Merge pull request pulibrary#398 from pulibrary/ocr_language

Set OCR Language

Gemfile

@@ -97,9 +97,11 @@ gem "omniauth-cas"
 gem 'ezid-client'
 gem 'sprockets-es6'
 gem 'sprockets-rails', '~> 2.3.3'
-gem 'browse-everything', github: 'projecthydra-labs/browse-everything'
+gem 'browse-everything', github: 'projecthydra-labs/browse-everything', branch: 'bootstrap-sprockets'
 gem 'aasm'
 gem 'newrelic_rpm'
+gem 'iso-639'
 source 'https://rails-assets.org' do
   gem 'rails-assets-babel-polyfill'
+  gem 'rails-assets-bootstrap-select', '1.9.4'
 end

Gemfile.lock

@@ -11,7 +11,8 @@ GIT
 
 GIT
   remote: git://github.com/projecthydra-labs/browse-everything.git
-  revision: 7f547a23d8fc20bd249b8f607e009b03aa112c5d
+  revision: 3c3f60472c2b31a3d5a07e8f2bf50f09f8220ed5
+  branch: bootstrap-sprockets
   specs:
     browse-everything (0.9.1)
       bootstrap-sass
@@ -325,34 +326,26 @@ GEM
     faraday (0.9.2)
       multipart-post (>= 1.2, < 3)
     ffi (1.9.10)
-    font-awesome-rails (4.4.0.0)
+    font-awesome-rails (4.5.0.0)
       railties (>= 3.2, < 5.0)
     globalid (0.3.6)
       activesupport (>= 4.1.0)
-    google-api-client (0.8.6)
-      activesupport (>= 3.2)
-      addressable (~> 2.3)
-      autoparse (~> 0.3)
-      extlib (~> 0.9)
-      faraday (~> 0.9)
-      googleauth (~> 0.3)
-      launchy (~> 2.4)
-      multi_json (~> 1.10)
-      retriable (~> 1.4)
-      signet (~> 0.6)
-    google_drive (1.0.2)
+    google-api-client (0.7.1)
+      addressable (>= 2.3.2)
+      autoparse (>= 0.3.3)
+      extlib (>= 0.9.15)
+      faraday (>= 0.9.0)
+      jwt (>= 0.1.5)
+      launchy (>= 2.1.1)
+      multi_json (>= 1.0.0)
+      retriable (>= 1.4)
+      signet (>= 0.5.0)
+      uuidtools (>= 2.1.0)
+    google_drive (1.0.5)
       google-api-client (>= 0.7.0, < 0.9)
-      minitest (>= 5.1.0)
       nokogiri (>= 1.4.4, != 1.5.2, != 1.5.1)
       oauth (>= 0.3.6)
       oauth2 (>= 0.5.0)
-    googleauth (0.4.2)
-      faraday (~> 0.9)
-      jwt (~> 1.4)
-      logging (~> 2.0)
-      memoist (~> 0.12)
-      multi_json (~> 1.11)
-      signet (~> 0.6)
     haml (4.0.7)
       tilt
     hashdiff (0.2.2)
@@ -401,6 +394,7 @@ GEM
       activesupport (>= 3.2.18)
       faraday (~> 0.9.0)
       json
+    iso-639 (0.2.5)
     jasmine-core (2.3.4)
     jasmine-jquery-rails (2.0.3)
     jasmine-rails (0.12.2)
@@ -417,7 +411,7 @@ GEM
       i18n
       logger
       rubyzip
-    jquery-rails (4.0.5)
+    jquery-rails (4.1.0)
       rails-dom-testing (~> 1.0)
       railties (>= 4.2.0)
       thor (>= 0.14, < 2.0)
@@ -460,22 +454,17 @@ GEM
       rdf-xsd (~> 1.1, >= 1.1.5)
       sparql (~> 1.99)
       sparql-client (~> 1.99)
-    little-plugger (1.1.4)
     logger (1.2.8)
-    logging (2.0.0)
-      little-plugger (~> 1.1)
-      multi_json (~> 1.10)
     loofah (2.0.3)
       nokogiri (>= 1.5.9)
     mail (2.6.3)
       mime-types (>= 1.16, < 3)
     marc (1.0.0)
       scrub_rb (>= 1.0.1, < 2)
       unf
-    memoist (0.12.0)
     method_source (0.8.2)
     mime-types (2.99)
-    mimemagic (0.3.0)
+    mimemagic (0.3.1)
     mini_magick (4.3.6)
     mini_portile (0.6.2)
     minitest (5.8.3)
@@ -563,6 +552,9 @@ GEM
       railties (= 4.2.4)
       sprockets-rails
     rails-assets-babel-polyfill (0.0.1)
+    rails-assets-bootstrap-select (1.9.4)
+      rails-assets-jquery (>= 1.8)
+    rails-assets-jquery (2.2.0)
     rails-deprecated_sanitizer (1.0.3)
       activesupport (>= 4.2.0.alpha)
     rails-dom-testing (1.0.7)
@@ -652,7 +644,7 @@ GEM
       http-cookie (>= 1.0.2, < 2.0)
       mime-types (>= 1.16, < 3.0)
       netrc (~> 0.7)
-    retriable (1.4.1)
+    retriable (2.1.0)
     rsolr (1.0.13)
       builder (>= 2.1.2)
     rspec-core (3.3.2)
@@ -705,9 +697,8 @@ GEM
       json (~> 1.0)
       redis (~> 3.2, >= 3.2.1)
       redis-namespace (~> 1.5, >= 1.5.2)
-    signet (0.6.1)
+    signet (0.7.2)
       addressable (~> 2.3)
-      extlib (~> 0.9)
       faraday (~> 0.9)
       jwt (~> 1.5)
       multi_json (~> 1.10)
@@ -782,6 +773,7 @@ GEM
     unf (0.1.4)
       unf_ext
     unf_ext (0.0.7.1)
+    uuidtools (2.1.5)
     vcr (2.9.3)
     vegas (0.1.11)
       rack (>= 1.0.0)
@@ -825,6 +817,7 @@ DEPENDENCIES
   hydra-role-management (= 0.1.0)
   hydra-works!
   iiif-presentation
+  iso-639
   jasmine-jquery-rails
   jasmine-rails
   jbuilder (~> 2.0)
@@ -845,6 +838,7 @@ DEPENDENCIES
   pul_metadata_services!
   rails (= 4.2.4)
   rails-assets-babel-polyfill!
+  rails-assets-bootstrap-select (= 1.9.4)!
   rsolr (~> 1.0.6)
   rspec-rails
   rubocop

