assembly-objectfile 1.8.3 → 1.8.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9fe45aaa147dfaac9e82820d38af8de6247116031979ff68f711954af0ff9bc8
-  data.tar.gz: 4dca70b7bf315d8520b523cfd516d78211201c5092dd36b44d8cf855ccf6247f
+  metadata.gz: 55d4581c113766a963826ca6e77b15edcab495dfe1a58d029515ac63f71dab61
+  data.tar.gz: 6345774179c1af7b1bb179dcb079c79f0c48d15418d15b6c9a8010d2a81d66a8
 SHA512:
-  metadata.gz: 4573713a41500722359af8a99e4972fcd6e7e566a864ddbc277bfecf6228bae9b2ea7194550ce43278189b13480e0e380ee72c1d7690305367f497442c87bf02
-  data.tar.gz: 5c09f8e544d3479c87caf0793bb632389bcb01a8b7c22ce369344ad99f43bd0f1f78f0b055dfe121a1f63c430c3ba8b82f1793c006e8a7d8306bb89306add142
+  metadata.gz: 834d7062c0f3ffa547d5c88e051a5148598cbd6ec0debc9b0318da8c97725e9b136ee5cacbd64b49d1cf5e0bd7481ae4eaa5061e9b247772915b821f2af1c0d3
+  data.tar.gz: 3c2b9d1496af4400a57b383de90df51b02775e6f15afb6ca869607a4c890dd489cd932f99f473fb346de7fea81121e731ab1e04fb7c58dd4f6ee77bedda37ae3

data/.github/ISSUE_TEMPLATE/bug_report.md

@@ -0,0 +1,30 @@
+---
+name: Bug report
+about: File a bug report
+title: ''
+labels: bug
+assignees: ''
+
+---
+
+**Describe the bug**
+A clear and concise description of what the bug is.
+
+**User Impact**
+A description of the impact of this bug on an end-user (or your own work).
+
+**To Reproduce**
+Steps to reproduce the behavior:
+1. Go to '...'
+2. Click on '....'
+3. Scroll down to '....'
+4. See error
+
+**Expected behavior**
+A clear and concise description of what you expected to happen.
+
+**Screenshots**
+If applicable, add screenshots to help explain your problem.
+
+**Additional context**
+Add any other context about the problem here.

data/.github/pull_request_template.md

@@ -0,0 +1,5 @@
+## Why was this change made?
+
+
+
+## Was the documentation updated?

data/.rubocop.yml

@@ -1,10 +1,6 @@
 inherit_from: .rubocop_todo.yml
 require: rubocop-rspec
 
-# Configuration parameters: AllowURI, URISchemes.
-Metrics/LineLength:
-  Max: 200
-
 RSpec/ContextWording:
   Enabled: false # too dogmatic
 
@@ -15,8 +11,5 @@ RSpec/ExampleLength:
 RSpec/MessageSpies:
   Enabled: false
 
-RSpec/MultipleExpectations:
-  Max: 5
-
 RSpec/NestedGroups:
   Max: 4 # default: 3

data/.rubocop_todo.yml

@@ -1,78 +1,62 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2018-09-13 12:44:21 -0700 using RuboCop version 0.59.0.
+# on 2019-10-17 23:28:35 -0500 using RuboCop version 0.75.1.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
 # versions of RuboCop, may require this file to be generated again.
 
 # Offense count: 1
-# Cop supports --auto-correct.
-# Configuration parameters: EnforcedStyleAlignWith, AutoCorrect, Severity.
-# SupportedStylesAlignWith: keyword, variable, start_of_line
-Layout/EndAlignment:
+Lint/AmbiguousBlockAssociation:
   Exclude:
-    - 'lib/assembly-objectfile/content_metadata.rb'
+    - 'spec/content_metadata_spec.rb'
 
