assembly-objectfile 1.12.0 → 2.1.0

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: assembly-objectfile
 version: !ruby/object:Gem::Version
-  version: 1.12.0
+  version: 2.1.0
 platform: ruby
 authors:
 - Peter Mangiafico
@@ -11,7 +11,7 @@ authors:
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2022-02-18 00:00:00.000000000 Z
+date: 2022-07-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -27,48 +27,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 5.2.0
-- !ruby/object:Gem::Dependency
-  name: deprecation
-  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: dry-struct
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.0'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.0'
-- !ruby/object:Gem::Dependency
-  name: dry-types
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.1'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - "~>"
-      - !ruby/object:Gem::Version
-        version: '1.1'
 - !ruby/object:Gem::Dependency
   name: mime-types
   requirement: !ruby/object:Gem::Requirement
@@ -98,21 +56,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: nokogiri
-  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: json
+  name: pry-byebug
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -157,16 +101,16 @@ dependencies:
   name: rubocop
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '1.25'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '1.25'
 - !ruby/object:Gem::Dependency
   name: rubocop-rspec
   requirement: !ruby/object:Gem::Requirement
@@ -185,16 +129,16 @@ dependencies:
   name: simplecov
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: 0.17.0
+        version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
-        version: 0.17.0
+        version: '0'
 description: Get exif data, file sizes and more.
 email:
 - pmangiafico@stanford.edu
@@ -202,14 +146,15 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- ".circleci/config.yml"
 - ".github/ISSUE_TEMPLATE/bug_report.md"
 - ".github/pull_request_template.md"
 - ".gitignore"
 - ".rubocop.yml"
 - ".rubocop_todo.yml"
 - ".rvmrc.example"
-- ".travis.yml"
 - Gemfile
+- Gemfile.lock
 - LICENSE
 - README.md
 - Rakefile
@@ -218,62 +163,27 @@ files:
 - bin/run_all_tests
 - config/boot.rb
 - lib/assembly-objectfile.rb
-- lib/assembly-objectfile/content_metadata.rb
-- lib/assembly-objectfile/content_metadata/config.rb
-- lib/assembly-objectfile/content_metadata/file.rb
-- lib/assembly-objectfile/content_metadata/file_set.rb
-- lib/assembly-objectfile/content_metadata/file_set_builder.rb
-- lib/assembly-objectfile/content_metadata/nokogiri_builder.rb
-- lib/assembly-objectfile/object_file.rb
-- lib/assembly-objectfile/object_fileable.rb
-- lib/assembly-objectfile/version.rb
-- profiles/AdobeRGB1998.icc
-- profiles/DotGain20.icc
-- profiles/sRGBIEC6196621.icc
-- spec/content_metadata_spec.rb
-- spec/object_file_spec.rb
+- lib/assembly/object_file.rb
+- lib/assembly/object_file/version.rb
+- spec/assembly/object_file_spec.rb
 - spec/spec_helper.rb
+- spec/test_data/empty.txt
 - spec/test_data/input/.empty
 - spec/test_data/input/file_with_no_exif.xml
-- spec/test_data/input/oo000oo0001/00/oo000oo0001_00_001.tif
-- spec/test_data/input/oo000oo0001/00/oo000oo0001_00_002.tif
-- spec/test_data/input/oo000oo0001/05/oo000oo0001_05_001.jp2
-- spec/test_data/input/oo000oo0001/05/oo000oo0001_05_002.jp2
-- spec/test_data/input/oo000oo0001/15/oo000oo0001_15_001.pdf
-- spec/test_data/input/oo000oo0001/15/oo000oo0001_15_002.pdf
-- spec/test_data/input/oo000oo0001/31/oo000oo0001_31_001.pdf
-- spec/test_data/input/oo000oo0001/50/oo000oo0001_50_001.tif
