carrierwave-docsplit 0.0.1

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 130d69ad01b1df6ed26e49629fdba48deb463cc0
+  data.tar.gz: f288596517c59c64764c5a4a9b64734c75ba326c
+SHA512:
+  metadata.gz: f78aa024c446ade564da2c3b18b46025bb5d2009e178142cdccee0ce5d68cecb7f66acacd25a93d79d4ee93ea620ccf7abe2ef37a34634c7b0dbfbe97ed02e3e
+  data.tar.gz: 8c5e6679a90e2da7e8658988bdf79d02cff814c51f65ba03f14125aad8ca153fc19765f051ff2f2fdd4672c3976bef1d89e71b924efce08f271b5d5e6a97207a

data/.gitignore

@@ -0,0 +1 @@
+test/uploads

data/Gemfile

@@ -0,0 +1,4 @@
+source "http://rubygems.org"
+
+# Specify your gem's dependencies in carrierwave-docsplit.gemspec
+gemspec

data/Gemfile.lock

@@ -0,0 +1,41 @@
+PATH
+  remote: .
+  specs:
+    carrierwave-docsplit (0.0.1)
+      carrierwave
+      docsplit
+
+GEM
+  remote: http://rubygems.org/
+  specs:
+    activemodel (3.2.2)
+      activesupport (= 3.2.2)
+      builder (~> 3.0.0)
+    activerecord (3.2.2)
+      activemodel (= 3.2.2)
+      activesupport (= 3.2.2)
+      arel (~> 3.0.2)
+      tzinfo (~> 0.3.29)
+    activesupport (3.2.2)
+      i18n (~> 0.6)
+      multi_json (~> 1.0)
+    ansi (1.4.2)
+    arel (3.0.2)
+    builder (3.0.0)
+    carrierwave (0.6.0)
+      activemodel (>= 3.2.0)
+      activesupport (>= 3.2.0)
+    docsplit (0.6.3)
+    i18n (0.6.0)
+    multi_json (1.2.0)
+    turn (0.9.4)
+      ansi
+    tzinfo (0.3.32)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  activerecord
+  carrierwave-docsplit!
+  turn

data/README.md

@@ -0,0 +1,17 @@
+# CarrierWave + Docsplit: A Loving Union
+
+carrierwave-docsplit is a thin wrapper around docsplit that knows how to talk to carrierwave.
+
+# Usage
+
+1. Require the file and drop it into your module.
+
+  ```ruby 
+  extend CarrierWave::DocsplitIntegration 
+  ```
+
+2. Hook in the integration.
+
+  ```ruby
+    extract_images :to => :thumbs, :sizes => { :large => "300x", :medium => "500x" }
+  ```

data/Rakefile

@@ -0,0 +1,12 @@
+require "bundler/gem_tasks"
+require 'rake/testtask'
+
+task :console do
+  sh "irb -r ./test/test_uploader"
+end
+
+Rake::TestTask.new do |t|
+  t.libs << "test"
+  t.test_files = FileList['test/test*.rb']
+  t.verbose = true
+end

data/carrierwave-docsplit.gemspec

@@ -0,0 +1,26 @@
+# -*- encoding: utf-8 -*-
+$:.push File.expand_path("../lib", __FILE__)
+require "carrierwave-docsplit/version"
+
+Gem::Specification.new do |s|
+  s.name        = "carrierwave-docsplit"
+  s.version     = Carrierwave::Docsplit::VERSION
+  s.authors     = ["Justin Woodbridge"]
+  s.email       = ["jwoodbridge@me.com"]
+  s.homepage    = ""
+  s.summary     = %q{Bring together docsplit and carrierwave in a loving union.}
+  s.description = %q{Bring together docsplit and carrierwave in a loving union.}
+
+  s.rubyforge_project = "carrierwave-docsplit"
+
+  s.files         = `git ls-files`.split("\n")
+  s.test_files    = `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.executables   = `git ls-files -- bin/*`.split("\n").map{ |f| File.basename(f) }
+  s.require_paths = ["lib"]
+
+  # specify any dependencies here; for example:
+  s.add_development_dependency "turn"
+
+  s.add_runtime_dependency "docsplit"
+  s.add_runtime_dependency "carrierwave"
+end