 # Offense count: 1
 Lint/UselessAssignment:
   Exclude:
     - 'config/boot.rb'
 
-# Offense count: 5
-# Configuration parameters: CheckForMethodsWithNoSideEffects.
-Lint/Void:
-  Exclude:
-    - 'spec/content_metadata_spec.rb'
-
 # Offense count: 3
 Metrics/AbcSize:
-  Max: 170
+  Max: 51
 
-# Offense count: 15
+# Offense count: 11
 # Configuration parameters: CountComments, ExcludedMethods.
 # ExcludedMethods: refine
 Metrics/BlockLength:
-  Max: 549
+  Max: 577
 
-# Offense count: 1
-# Configuration parameters: CountComments.
-Metrics/ClassLength:
-  Max: 137
-
-# Offense count: 1
+# Offense count: 3
 Metrics/CyclomaticComplexity:
-  Max: 63
-
-# Offense count: 26
-# Configuration parameters: AllowHeredoc, AllowURI, URISchemes, IgnoreCopDirectives, IgnoredPatterns.
-# URISchemes: http, https
-Metrics/LineLength:
-  Max: 304
+  Max: 14
 
-# Offense count: 1
-# Configuration parameters: CountComments.
+# Offense count: 5
+# Configuration parameters: CountComments, ExcludedMethods.
 Metrics/MethodLength:
-  Max: 131
+  Max: 30
 
 # Offense count: 1
 # Configuration parameters: CountComments.
 Metrics/ModuleLength:
-  Max: 106
+  Max: 107
 
 # Offense count: 1
-Metrics/PerceivedComplexity:
-  Max: 59
+# Configuration parameters: CountKeywordArgs.
+Metrics/ParameterLists:
+  Max: 11
 
 # Offense count: 2
+Metrics/PerceivedComplexity:
+  Max: 12
+
+# Offense count: 1
 # Configuration parameters: ExpectMatchingDefinition, Regex, IgnoreExecutableScripts, AllowedAcronyms.
 # AllowedAcronyms: CLI, DSL, ACL, API, ASCII, CPU, CSS, DNS, EOF, GUID, HTML, HTTP, HTTPS, ID, IP, JSON, LHS, QPS, RAM, RHS, RPC, SLA, SMTP, SQL, SSH, TCP, TLS, TTL, UDP, UI, UID, UUID, URI, URL, UTF8, VM, XML, XMPP, XSRF, XSS
 Naming/FileName:
   Exclude:
-    - 'assembly-objectfile.gemspec'
     - 'lib/assembly-objectfile.rb'
 
-# Offense count: 2
+# Offense count: 1
 # Configuration parameters: NamePrefix, NamePrefixBlacklist, NameWhitelist, MethodDefinitionMacros.
 # NamePrefix: is_, has_, have_
 # NamePrefixBlacklist: is_, has_, have_
@@ -81,7 +65,6 @@ Naming/FileName:
 Naming/PredicateName:
   Exclude:
     - 'spec/**/*'
-    - 'lib/assembly-objectfile/content_metadata.rb'
     - 'lib/assembly-objectfile/object_fileable.rb'
 
 # Offense count: 8
@@ -97,13 +80,13 @@ RSpec/FilePath:
     - 'spec/content_metadata_spec.rb'
     - 'spec/object_file_spec.rb'
 
-# Offense count: 60
+# Offense count: 67
 # Configuration parameters: AssignmentOnly.
 RSpec/InstanceVariable:
   Exclude:
     - 'spec/object_file_spec.rb'
 
-# Offense count: 28
+# Offense count: 38
 # Configuration parameters: AggregateFailuresByDefault.
 RSpec/MultipleExpectations:
   Max: 29
@@ -113,20 +96,25 @@ RSpec/RepeatedDescription:
   Exclude:
     - 'spec/object_file_spec.rb'
 
-# Offense count: 6
-Style/CommentedKeyword:
+# Offense count: 2
+RSpec/RepeatedExample:
   Exclude:
-    - 'lib/assembly-objectfile/content_metadata.rb'
+    - 'spec/object_file_spec.rb'
 