app/assets/javascripts/application.js

@@ -19,6 +19,7 @@
 //= require blacklight/blacklight
 //= require browse_everything
 //= require nestedSortable/jquery.mjs.nestedSortable
+//= require bootstrap-select
 
 //= require_tree .
 

app/assets/javascripts/bootstrap_select_initialize.es6

@@ -0,0 +1,4 @@
+jQuery(() => {
+  $("select[multiple='multiple']").selectpicker({
+  })
+})

app/assets/javascripts/bulk_label.es6

@@ -55,7 +55,7 @@ class BulkLabeler {
   }
   track_resource_form() {
     let master = this
-    this.element.find("input.resource-radio-button").change(function() {
+    this.element.find(".resource-radio-button").change(function() {
       $("#resource-form").attr("changed", "true")
       master.check_save_button()
     })

app/assets/stylesheets/application.css.scss

@@ -43,3 +43,4 @@
 @import "browse_everything";
 
 @import "components/breadcrumb";
+@import "bootstrap-select";

app/controllers/concerns/curation_concerns/remote_metadata.rb

@@ -1,24 +1,14 @@
 module CurationConcerns::RemoteMetadata
   extend ActiveSupport::Concern
 
-  included do
-    def curation_concern
-      if wants_to_update_remote_metadata?
-        decorated_concern
-      else
-        @curation_concern
-      end
-    end
-  end
-
   private
 
-    def decorated_concern
-      decorator.new(@curation_concern)
-    end
-
     def decorator
-      UpdatesMetadata
+      if wants_to_update_remote_metadata?
+        CompositeDecorator.new(UpdatesMetadata, super)
+      else
+        super
+      end
     end
 
     def wants_to_update_remote_metadata?

app/controllers/concerns/curation_concerns/update_ocr.rb

@@ -0,0 +1,8 @@
+module CurationConcerns
+  module UpdateOCR
+    extend ActiveSupport::Concern
+    def decorator
+      ::UpdatesOCR
+    end
+  end
+end

app/controllers/curation_concerns/curation_concerns_controller.rb

@@ -3,14 +3,15 @@ class CurationConcerns::CurationConcernsController < ApplicationController
   include CurationConcerns::Collectible
   include CurationConcerns::Manifest
   include CurationConcerns::MemberManagement
+  include CurationConcerns::UpdateOCR
   include CurationConcerns::RemoteMetadata
 
   def curation_concern_name
     curation_concern.class.name.underscore
   end
 
   def update
-    authorize!(:complete, @curation_concern, message: 'Unable to mark resource complete') if @curation_concern.state != 'complete' && params[curation_concern_name][:state] == 'complete'
+    authorize!(:complete, curation_concern, message: 'Unable to mark resource complete') if curation_concern.state != 'complete' && params[curation_concern_name][:state] == 'complete'
     add_to_collections(params[curation_concern_name].delete(:collection_ids))
     super
   end
@@ -41,6 +42,17 @@ def browse_everything_files
 
