moab-versioning 4.4.2 → 5.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4b576c73187205ae4548c6460bac2b68cedab5dc7c3b94a66a5d8ec13729c6ae
-  data.tar.gz: e714a158d2e1388dec94482c1fa858a2ee29a6eca6e0f9004658446801088476
+  metadata.gz: 6e9810b925a8e475771162b8f367bea2555e1a9e663993040f84256d1897a1bc
+  data.tar.gz: b329ea89a3b4ba5705cb63a5201d35125772d2dc872942b0df7cf62fc3c260c6
 SHA512:
-  metadata.gz: 251925f9e4c45f20f6ae4270fc73adfe9f4645689bda08ebd42c2b707909177ab575e0427c83e321c2ebc039c735f9bc141b29037f28cb65249c25bd9e76d520
-  data.tar.gz: '053391d53a3a3a46bbfedce8e7a04ea0f28a7f5c43689c2bde3c1901f9303eae7356d507904ec1892d9da61f2d4c9c6e58ed3fec894985e19b9ab85dd5fa65fd'
+  metadata.gz: 399fdfdef956fbec356495e4346b7a4114c921bf84df1f24bb1595795d045c6ef2ad5618b148c53364375ef7178fd99ac1be57a9aac4c9a2e184334387031d57
+  data.tar.gz: 29610e9c9a72e57ca5340e5685d6582897bda53f79904bdeccc0bc62f485827eb72eff4afb86ca2258bce47052f56f204c8ec987b94bef55579002fa045caa3c

data/lib/moab/file_inventory_difference.rb

@@ -100,10 +100,10 @@ module Moab
     # @param (see #compare)
     # @return [String] Returns either the common digitial object ID, or a concatenation of both inventory's IDs
     def common_object_id(basis_inventory, other_inventory)
-      if basis_inventory.digital_object_id != other_inventory.digital_object_id
-        "#{basis_inventory.digital_object_id}|#{other_inventory.digital_object_id}"
-      else
+      if basis_inventory.digital_object_id == other_inventory.digital_object_id
         basis_inventory.digital_object_id.to_s
+      else
+        "#{basis_inventory.digital_object_id}|#{other_inventory.digital_object_id}"
       end
     end
 

data/lib/moab/file_signature.rb

@@ -77,7 +77,7 @@ module Moab
     def self.from_file(pathname, algos_to_use = active_algos)
       raise(MoabRuntimeError, 'Unrecognized algorithm requested') unless algos_to_use.all? { |a| KNOWN_ALGOS.include?(a) }
 
-      signatures = algos_to_use.map { |k| [k, KNOWN_ALGOS[k].call] }.to_h
+      signatures = algos_to_use.to_h { |k| [k, KNOWN_ALGOS[k].call] }
 
       pathname.open("r") do |stream|
         while (buffer = stream.read(8192))

data/lib/moab/signature_catalog.rb

@@ -84,7 +84,7 @@ module Moab
 
     def block_count
       block_size = 1024
-      entries.inject(0) { |sum, entry| sum + (entry.signature.size.to_i + block_size - 1) / block_size }
+      entries.inject(0) { |sum, entry| sum + ((entry.signature.size.to_i + block_size - 1) / block_size) }
     end
 
     # @return [Array<String>] The data fields to include in summary reports

data/lib/moab/stanford.rb

@@ -2,20 +2,10 @@
 
 require 'moab'
 require 'stanford/content_inventory'
-require 'stanford/dor_metadata'
 require 'stanford/storage_repository'
 require 'stanford/storage_services'
-require 'stanford/active_fedora_object'
-require 'stanford/moab_storage_directory'
 require 'stanford/storage_object_validator'
 