-# Offense count: 1
-Style/Documentation:
+# Offense count: 2
+Style/CommentedKeyword:
   Exclude:
-    - 'spec/**/*'
-    - 'test/**/*'
-    - 'lib/assembly-objectfile.rb'
+    - 'lib/assembly-objectfile/content_metadata.rb'
 
 # Offense count: 1
 # Configuration parameters: MinBodyLength.
 Style/GuardClause:
   Exclude:
     - 'lib/assembly-objectfile/object_file.rb'
+
+# Offense count: 346
+# Cop supports --auto-correct.
+# Configuration parameters: AutoCorrect, AllowHeredoc, AllowURI, URISchemes, IgnoreCopDirectives, IgnoredPatterns.
+# URISchemes: http, https
+Metrics/LineLength:
+  Max: 304

data/.travis.yml

@@ -1,9 +1,7 @@
-
 language: ruby
 rvm:
-  - 2.2.5
-  - 2.3.1
-  - 2.5.3
+  - 2.5.7
+  - 2.6.5
 addons:
   apt:
     packages:
@@ -13,6 +11,7 @@ before_script:
   - 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

data/Gemfile

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 source 'http://rubygems.org'
 
 # Specify your gem's dependencies in .gemspec

data/Rakefile

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'bundler/gem_tasks'
 
 desc 'Run console with irb (default), pry, etc.'

data/assembly-objectfile.gemspec

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 $LOAD_PATH.push File.expand_path('lib', __dir__)
 require 'assembly-objectfile/version'
 
@@ -12,14 +14,15 @@ Gem::Specification.new do |s|
   s.description = 'Get exif data, file sizes and more.'
   s.license     = 'ALv2'
 
-  s.rubyforge_project = 'assembly-objectfile'
-
   s.files         = `git ls-files`.split("\n")
   s.test_files    = `git ls-files -- spec/*`.split("\n")
   s.bindir        = 'exe'
   s.executables   = s.files.grep(%r{^exe/}) { |f| File.basename(f) }
   s.require_paths = ['lib']
 
+  s.add_dependency 'activesupport', '>= 5.2.0'
+  s.add_dependency 'dry-struct', '~> 1.0'
+  s.add_dependency 'dry-types', '~> 1.1'
   s.add_dependency 'mime-types', '> 3'
   s.add_dependency 'mini_exiftool'
   s.add_dependency 'nokogiri'

data/config/boot.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'rubygems'
 
 environment  = ENV['ENVIRONMENT'] ||= 'development'

data/lib/assembly-objectfile.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module Assembly
   # the path to the gem, used to access profiles stored with the gem
   PATH_TO_GEM = File.expand_path(File.dirname(__FILE__) + '/..')
@@ -14,23 +16,24 @@ module Assembly
   TRUSTED_MIMETYPES = ['text/plain', 'plain/text', 'application/pdf', 'text/html', 'application/xml'].freeze
 
   # default publish/preserve/shelve attributes used in content metadata
-  FILE_ATTRIBUTES = {}
   # if no mimetype specific attributes are specified for a given file, define some defaults, and override for specific mimetypes below
-  FILE_ATTRIBUTES['default'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['image/tif'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['image/tiff'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['image/jp2'] = { preserve: 'no', shelve: 'yes', publish: 'yes' }
-  FILE_ATTRIBUTES['image/jpeg'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['audio/wav'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['audio/x-wav'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['audio/mp3'] = { preserve: 'no', shelve: 'yes', publish: 'yes' }
-  FILE_ATTRIBUTES['audio/mpeg'] = { preserve: 'no', shelve: 'yes', publish: 'yes' }
-  FILE_ATTRIBUTES['application/pdf'] = { preserve: 'yes', shelve: 'yes', publish: 'yes' }
-  FILE_ATTRIBUTES['plain/text'] = { preserve: 'yes', shelve: 'yes', publish: 'yes' }
-  FILE_ATTRIBUTES['text/plain'] = { preserve: 'yes', shelve: 'yes', publish: 'yes' }
-  FILE_ATTRIBUTES['image/png'] = { preserve: 'yes', shelve: 'yes', publish: 'no'}
-  FILE_ATTRIBUTES['application/zip'] = { preserve: 'yes', shelve: 'no', publish: 'no' }
-  FILE_ATTRIBUTES['application/json'] = { preserve: 'yes', shelve: 'yes', publish: 'yes' }
+  FILE_ATTRIBUTES = {
+    '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
 end
 