   private
 
+    def curation_concern
+      @decorated_concern ||=
+        begin
+          @curation_concern = decorator.new(@curation_concern)
+        end
+    end
+
+    def decorator
+      CompositeDecorator.new(super, NullDecorator)
+    end
+
     def selected_files_params
       params[:selected_files]
     end

app/decorators/composite_decorator.rb

@@ -0,0 +1,13 @@
+class CompositeDecorator
+  attr_reader :decorators
+  def initialize(*decorators)
+    @decorators = decorators.flatten.compact
+  end
+
+  def new(obj)
+    decorators.each do |decorator|
+      obj = decorator.new(obj)
+    end
+    obj
+  end
+end

app/decorators/decorator.rb

@@ -0,0 +1,3 @@
+class Decorator < SimpleDelegator
+  delegate :class, :is_a?, :instance_of?, to: :__getobj__
+end

app/decorators/null_decorator.rb

@@ -0,0 +1,7 @@
+class NullDecorator
+  class << self
+    def new(obj)
+      obj
+    end
+  end
+end

app/decorators/updates_metadata.rb

@@ -1,6 +1,4 @@
-class UpdatesMetadata < SimpleDelegator
-  delegate :class, to: :__getobj__
-
+class UpdatesMetadata < Decorator
   def save
     apply_remote_metadata
     super

app/decorators/updates_ocr.rb

@@ -0,0 +1,18 @@
+class UpdatesOCR < Decorator
+  delegate :class, to: :__getobj__
+
+  def save
+    changed = ocr_language_changed?
+    super.tap do
+      regenerate_derivatives if changed
+    end
+  end
+
+  private
+
+    def regenerate_derivatives
+      file_set_ids.each do |f|
+        RunOCRJob.perform_later(f)
+      end
+    end
+end

app/forms/curation_concerns/curation_concerns_form.rb

@@ -1,6 +1,6 @@
 module CurationConcerns
   class CurationConcernsForm < CurationConcerns::Forms::WorkForm
-    self.terms += [:access_policy, :holding_location, :rights_statement, :rights_note, :source_metadata_identifier, :portion_note, :description, :state, :workflow_note, :collection_ids]
+    self.terms += [:access_policy, :holding_location, :rights_statement, :rights_note, :source_metadata_identifier, :portion_note, :description, :state, :workflow_note, :collection_ids, :ocr_language]
 
     def notable_rights_statement?
       RightsStatementService.notable?(model.rights_statement)

app/jobs/run_ocr_job.rb

@@ -0,0 +1,8 @@
+class RunOCRJob < ActiveJob::Base
+  queue_as :default
+
+  def perform(file_set_id)
+    file_set = FileSet.find(file_set_id)
+    OCRRunner.new(file_set).from_datastream
+  end
+end

app/models/concerns/common_metadata.rb

@@ -32,6 +32,7 @@ module CommonMetadata
     property :holding_location, predicate: ::RDF::Vocab::Bibframe.heldBy, multiple: false do |index|
       index.as :stored_searchable
     end
+    property :ocr_language, predicate: ::PULTerms.ocr_language
 
     # IIIF
     apply_schema IIIFBookSchema, ActiveFedora::SchemaIndexingStrategy.new(

app/models/file_set.rb

@@ -27,14 +27,7 @@ def create_derivatives(filename)
           url: derivative_url('intermediate_file')
         ]
       )
-      OCRCreator.create(
-        filename,
-        outputs: [
-          label: 'ocr',
-          url: ocr_file,
-          format: :hocr
-        ]
-      )
+      OCRRunner.new(self).from_file(filename)
     end
     super
   end
@@ -52,14 +45,10 @@ def ocr_file
     end
 
     def ocr_text
-      @ocr_text ||=
-        begin
-          if persisted? && File.exist?(ocr_file.gsub("file:", ""))
-            file = File.open(ocr_file.gsub("file:", ""))
-            ocr_doc = HOCRDocument.new(file)
-            ocr_doc.text.strip
-          end
-        end
+      return unless persisted? && File.exist?(ocr_file.gsub("file:", ""))
+      file = File.open(ocr_file.gsub("file:", ""))
+      ocr_doc = HOCRDocument.new(file)
+      ocr_doc.text.strip
     end
 
     # The destination_name parameter has to match up with the file parameter

app/models/vocab/pul_terms.rb

@@ -3,4 +3,5 @@ class PULTerms < RDF::StrictVocabulary('http://library.princeton.edu/terms/')
   term :exhibit_id, label: 'Exhibit ID'.freeze, type: 'rdf:Property'.freeze
   term :metadata_id, label: 'Metadata ID'.freeze, type: 'rdf:Property'.freeze
   term :source_metadata, label: 'Source Metadata'.freeze, type: 'rdf:Property'.freeze
+  term :ocr_language, label: "OCR Language".freeze, type: 'rdf:Property'.freeze
 end