 # Stanford is a module that isolates classes specific to the Stanford Digital Repository
-#
-# ====Data Model
-# * <b>{DorMetadata} = utility methods for interfacing with Stanford metadata files (esp contentMetadata)</b>
-#   * {ActiveFedoraObject} [1..*] = utility for extracting content or other information from a Fedora Instance
-#
-# @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
-#   All rights reserved.  See {file:LICENSE.rdoc} for details.
 module Stanford
 end

data/lib/moab/storage_object.rb

@@ -57,9 +57,10 @@ module Moab
 
     # @api external
     # @param bag_dir [Pathname,String] The location of the bag to be ingested
+    # @param use_links [Boolean] If true, use hard links; if false, make copies
     # @return [void] Ingest a new object version contained in a bag into this objects storage area
     # @example {include:file:spec/features/storage/ingest_spec.rb}
-    def ingest_bag(bag_dir = deposit_bag_pathname)
+    def ingest_bag(bag_dir = deposit_bag_pathname, use_links: true)
       bag_dir = Pathname(bag_dir)
       current_version = StorageObjectVersion.new(self, current_version_id)
       current_inventory = current_version.file_inventory('version')
@@ -70,7 +71,7 @@ module Moab
         new_inventory = versionize_bag(bag_dir, current_version, new_version)
       end
       validate_new_inventory(new_inventory)
-      new_version.ingest_bag_data(bag_dir)
+      new_version.ingest_bag_data(bag_dir, use_links: use_links)
       new_version.update_catalog(current_version.signature_catalog, new_inventory)
       new_version.generate_differences_report(current_inventory, new_inventory)
       new_version.generate_manifest_inventory

data/lib/moab/storage_object_version.rb

@@ -137,15 +137,16 @@ module Moab
 
     # @api internal
     # @param bag_dir [Pathname,String] The location of the bag to be ingested
+    # @param use_links [Boolean] If true, use hard links; if false, make copies
     # @return [void] Create the version subdirectory and move files into it
-    def ingest_bag_data(bag_dir)
+    def ingest_bag_data(bag_dir, use_links: true)
       raise(MoabRuntimeError, "Version already exists: #{@version_pathname}") if @version_pathname.exist?
 
       @version_pathname.join('manifests').mkpath
       bag_dir = Pathname(bag_dir)
-      ingest_dir(bag_dir.join('data'), @version_pathname.join('data'))
-      ingest_file(bag_dir.join(FileInventory.xml_filename('version')), @version_pathname.join('manifests'))
-      ingest_file(bag_dir.join(FileInventory.xml_filename('additions')), @version_pathname.join('manifests'))
+      ingest_dir(bag_dir.join('data'), @version_pathname.join('data'), use_links)
+      ingest_file(bag_dir.join(FileInventory.xml_filename('version')), @version_pathname.join('manifests'), use_links)
+      ingest_file(bag_dir.join(FileInventory.xml_filename('additions')), @version_pathname.join('manifests'), use_links)
     end
 
     # @api internal
@@ -286,10 +287,10 @@ module Moab
           file.instances.each do |instance|
             relative_path = File.join(group.group_id, instance.path)
             catalog_entry = signature_catalog.signature_hash[file.signature]
-            if !catalog_entry.nil?
-              found += 1
-            else
+            if catalog_entry.nil?
               missing << relative_path.to_s
+            else
+              found += 1
             end
           end
         end

data/lib/moab/storage_services.rb

@@ -83,12 +83,6 @@ module Moab
       @@repository.storage_object(object_id).current_version_id
     end
 
-    # @param [String] object_id The digital object identifier of the object
-    # @return [Pathname] Pathname object containing the full path for the specified file
-    def self.version_metadata(object_id)
-      retrieve_file('metadata', 'versionMetadata.xml', object_id)
-    end
-
     # @param [String] object_id The digital object identifier of the object
     # @param [Integer] version_id The ID of the version, if nil use latest version
     # @return [FileInventory] the file inventory for the specified object version

data/lib/moab/utc_time.rb

@@ -7,9 +7,7 @@ module Moab
     # @return [void] Convert input datetime to a Time object, or nil if input is empty.
     def self.input(datetime)
       case datetime
-      when nil
-        nil
-      when ""
+      when nil, ""
         nil
       when String
         Time.parse(datetime)

data/lib/moab/verification_result.rb

@@ -66,7 +66,7 @@ module Moab
     # @param level [Integer] Used to increment the depth of recursion
     # @return [Hash] The verification result of subentities serialized to a hash
     def subentities_to_hash(verbose, level)
-      subentities.map { |s| [s.entity, s.to_hash(verbose, level + 1)] }.to_h
+      subentities.to_h { |s| [s.entity, s.to_hash(verbose, level + 1)] }
     end
   end
 end