 require 'assembly-objectfile/content_metadata'

data/lib/assembly-objectfile/content_metadata.rb

@@ -1,4 +1,12 @@
+# frozen_string_literal: true
+
 require 'nokogiri'
+require 'active_support'
+require 'assembly-objectfile/content_metadata/file'
+require 'assembly-objectfile/content_metadata/file_set'
+require 'assembly-objectfile/content_metadata/file_set_builder'
+require 'assembly-objectfile/content_metadata/config'
+require 'assembly-objectfile/content_metadata/nokogiri_builder'
 
 module Assembly
   SPECIAL_DPG_FOLDERS = %w[31 44 50].freeze # these special dpg folders will force any files contained in them into their own resources, regardless of filenaming convention
@@ -43,195 +51,69 @@ module Assembly
     #   :auto_labels = optional - Will add automated resource labels (e.g. "File 1") when labels are not provided by the user.  The default is true.
     # Example:
     #    Assembly::ContentMetadata.create_content_metadata(:druid=>'druid:nx288wh8889',:style=>:simple_image,:objects=>object_files,:add_file_attributes=>false)
-    def self.create_content_metadata(params = {})
-      druid = params[:druid]
-      objects = params[:objects]
-
-      raise 'No objects and/or druid supplied' if druid.nil? || objects.nil?
-
-      pid = druid.gsub('druid:', '') # remove druid prefix when creating IDs
-
-      style = params[:style] || :simple_image
-      bundle = params[:bundle] || :default
-      add_exif = params[:add_exif] || false
-      auto_labels = (params[:auto_labels].nil? ? true : params[:auto_labels])
-      add_file_attributes = params[:add_file_attributes] || false
-      file_attributes = params[:file_attributes] || {}
-      preserve_common_paths = params[:preserve_common_paths] || false
-      flatten_folder_structure = params[:flatten_folder_structure] || false
-      include_root_xml = params[:include_root_xml]
-
-      all_paths = []
-      objects.flatten.each do |obj|
-        raise "File '#{obj.path}' not found" unless obj.file_exists?
+    def self.create_content_metadata(druid:, objects:, auto_labels: true,
+                                     add_exif: false, bundle: :default, style: :simple_image,
+                                     add_file_attributes: false, file_attributes: {},
+                                     preserve_common_paths: false, flatten_folder_structure: false,
+                                     include_root_xml: nil)
 
-        all_paths << obj.path unless preserve_common_paths # collect all of the filenames into an array
-      end
+      common_path = find_common_path(objects) unless preserve_common_paths # find common paths to all files provided if needed
+
+      filesets = FileSetBuilder.build(bundle: bundle, objects: objects, style: style)
+
+      config = Config.new(auto_labels: auto_labels,
+                          flatten_folder_structure: flatten_folder_structure,
+                          add_file_attributes: add_file_attributes,
+                          file_attributes: file_attributes,
+                          add_exif: add_exif,
+                          type: object_level_type(style))
+
+      builder = NokogiriBuilder.build(druid: druid,
+                                      filesets: filesets,
+                                      common_path: common_path,
+                                      config: config)
+
+      result = if include_root_xml == false
+                 builder.doc.root.to_xml
+               else
+                 builder.to_xml
+               end
+
+      result
+    end
 
-      common_path = Assembly::ObjectFile.common_path(all_paths) unless preserve_common_paths # find common paths to all files provided if needed
+    def self.special_dpg_folder?(folder)
+      SPECIAL_DPG_FOLDERS.include?(folder)
+    end
 