-- spec/test_data/input/oo000oo0001/oo000oo0001_book.pdf
-- spec/test_data/input/res1_image1.jp2
 - spec/test_data/input/res1_image1.tif
-- spec/test_data/input/res1_image2.jp2
-- spec/test_data/input/res1_image2.tif
-- spec/test_data/input/res1_teifile.txt
 - spec/test_data/input/res1_textfile.txt
 - spec/test_data/input/res1_transcript.pdf
-- spec/test_data/input/res2_image1.jp2
-- spec/test_data/input/res2_image1.tif
-- spec/test_data/input/res2_image2.jp2
-- spec/test_data/input/res2_image2.tif
-- spec/test_data/input/res2_teifile.txt
-- spec/test_data/input/res2_textfile.txt
-- spec/test_data/input/res3_image1.jp2
-- spec/test_data/input/res3_image1.tif
-- spec/test_data/input/res3_teifile.txt
 - spec/test_data/input/someobject.obj
 - spec/test_data/input/someobject.ply
 - spec/test_data/input/test.jp2
 - spec/test_data/input/test.json
-- spec/test_data/input/test.pdf
-- spec/test_data/input/test.svg
 - spec/test_data/input/test.tif
-- spec/test_data/input/test2.jp2
-- spec/test_data/input/test2.tif
 - spec/test_data/input/test_no_color_profile.tif
 homepage: https://github.com/sul-dlss/assembly-objectfile
 licenses:
 - ALv2
-metadata: {}
+metadata:
+  rubygems_mfa_required: 'true'
 post_install_message:
 rdoc_options: []
 require_paths:
@@ -282,10 +192,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.5'
-  - - "<"
-    - !ruby/object:Gem::Version
-      version: '4'
+      version: '3.0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
@@ -295,46 +202,6 @@ requirements: []
 rubygems_version: 3.2.32
 signing_key:
 specification_version: 4
-summary: Ruby immplementation of file services needed to prepare objects to be accessioned
-  in SULAIR digital library
-test_files:
-- spec/content_metadata_spec.rb
-- spec/object_file_spec.rb
-- spec/spec_helper.rb
-- spec/test_data/input/.empty
-- spec/test_data/input/file_with_no_exif.xml
-- spec/test_data/input/oo000oo0001/00/oo000oo0001_00_001.tif
-- spec/test_data/input/oo000oo0001/00/oo000oo0001_00_002.tif
-- spec/test_data/input/oo000oo0001/05/oo000oo0001_05_001.jp2
-- spec/test_data/input/oo000oo0001/05/oo000oo0001_05_002.jp2
-- spec/test_data/input/oo000oo0001/15/oo000oo0001_15_001.pdf
-- spec/test_data/input/oo000oo0001/15/oo000oo0001_15_002.pdf
-- spec/test_data/input/oo000oo0001/31/oo000oo0001_31_001.pdf
-- spec/test_data/input/oo000oo0001/50/oo000oo0001_50_001.tif
-- spec/test_data/input/oo000oo0001/oo000oo0001_book.pdf
-- spec/test_data/input/res1_image1.jp2
-- spec/test_data/input/res1_image1.tif
-- spec/test_data/input/res1_image2.jp2
-- spec/test_data/input/res1_image2.tif
-- spec/test_data/input/res1_teifile.txt
-- spec/test_data/input/res1_textfile.txt
-- spec/test_data/input/res1_transcript.pdf
-- spec/test_data/input/res2_image1.jp2
-- spec/test_data/input/res2_image1.tif
-- spec/test_data/input/res2_image2.jp2
-- spec/test_data/input/res2_image2.tif
-- spec/test_data/input/res2_teifile.txt
-- spec/test_data/input/res2_textfile.txt
-- spec/test_data/input/res3_image1.jp2
-- spec/test_data/input/res3_image1.tif
-- spec/test_data/input/res3_teifile.txt
-- spec/test_data/input/someobject.obj
-- spec/test_data/input/someobject.ply
-- spec/test_data/input/test.jp2
-- spec/test_data/input/test.json
-- spec/test_data/input/test.pdf
-- spec/test_data/input/test.svg
-- spec/test_data/input/test.tif
-- spec/test_data/input/test2.jp2
-- spec/test_data/input/test2.tif
-- spec/test_data/input/test_no_color_profile.tif
+summary: Ruby implementation of file services needed to prepare objects to be accessioned
+  into the Stanford Digital Repository
+test_files: []

