vendor/phantomjs/bin/phantomjs"
  # config.s3_key       = "your_S3_KEY"
  # config.s3_secret    = "your_S3_SECRET_DONT_COMMIT_IT"
  # config.s3_bin       = "your_S3_bucket_name"
end

app = proc do |env|
  [200, { 'Content-Type' => 'text/html' }, ['move along']]
end

run app

And a Procfile which looks like this:

web: bundle exec rackup config.ru -p $PORT

Changes

• December 3, 2015 – v 0.5.9

  • S3 storage provider explicitly removes temporary files.

• October 6, 2015 – v 0.5.8

• Add error_trigger handler, to verify exception reporting.

• Remove Ruby 1.9x travis tests, because of Fog and Net-SSH deps on Ruby 2.0

• September 4, 2015 – v 0.5.6

• Prevent some iFrame data-url decoding problems by inserting magical "#". For reasons we don't yet understand, iframes with base hrefs without hash (#) symbols can sometimes cause the data-urls to decode incorrectly. Our Longterm solution will be to base64 encode iFrame data-urls.

• August 25, 2015 – v 0.5.5

• More aggressive single quote replacement in phantom_job.rb

• August 14, 2015 – v 0.5.4

• Fix single quote encoding issue in phantom_job.rb. DM:

Phantom's html entity decoder borks on single quotes when you have an encoded block with single quotes embedded in another encoded block with single quotes. The decoder will decode correctly up to the first single quote but then will leave the rest of the string encoded. If the single quotes are changed to double quotes then they are properly encoded within the enclosing block.

• June 4, 2105 – v 0.5.3

  • Rasterize.js: Also loads non-base64 encoded iframe src content when walking the iFrames. Also, better mechanics for waiting for page rendering.
  • s3_storage.rb: Look for existing S3 bucket using get, to avoid 404 errors when bucket limit of 100 is hit. (github issue #15)

• May 28, 2015 – v 0.5.2

  • Rasterize.js: "Loads an html page and does a depth-first walk of its iframes. As the walk returns to the root each iframe's src is set to the base64 png rendering of its contents." The purpose of this change was to
    allow SVG document relative resource links (such as gradients) to work,
    this required the removal of the base tag to work in Phantom 2.

• December 11, 2014 – v 0.5.0

  • JS lib, JS handler and demos removed (as they are part of the new shutterbug.js repo now).

• December 4, 2014 – v 0.4.3

  • Added support of various image formats and quality settings.

• December 2, 2014 – v 0.3.0

  • Improved canvas snapshot - data is uploaded directly to S3 from the browser (no PhantomJS rendering).

• November 12, 2014 – v 0.2.5

  • Added setFailureCallback to shutterbug.js for gracefully handling ajax failures in some custom way.
  • Updated CORS configuration in config.ru to allow [:options] requests.

TODO:

• Configuration of the rack paths.
• Fix web-font bugs in phantom js.
• Better abstraction phantomjs command line invocation. Use phantomjs.rb ?
• Use sprockets for and coffee.erb for shutterbug.js
• Write Tests.

Contributing

• Join the mailing list: email or web
• Fork this project.
• Create your feature branch (git checkout -b my-new-feature)
• Commit your changes (git commit -am 'Add some feature')
• Push to the branch (git push origin my-new-feature)
• Create new Pull Request

© 2013 The Concord Consortium.