-      # these are the valid strings for each type of document to be use contentMetadata type and resourceType
-      content_type_descriptions = { file: 'file', image: 'image', book: 'book', map: 'map', '3d': '3d' }
-      resource_type_descriptions = { object: 'object', file: 'file', image: 'image', book: 'page', map: 'image', '3d': '3d' }
+    def self.find_common_path(objects)
+      all_paths = objects.flatten.map do |obj|
+        raise "File '#{obj.path}' not found" unless obj.file_exists?
+
+        obj.path # collect all of the filenames into an array
+      end
 
-      # global sequence for resource IDs
-      sequence = 0
+      Assembly::ObjectFile.common_path(all_paths) # find common paths to all files provided if needed
+    end
+    private_class_method :find_common_path
 
-      # a counter to use when creating auto-labels for resources, with incremenets for each type
-      resource_type_counters = Hash.new(0)
+    def self.object_level_type(style)
+      puts "WARNING - the style #{style} is now deprecated and should not be used." if DEPRECATED_STYLES.include? style
 
-      # set the object level content type id
       case style
       when :simple_image
-        content_type_description = content_type_descriptions[:image]
+        'image'
       when :file
-        content_type_description = content_type_descriptions[:file]
+        'file'
       when :simple_book, :book_with_pdf, :book_as_image
-        content_type_description = content_type_descriptions[:book]
+        'book'
       when :map
-        content_type_description = content_type_descriptions[:map]
+        'map'
       when :'3d'
-        content_type_description = content_type_descriptions[:'3d']
+        '3d'
       else
         raise "Supplied style (#{style}) not valid"
       end
-
-      puts "WARNING - the style #{style} is now deprecated and should not be used." if DEPRECATED_STYLES.include? style
-
-      # determine how many resources to create
-      # setup an array of arrays, where the first array is the number of resources, and the second array is the object files containined in that resource
-      case bundle
-      when :default # one resource per object
-        resources = objects.collect { |obj| [obj] }
-      when :filename # one resource per distinct filename (excluding extension)
-        # 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
-        resources = []
-        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.each { |distinct_filename| resources << objects.collect { |obj| obj if obj.filename_without_ext == distinct_filename }.compact }
-      when :dpg # group by DPG filename
-        # 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
-        resources = []
-        distinct_filenames = objects.collect(&:dpg_basename).uniq # find all the unique DPG filenames in the set of objects
-        distinct_filenames.each do |distinct_filename|
-          resources << objects.collect { |obj| obj if obj.dpg_basename == distinct_filename && !is_special_dpg_folder?(obj.dpg_folder) }.compact
-        end
-        objects.each { |obj| resources << [obj] if is_special_dpg_folder?(obj.dpg_folder) } # certain subfolders require individual resources for files within them regardless of file-naming convention
-      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
-        resources = objects
-      else
-        raise 'Invalid bundle method'
-      end
-
-      resources.delete([]) # delete any empty elements
-
-      builder = Nokogiri::XML::Builder.new do |xml|
-        xml.contentMetadata(objectId: druid.to_s, type: content_type_description) do
-          resources.each do |resource_files| # iterate over all the resources
-            # start a new resource element
-            sequence += 1
-            resource_id = "#{pid}_#{sequence}"
-
-            # 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?
-            resource_from_special_dpg_folder = resource_files.collect { |obj| is_special_dpg_folder?(obj.dpg_folder) }.uniq
-
-            if bundle == :dpg && resource_from_special_dpg_folder.include?(true) # objects in the special DPG folders are always type=object when we using :bundle=>:dpg
-              resource_type_description = resource_type_descriptions[:object]
-            else # otherwise look at the style to determine the resource_type_description
-              case style
-              when :simple_image
-                resource_type_description = resource_type_descriptions[:image]
-              when :file
-                resource_type_description = resource_type_descriptions[: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_type_description = resource_has_non_images && resource_file_types.include?(:image) == false ? resource_type_descriptions[:object] : resource_type_descriptions[:book]