data/.travis.yml

@@ -1,20 +0,0 @@
-language: ruby
-rvm:
-  - 2.5.7
-  - 2.6.5
-addons:
-  apt:
-    packages:
-      - libimage-exiftool-perl
-before_script:
-  - curl -L https://codeclimate.com/downloads/test-reporter/test-reporter-latest-linux-amd64 > ./cc-test-reporter
-  - chmod +x ./cc-test-reporter
-  - ./cc-test-reporter before-build
-script:
-  - bundle exec rubocop
-  - bundle exec rake
-after_script:
-  - ./cc-test-reporter after-build --exit-code $TRAVIS_TEST_RESULT
-
-notifications:
-  email: false

data/lib/assembly-objectfile/content_metadata/config.rb

@@ -1,26 +0,0 @@
-# frozen_string_literal: true
-
-require 'dry-struct'
-require 'dry-types'
-
-module Assembly
-  class ContentMetadata
-    # Types for the configuration
-    module Types
-      include Dry.Types()
-    end
-
-    # Represents a configuration for generating the content metadata
-    class Config < Dry::Struct
-      STYLES = %w[image file book map 3d document webarchive-seed].freeze
-      READING_ORDERS = %w[ltr rtl].freeze
-      attribute :auto_labels, Types::Strict::Bool.default(true)
-      attribute :flatten_folder_structure, Types::Strict::Bool.default(false)
-      attribute :add_file_attributes, Types::Strict::Bool.default(false)
-      attribute :add_exif, Types::Strict::Bool.default(false)
-      attribute :file_attributes, Types::Strict::Hash.default({}.freeze)
-      attribute :type, Types::Strict::String.enum(*STYLES)
-      attribute :reading_order, Types::Strict::String.default('ltr').enum(*READING_ORDERS)
-    end
-  end
-end

data/lib/assembly-objectfile/content_metadata/file.rb

@@ -1,63 +0,0 @@
-# frozen_string_literal: true
-
-require 'active_support/core_ext/module/delegation'
-
-module Assembly
-  class ContentMetadata
-    # Represents a single File
-    class File
-      # default publish/preserve/shelve attributes used in content metadata
-      # if no mimetype specific attributes are specified for a given file, define some defaults, and override for specific mimetypes below
-      ATTRIBUTES_FOR_TYPE = {
-        'default' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'image/tif' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'image/tiff' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'image/jp2' => { preserve: 'no', shelve: 'yes', publish: 'yes' },
-        'image/jpeg' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'audio/wav' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'audio/x-wav' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'audio/mp3' => { preserve: 'no', shelve: 'yes', publish: 'yes' },
-        'audio/mpeg' => { preserve: 'no', shelve: 'yes', publish: 'yes' },
-        'application/pdf' => { preserve: 'yes', shelve: 'yes', publish: 'yes' },
-        'plain/text' => { preserve: 'yes', shelve: 'yes', publish: 'yes' },
-        'text/plain' => { preserve: 'yes', shelve: 'yes', publish: 'yes' },
-        'image/png' => { preserve: 'yes', shelve: 'yes', publish: 'no' },
-        'application/zip' => { preserve: 'yes', shelve: 'no', publish: 'no' },
-        'application/json' => { preserve: 'yes', shelve: 'yes', publish: 'yes' }
-      }.freeze
-
-      # @param [Symbol] bundle
-      # @param [Assembly::ObjectFile] file
-      # @param style
-      def initialize(file:, bundle: nil, style: nil)
-        @bundle = bundle
-        @file = file
-        @style = style
-      end
-
-      delegate :sha1, :md5, :provider_md5, :provider_sha1, :mimetype, :filesize, :image?, :valid_image?, to: :file
-
-      def file_id(common_path:, flatten_folder_structure:)
-        # set file id attribute, first check the relative_path parameter on the object, and if it is set, just use that
-        return file.relative_path if file.relative_path
-
-        # if the relative_path attribute is not set, then use the path attribute and check to see if we need to remove the common part of the path
-        file_id = common_path ? file.path.gsub(common_path, '') : file.path
-        file_id = ::File.basename(file_id) if flatten_folder_structure
-        file_id
-      end
-
-      def file_attributes(provided_file_attributes)
-        file.file_attributes || provided_file_attributes[mimetype] || provided_file_attributes['default'] || ATTRIBUTES_FOR_TYPE[mimetype] || ATTRIBUTES_FOR_TYPE['default']
-      end
-
-      def image_data
-        { height: file.exif.imageheight, width: file.exif.imagewidth }
-      end
-
-      private
-
-      attr_reader :file
-    end
-  end
-end