data/lib/carrierwave-docsplit.rb

@@ -0,0 +1,127 @@
+require "carrierwave-docsplit/version"
+
+require 'carrierwave'
+require 'docsplit'
+require 'json'
+require 'pathname'
+require 'fileutils'
+
+module CarrierWave
+  module DocsplitIntegration
+
+    class NoModelError < Exception
+      def message
+        "Text extraction requires a model for the uploader to write results to."
+      end
+    end
+
+    # Setup the extraction.
+    #
+    #
+
+    def extract(options = {})
+      if options[:images]
+        self.setup_image_extraction options[:images]
+      end
+
+      if options[:text]
+        self.setup_text_extraction options[:text]
+      end
+    end
+
+    def setup_text_extraction(options)
+      self.instance_eval do
+        process :enact_text_extraction
+
+        define_method :enact_text_extraction do
+          raise NoModelError if @model.nil?
+
+          out = File.join self.store_dir, self.file.basename
+          FileUtils.mkdir_p out
+          Docsplit.extract_text self.file.path, :ocr => false, :output => out
+          text = File.read Dir.glob(File.join(out, '*.txt')).first
+
+          @model.send "#{options[:to]}=", text
+        end
+      end
+    end
+
+    def setup_image_extraction(options)
+       self.instance_eval do
+
+        define_method :output_path do
+          return nil if self.file.nil?
+          File.join self.store_dir, self.file.basename
+        end
+
+        # Latch our extraction method into the processing queue.
+        process :enact_extraction => options[:sizes]
+
+
+        # Define a reader to access the thumbnails stored on disk.
+        #
+        # Returns a hash structured like so.
+        #
+        # {
+        #   '700x700' => ['/uploads/w9/700x/w9_1.png']
+        # }
+        #
+        # This allows for accessing the sizes like this:
+        #
+        # file.thumbs['700x700']
+
+        define_method options[:to] do
+          path = File.join(self.output_path, '*')
+
+          dirs_or_files = Dir.glob(path)
+          reduced = {}
+
+          # Multiple Sizes
+          if dirs_or_files.any? { |entry| entry.match /\d+x\d*/ }
+
+            Dir.glob(path) do |dirs|
+              if dirs.is_a?(String)
+                dirs = [] << dirs
+              end
+
+              dirs.each do |dir|
+                thumbs = Dir.glob(File.join(dir, '*'))
+                key = File.basename(dir)
+                reduced[key] = thumbs
+              end
+            end
+
+          # Only as single size supplied
+          else
+            options.delete :to
+            size = options.values.first
+            reduced[size] = dirs_or_files
+          end
+
+          reduced
+        end
+
+        define_method :enact_extraction do |*args|
+          args.flatten!
+
+          # zip the args back into the hash the user wrote them in.
+          if args.size == 0 || (args.size % 2) > 0
+            raise ArgumentError, "Need an even amount of arguments (Given #{args.size} #{args.size % 2})"
+          end
+
+          sizes = args.each_slice(2).reduce({}) { |mem, pair| mem.merge({ pair.first => pair.last }) }
+
+          self.class.class_eval do
+            sizes.keys.each { |size| version size }
+          end
+
+          out = self.output_path
+
+          FileUtils.mkdir_p out
+
+          Docsplit.extract_images self.file.path, :size => sizes.values, :output => out
+        end
+      end
+    end
+  end
+end

data/lib/carrierwave-docsplit/version.rb

@@ -0,0 +1,5 @@
+module Carrierwave
+  module Docsplit
+    VERSION = "0.0.1"
+  end
+end