-              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_type_description = resource_has_non_images && resource_file_types.include?(:image) == false ? resource_type_descriptions[:object] : resource_type_descriptions[: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_type_description = resource_has_non_images ? resource_type_descriptions[:object] : resource_type_descriptions[:book]
-              when :map
-                resource_type_description = resource_type_descriptions[:map]
-              when :'3d'
-                 resource_extensions = resource_files.collect {|obj| obj.ext}
-                 if (resource_extensions & VALID_THREE_DIMENSION_EXTENTIONS).empty? # if this resource contains no known 3D file extensions, the resource type is file
-                   resource_type_description = resource_type_descriptions[:file]
-                 else # otherwise the resource type is 3d
-                   resource_type_description = resource_type_descriptions[:'3d']
-                 end
-               end
-            end
-
-            resource_type_counters[resource_type_description.to_sym] += 1 # each resource type description gets its own incrementing counter
-
-            xml.resource(id: resource_id, sequence: sequence, type: resource_type_description) do
-              # create a generic resource label if needed
-              resource_label = (auto_labels == true ? "#{resource_type_description.capitalize} #{resource_type_counters[resource_type_description.to_sym]}" : '')
-
-              # but if one of the files has a label, use it instead
-              resource_files.each { |obj| resource_label = obj.label unless obj.label.nil? || obj.label.empty? }
-
-              xml.label(resource_label) unless resource_label.empty?
-
-              resource_files.each do |obj| # iterate over all the files in a resource
-                mimetype = obj.mimetype if add_file_attributes || add_exif # we only need to compute the mimetype if we are adding file attributes or exif info, otherwise skip it for performance reasons
-
-                # set file id attribute, first check the relative_path parameter on the object, and if it is set, just use that
-                if obj.relative_path
-                  file_id = obj.relative_path
-                else
-                  # 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 = preserve_common_paths ? obj.path : obj.path.gsub(common_path, '')
-                  file_id = File.basename(file_id) if flatten_folder_structure
-                end
-
-                xml_file_params = { id: file_id }
-
-                if add_file_attributes
-                  file_attributes_hash = obj.file_attributes || file_attributes[mimetype] || file_attributes['default'] || Assembly::FILE_ATTRIBUTES[mimetype] || Assembly::FILE_ATTRIBUTES['default']
-                  xml_file_params.merge!(
-                    preserve: file_attributes_hash[:preserve],
-                    publish: file_attributes_hash[:publish],
-                    shelve: file_attributes_hash[:shelve],
-                    role: file_attributes_hash[:role]
-                  )
-                  xml_file_params.reject! { |_k, v| v.nil? || v.empty? }
-                end
-
-                if add_exif
-                  xml_file_params[:mimetype] = mimetype
-                  xml_file_params[:size] = obj.filesize
-                end
-                xml.file(xml_file_params) do
-                  if add_exif # add exif info if the user requested it
-                    xml.checksum(obj.sha1, type: 'sha1')
-                    xml.checksum(obj.md5, type: 'md5')
-                    xml.imageData(height: obj.exif.imageheight, width: obj.exif.imagewidth) if obj.image? # add image data for an image
-                  elsif obj.provider_md5 || obj.provider_sha1 # if we did not add exif info, see if there are user supplied checksums to add
-                    xml.checksum(obj.provider_sha1, type: 'sha1') if obj.provider_sha1
-                    xml.checksum(obj.provider_md5, type: 'md5') if obj.provider_md5
-                  end # add_exif
-                end
-              end # end resource_files.each
-            end
-          end # resources.each
-        end
-      end
-
-      result = if include_root_xml == false
-                 builder.doc.root.to_xml
-               else
-                 builder.to_xml
-               end
-
-      result
-    end # create_content_metadata
-
-    def self.is_special_dpg_folder?(folder)
-      SPECIAL_DPG_FOLDERS.include?(folder)
     end
   end # class
 end # module