data/lib/assembly-objectfile/content_metadata/file_set.rb

@@ -1,73 +0,0 @@
-# frozen_string_literal: true
-
-require 'active_support/core_ext/object/blank'
-
-module Assembly
-  class ContentMetadata
-    # Represents a groups of related Files, such as a single master file and the derivatives
-    class FileSet
-      # @param [Boolean] dpg (false) is it a dpg bundle?
-      # @param [Array<Assembly::ObjectFile>] resource_files
-      # @param style
-      def initialize(resource_files:, style:, dpg: false)
-        @dpg = dpg
-        @resource_files = resource_files
-        @style = style
-      end
-
-      # objects in the special DPG folders are always type=object when we using :bundle=>:dpg
-      # otherwise look at the style to determine the resource_type_description
-      def resource_type_description
-        @resource_type_description ||= special_dpg_resource? ? 'object' : resource_type_descriptions
-      end
-
-      def label_from_file(default:)
-        resource_files.find { |obj| obj.label.present? }&.label || default
-      end
-
-      def files
-        resource_files.map { |file| File.new(file: file) }
-      end
-
-      private
-
-      attr_reader :dpg, :resource_files, :style
-
-      def special_dpg_resource?
-        return false unless dpg
-
-        resource_files.collect { |obj| ContentMetadata.special_dpg_folder?(obj.dpg_folder) }.include?(true)
-      end
-
-      # rubocop:disable Metrics/CyclomaticComplexity
-      def resource_type_descriptions
-        # grab all of the file types within a resource into an array so we can decide what the resource type should be
-        resource_file_types = resource_files.collect(&:object_type)
-        resource_has_non_images = !(resource_file_types - [:image]).empty?
-
-        case style
-        when :simple_image, :map, :'webarchive-seed'
-          'image'
-        when :file
-          'file'
-        when :simple_book # in a simple book project, all resources are pages unless they are *all* non-images -- if so, switch the type to object
-          resource_has_non_images && resource_file_types.include?(:image) == false ? 'object' : 'page'
-        when :book_as_image # same as simple book, but all resources are images instead of pages, unless we need to switch them to object type
-          resource_has_non_images && resource_file_types.include?(:image) == false ? 'object' : 'image'
-        when :book_with_pdf # in book with PDF type, if we find a resource with *any* non images, switch it's type from book to object
-          resource_has_non_images ? 'object' : 'page'
-        when :document
-          'document'
-        when :'3d'
-          resource_extensions = resource_files.collect(&:ext)
-          if (resource_extensions & VALID_THREE_DIMENSION_EXTENTIONS).empty? # if this resource contains no known 3D file extensions, the resource type is file
-            'file'
-          else # otherwise the resource type is 3d
-            '3d'
-          end
-        end
-      end
-      # rubocop:enable Metrics/CyclomaticComplexity
-    end
-  end
-end

data/lib/assembly-objectfile/content_metadata/file_set_builder.rb

@@ -1,65 +0,0 @@
-# frozen_string_literal: true
-
-module Assembly
-  class ContentMetadata
-    # Builds a groups of related Files, based on bundle
-    class FileSetBuilder
-      # @param [Symbol] bundle one of: :default, :filename, :dpg or :prebundled
-      # @param [Array<Assembly::ObjectFile>] objects
-      # @param [Symbol] style one of: :simple_image, :file, :simple_book, :book_as_image, :book_with_pdf, :map, or :'3d'
-      def self.build(bundle:, objects:, style:)
-        new(bundle: bundle, objects: objects, style: style).build
-      end
-
-      def initialize(bundle:, objects:, style:)
-        @bundle = bundle
-        @objects = objects
-        @style = style
-      end
-
-      # @return [Array<FileSet>] a list of filesets in the object
-      def build
-        case bundle
-        when :default # one resource per object
-          objects.collect { |obj| FileSet.new(resource_files: [obj], style: style) }
-        when :filename # one resource per distinct filename (excluding extension)
-          build_for_filename
-        when :dpg # group by DPG filename
-          build_for_dpg
-        when :prebundled
-          # if the user specifies this method, they will pass in an array of arrays, indicating resources, so we don't need to bundle in the gem
-          # This is used by the assemblyWF if you have stubContentMetadata.xml
-          objects.map { |inner| FileSet.new(resource_files: inner, style: style) }
-        else
-          raise 'Invalid bundle method'
-        end
-      end
-
-      private
-
-      attr_reader :bundle, :objects, :style
-
-      def build_for_filename
-        # loop over distinct filenames, this determines how many resources we will have and
-        # create one resource node per distinct filename, collecting the relevant objects with the distinct filename into that resource
-        distinct_filenames = objects.collect(&:filename_without_ext).uniq # find all the unique filenames in the set of objects, leaving off extensions and base paths
-        distinct_filenames.map do |distinct_filename|
-          FileSet.new(resource_files: objects.collect { |obj| obj if obj.filename_without_ext == distinct_filename }.compact,
-                      style: style)
-        end
-      end
-
-      def build_for_dpg
-        # loop over distinct dpg base names, this determines how many resources we will have and
-        # create one resource node per distinct dpg base name, collecting the relevant objects with the distinct names into that resource
-
-        distinct_filenames = objects.collect(&:dpg_basename).uniq # find all the unique DPG filenames in the set of objects
-        resources = distinct_filenames.map do |distinct_filename|
-          FileSet.new(dpg: true, resource_files: objects.collect { |obj| obj if obj.dpg_basename == distinct_filename && !ContentMetadata.special_dpg_folder?(obj.dpg_folder) }.compact, style: style)
-        end
-        objects.each { |obj| resources << FileSet.new(dpg: true, resource_files: [obj], style: style) if ContentMetadata.special_dpg_folder?(obj.dpg_folder) } # certain subfolders require individual resources for files within them regardless of file-naming convention
-        resources
-      end
-    end
-  end
-end