data/test/test_uploader.rb

@@ -0,0 +1,143 @@
+require 'carrierwave'
+require 'fileutils'
+
+ROOT = File.dirname(__FILE__)
+
+$:.unshift File.join "../lib", ROOT
+
+require 'carrierwave-docsplit'
+
+require 'minitest/unit'
+
+begin; require 'turn/autorun'; rescue LoadError; end
+
+class TestUploader < CarrierWave::Uploader::Base
+  extend CarrierWave::DocsplitIntegration
+
+  def store_dir
+    File.join(ROOT, 'uploads')
+  end
+
+  def self.file_path
+    file = File.open(File.join(ROOT, 'data/w9.pdf'))
+  end
+
+  storage :file
+
+  extract :images => { :to => :thumbs, :sizes => { :large => "300x", :medium => "500x" } }
+end
+
+class SingleSizeUploader < CarrierWave::Uploader::Base
+  extend CarrierWave::DocsplitIntegration
+
+  def store_dir
+    File.join(ROOT, 'uploads')
+  end
+
+  storage :file
+
+  def self.sizes
+    { :large => "300x" }
+  end
+
+  def self.file_path
+    file = File.open(File.join(ROOT, 'data/w9_single.pdf'))
+  end
+
+  image_options = {:to => :thumbs }.merge self.sizes
+
+  extract :images => image_options
+end
+
+class TextExtractionUploader < CarrierWave::Uploader::Base
+  extend CarrierWave::DocsplitIntegration
+
+  def store_dir
+    File.join(ROOT, 'uploads')
+  end
+
+  storage :file
+
+  extract :text => { :to => :tail }
+end
+
+class Pig
+  extend CarrierWave::Mount
+  attr_accessor :tail
+end
+
+TEST_OUTPUT_PATH = File.join ROOT, 'uploads/w9'
+SINGLE_OUTPUT_PATH = File.join ROOT, 'uploads/w9_single'
+
+class TestCarrierWaveDocsplit < MiniTest::Unit::TestCase
+  include CarrierWave::DocsplitIntegration
+
+  def store_dir
+    File.join(ROOT, 'uploads')
+  end
+
+  def setup
+    @uploader = TestUploader.new
+    @uploader.retrieve_from_store! File.join(ROOT, 'w9.pdf')
+
+    CarrierWave.configure do |config|
+      config.root = ROOT
+    end
+  end
+
+  def extracted_images_exist?(uploader, output_path)
+    File.exist?(output_path) && Dir.glob(File.join(output_path, "*")).any?
+  end
+
+  def test_that_read_accessor_is_being_generated
+    assert @uploader.respond_to? :thumbs
+  end
+
+  def test_that_reader_returns_valid_hash
+    if extracted_images_exist? @uploader, TEST_OUTPUT_PATH
+      @uploader.retrieve_from_store!('w9.pdf')
+    else
+      file = File.open(TestUploader.file_path)
+      @uploader.store! file
+    end
+
+    thumbs = @uploader.thumbs
+
+    assert thumbs.include?('300x'), "Thumbs does not include 300x"
+    assert thumbs.include?('500x'), "Thumbs does not include 500x"
+
+    assert thumbs.values.all? { |val| val.is_a? Array }
+  end
+
+  def test_that_output_path_returns_nil_if_no_file_stored
+    uploader = TestUploader.new
+    assert_equal nil, uploader.output_path
+  end
+
+  def test_should_handle_one_size_gracefully
+    uploader = SingleSizeUploader.new
+
+    if extracted_images_exist? uploader, SINGLE_OUTPUT_PATH
+      uploader.retrieve_from_store! 'w9_single.pdf'
+    else
+      file = File.open SingleSizeUploader.file_path
+      uploads.store! file
+    end
+
+    assert uploader.thumbs.include?(SingleSizeUploader.sizes.values.first)
+  end
+
+  def test_that_text_extraction_should_raise_error_if_no_model
+    assert_raises NoModelError do
+      uploader = TextExtractionUploader.new
+      uploader.enact_text_extraction
+    end
+  end
+
+  def test_that_text_is_assigned_to_chosen_attribute
+    Pig.mount_uploader :description, TextExtractionUploader
+    pig = Pig.new
+    pig.description = File.open(File.join(ROOT, 'data/w9.pdf'))
+    assert_equal pig.tail, File.read(File.join(ROOT,'uploads/w9/w9.txt'))
+  end
+end

metadata

@@ -0,0 +1,99 @@
+--- !ruby/object:Gem::Specification
+name: carrierwave-docsplit
+version: !ruby/object:Gem::Version
+  version: 0.0.1
+platform: ruby
+authors:
+- Justin Woodbridge
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2013-07-28 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: turn
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: docsplit
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: carrierwave
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+description: Bring together docsplit and carrierwave in a loving union.
+email:
+- jwoodbridge@me.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- .gitignore
+- Gemfile
+- Gemfile.lock
+- README.md
+- Rakefile
+- carrierwave-docsplit.gemspec
+- lib/carrierwave-docsplit.rb
+- lib/carrierwave-docsplit/version.rb
+- test/data/w9.pdf
+- test/data/w9_single.pdf
+- test/test_uploader.rb
+homepage: ''
+licenses: []
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: carrierwave-docsplit
+rubygems_version: 2.0.0
+signing_key: 
+specification_version: 4
+summary: Bring together docsplit and carrierwave in a loving union.
+test_files:
+- test/data/w9.pdf
+- test/data/w9_single.pdf
+- test/test_uploader.rb