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

@@ -0,0 +1,24 @@
+# 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].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)
+    end
+  end
+end

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

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+
+require 'active_support/core_ext/module/delegation'
+
+module Assembly
+  class ContentMetadata
+    # Represents a single File
+    class File
+      # @param [Symbol] bundle
+      # @param [Assembly::ObjectFile] file
+      # @param style
+      def initialize(bundle: nil, file:, style: nil)
+        @bundle = bundle
+        @file = file
+        @style = style
+      end
+
+      delegate :sha1, :md5, :provider_md5, :provider_sha1, :mimetype, :filesize, :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'] || Assembly::FILE_ATTRIBUTES[mimetype] || Assembly::FILE_ATTRIBUTES['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

@@ -0,0 +1,71 @@
+# 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] resource_files
+      # @param style
+      def initialize(dpg: false, resource_files:, style:)
+        @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
+
+      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
+          '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 :map
+          'image'
+        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
+    end
+  end
+end

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

@@ -0,0 +1,61 @@
+# frozen_string_literal: true
+
+module Assembly
+  class ContentMetadata
+    # Builds a groups of related Files, based on bundle
+    class FileSetBuilder
+      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
+          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

@@ -0,0 +1,55 @@
+# 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
+            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 |obj| # iterate over all the files in a resource
+                  xml_file_params = { id: obj.file_id(common_path: common_path, flatten_folder_structure: config.flatten_folder_structure) }
+                  xml_file_params.merge!(obj.file_attributes(config.file_attributes)) if config.add_file_attributes
+                  xml_file_params.merge!(mimetype: obj.mimetype, size: obj.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(obj.sha1, type: 'sha1')
+                      xml.checksum(obj.md5, type: 'md5')
+                      xml.imageData(obj.image_data) if obj.image? # add image data for an image
+                    elsif obj.provider_md5 || obj.provider_sha1 # if we did not add exif info, see if there are user supplied checksums to add
+                      xml.checksum(obj.provider_sha1, type: 'sha1') if obj.provider_sha1
+                      xml.checksum(obj.provider_md5, type: 'md5') if obj.provider_md5
+                    end
+                  end
+                end
+              end
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/assembly-objectfile/object_file.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 module Assembly
   # This class contains generic methods to operate on any file.
   class ObjectFile
@@ -18,9 +20,9 @@ module Assembly
       n += 1 while strings.all? { |s| s[n] && (s[n] == x[n]) }
       common_prefix = x[0...n]
       if common_prefix[-1, 1] != '/' # check if last element of the common string is the end of a directory
-        return common_prefix.split('/')[0..-2].join('/') + '/' # if not, split string along directories, and reject last one
+        common_prefix.split('/')[0..-2].join('/') + '/' # if not, split string along directories, and reject last one
       else
-        return common_prefix # if it was, then return the common prefix directly
+        common_prefix # if it was, then return the common prefix directly
       end
     end
   end

data/lib/assembly-objectfile/object_fileable.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'mini_exiftool'
 require 'mime/types'
 # require 'checksum-tools'
@@ -120,7 +122,7 @@ module Assembly
           exif_mimetype
         elsif file_mimetype # next, try exif/unix file system command
           file_mimetype
-        else  # finally, get it from the mime-types gem (using the file extension) if both of those failed for some reason
+        else # finally, get it from the mime-types gem (using the file extension) if both of those failed for some reason
           mtype = MIME::Types.type_for(path).first
           mtype ? mtype.content_type : ''
         end
@@ -148,7 +150,8 @@ module Assembly
       @exif_mimetype ||= begin
         check_for_file
         prefer_exif = !Assembly::TRUSTED_MIMETYPES.include?(file_mimetype) # if it's not a "trusted" mimetype and there is exif data; get the mimetype from the exif
-        exif.mimetype if exif && exif.mimetype && prefer_exif
+        exif.mimetype if
+exif&.mimetype && prefer_exif
       end
     end
 