data/lib/assembly-objectfile/content_metadata/nokogiri_builder.rb

@@ -1,57 +0,0 @@
-# frozen_string_literal: true
-
-module Assembly
-  class ContentMetadata
-    # Builds a nokogiri representation of the content metadata
-    class NokogiriBuilder
-      # @param [Array<Fileset>] filesets
-      # @param [String] druid
-      # @param [String] common_path
-      # @param [Config] config
-      def self.build(filesets:, druid:, common_path:, config:)
-        # a counter to use when creating auto-labels for resources, with incremenets for each type
-        resource_type_counters = Hash.new(0)
-        pid = druid.gsub('druid:', '') # remove druid prefix when creating IDs
-
-        Nokogiri::XML::Builder.new do |xml|
-          xml.contentMetadata(objectId: druid.to_s, type: config.type) do
-            xml.bookData(readingOrder: config.reading_order) if config.type == 'book'
-
-            filesets.each_with_index do |fileset, index| # iterate over all the resources
-              # start a new resource element
-              sequence = index + 1
-
-              resource_type_counters[fileset.resource_type_description] += 1 # each resource type description gets its own incrementing counter
-
-              xml.resource(id: "#{pid}_#{sequence}", sequence: sequence, type: fileset.resource_type_description) do
-                # create a generic resource label if needed
-                default_label = config.auto_labels ? "#{fileset.resource_type_description.capitalize} #{resource_type_counters[fileset.resource_type_description]}" : ''
-
-                # but if one of the files has a label, use it instead
-                resource_label = fileset.label_from_file(default: default_label)
-
-                xml.label(resource_label) unless resource_label.empty?
-                fileset.files.each do |cm_file| # iterate over all the files in a resource
-                  xml_file_params = { id: cm_file.file_id(common_path: common_path, flatten_folder_structure: config.flatten_folder_structure) }
-                  xml_file_params.merge!(cm_file.file_attributes(config.file_attributes)) if config.add_file_attributes
-                  xml_file_params.merge!(mimetype: cm_file.mimetype, size: cm_file.filesize) if config.add_exif
-
-                  xml.file(xml_file_params) do
-                    if config.add_exif # add exif info if the user requested it
-                      xml.checksum(cm_file.sha1, type: 'sha1')
-                      xml.checksum(cm_file.md5, type: 'md5')
-                      xml.imageData(cm_file.image_data) if cm_file.valid_image? # add image data for an image
-                    elsif cm_file.provider_md5 || cm_file.provider_sha1 # if we did not add exif info, see if there are user supplied checksums to add
-                      xml.checksum(cm_file.provider_sha1, type: 'sha1') if cm_file.provider_sha1
-                      xml.checksum(cm_file.provider_md5, type: 'md5') if cm_file.provider_md5
-                    end
-                  end
-                end
-              end
-            end
-          end
-        end
-      end
-    end
-  end
-end