data/lib/assembly-objectfile/version.rb

@@ -1,7 +1,9 @@
+# frozen_string_literal: true
+
 # Main Assembly namespace
 module Assembly
   class ObjectFile
     # Project version number
-    VERSION = '1.8.3'.freeze
+    VERSION = '1.8.4'
   end
 end

data/spec/content_metadata_spec.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'spec_helper'
 
 describe Assembly::ContentMetadata do
@@ -512,8 +514,8 @@ describe Assembly::ContentMetadata do
   end
 
   it 'generates valid content metadata for a 3d object with one 3d type files and three other supporting files (where one supporting file is a non-viewable but downloadable 3d file)' do
-    objects=[Assembly::ObjectFile.new(TEST_OBJ_FILE),Assembly::ObjectFile.new(TEST_PLY_FILE),Assembly::ObjectFile.new(TEST_TIF_INPUT_FILE),Assembly::ObjectFile.new(TEST_PDF_FILE)]
-    result = Assembly::ContentMetadata.create_content_metadata(:druid=>TEST_DRUID,:style=>:'3d',:objects=>objects)
+    objects = [Assembly::ObjectFile.new(TEST_OBJ_FILE), Assembly::ObjectFile.new(TEST_PLY_FILE), Assembly::ObjectFile.new(TEST_TIF_INPUT_FILE), Assembly::ObjectFile.new(TEST_PDF_FILE)]
+    result = described_class.create_content_metadata(druid: TEST_DRUID, style: :'3d', objects: objects)
     expect(result.class).to be String
     xml = Nokogiri::XML(result)
     expect(xml.errors.size).to be 0
@@ -600,9 +602,9 @@ describe Assembly::ContentMetadata do
 
   it 'generates an error message when an unknown style is passed in' do
     objects = []
-    expect {
+    expect do
       described_class.create_content_metadata(druid: TEST_DRUID, bundle: :prebundled, style: :borked, objects: objects)
-    }.to raise_error { |error|
+    end.to raise_error { |error|
       expect(error.message).to eq('Supplied style (borked) not valid')
     }
   end

data/spec/object_file_spec.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'spec_helper'
 
 describe Assembly::ObjectFile do

data/spec/spec_helper.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'simplecov'
 SimpleCov.start
 
@@ -52,7 +54,7 @@ TEST_RES3_TEI = File.join(TEST_INPUT_DIR, 'res3_teifile.txt')
 
 TEST_FILE_NO_EXIF = File.join(TEST_INPUT_DIR, 'file_with_no_exif.xml')
 
-TEST_OBJ_FILE=File.join(TEST_INPUT_DIR,'someobject.obj')
-TEST_PLY_FILE=File.join(TEST_INPUT_DIR,'someobject.ply')
+TEST_OBJ_FILE = File.join(TEST_INPUT_DIR, 'someobject.obj')
+TEST_PLY_FILE = File.join(TEST_INPUT_DIR, 'someobject.ply')
 
-TEST_DRUID = 'nx288wh8889'.freeze
+TEST_DRUID = 'nx288wh8889'

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: assembly-objectfile
 version: !ruby/object:Gem::Version
-  version: 1.8.3
+  version: 1.8.4
 platform: ruby
 authors:
 - Peter Mangiafico
@@ -11,8 +11,50 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-08-29 00:00:00.000000000 Z
+date: 2020-01-21 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: activesupport
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 5.2.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 5.2.0
+- !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
@@ -146,6 +188,8 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- ".github/ISSUE_TEMPLATE/bug_report.md"
+- ".github/pull_request_template.md"
 - ".gitignore"
 - ".rubocop.yml"
 - ".rubocop_todo.yml"
@@ -161,6 +205,11 @@ files:
 - 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
@@ -224,7 +273,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.4
+rubygems_version: 3.1.2
 signing_key: 
 specification_version: 4
 summary: Ruby immplementation of file services needed to prepare objects to be accessioned