data/lib/moab.rb

@@ -20,10 +20,6 @@
 #       * {FileInstanceDifference} [1..*] = contains difference information at the file level
 #         * {FileSignature} [1..2] = contains the file signature(s) of two file instances being compared
 #
-# * <b>{VersionMetadata} = descriptive information about a digital object's versions</b>
-#   * {VersionMetadataEntry} [1..*] = attributes of a digital object version
-#     * {VersionMetadataEvent} [1..*] = object version lifecycle events with timestamps
-#
 # * <b>{StorageObject} = represents a digital object's repository storage location and ingest/dissemination methods</b>
 #   * {StorageObjectVersion} [1..*] = represents a version subdirectory within an object's home directory
 #     * {Bagger} [1] = utility for creating bagit packages for ingest or dissemination
@@ -48,9 +44,6 @@ require 'moab/file_instance_difference'
 require 'moab/file_group_difference_subset'
 require 'moab/file_group_difference'
 require 'moab/file_inventory_difference'
-require 'moab/version_metadata_event'
-require 'moab/version_metadata_entry'
-require 'moab/version_metadata'
 require 'moab/bagger'
 require 'moab/storage_object'
 require 'moab/storage_object_version'
@@ -59,4 +52,3 @@ require 'moab/storage_services'
 require 'moab/exceptions'
 require 'moab/verification_result'
 require 'moab/storage_object_validator'
-require 'moab/deposit_bag_validator'

data/lib/serializer/serializable.rb

@@ -157,7 +157,7 @@ module Serializer
           diff[k] = if left[k].is_a?(Hash) && right[k].is_a?(Hash)
                       deep_diff(ltag, left[k], rtag, right[k])
                     else
-                      Hash.[](ltag, left[k], rtag, right[k])
+                      { ltag => left[k], rtag => right[k] }
                     end
         end
       end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: moab-versioning
 version: !ruby/object:Gem::Version
-  version: 4.4.2
+  version: 5.1.0
 platform: ruby
 authors:
 - Darren Weber
@@ -11,7 +11,7 @@ authors:
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-06-01 00:00:00.000000000 Z
+date: 2022-07-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: druid-tools
@@ -69,20 +69,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: coveralls
-  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: equivalent-xml
   requirement: !ruby/object:Gem::Requirement
@@ -167,6 +153,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '2.1'
+- !ruby/object:Gem::Dependency
+  name: simplecov
+  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'
 description: Contains classes to process digital object version content and metadata
 email:
 - darren.weber@stanford.edu
@@ -177,7 +177,6 @@ files:
 - lib/moab.rb
 - lib/moab/bagger.rb
 - lib/moab/config.rb
-- lib/moab/deposit_bag_validator.rb
 - lib/moab/exceptions.rb
 - lib/moab/file_group.rb
 - lib/moab/file_group_difference.rb
@@ -198,23 +197,19 @@ files:
 - lib/moab/storage_services.rb
 - lib/moab/utc_time.rb
 - lib/moab/verification_result.rb
-- lib/moab/version_metadata.rb
 - lib/moab/version_metadata_entry.rb
-- lib/moab/version_metadata_event.rb
 - lib/serializer.rb
 - lib/serializer/manifest.rb
 - lib/serializer/serializable.rb
-- lib/stanford/active_fedora_object.rb
 - lib/stanford/content_inventory.rb
-- lib/stanford/dor_metadata.rb
-- lib/stanford/moab_storage_directory.rb
 - lib/stanford/storage_object_validator.rb
 - lib/stanford/storage_repository.rb
 - lib/stanford/storage_services.rb
 homepage: https://github.com/sul-dlss/moab-versioning
 licenses:
 - Apache-2.0
-metadata: {}
+metadata:
+  rubygems_mfa_required: 'true'
 post_install_message:
 rdoc_options: []
 require_paths:
@@ -223,14 +218,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '2.6'
+      version: '2.7'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.4
+rubygems_version: 3.2.32
 signing_key:
 specification_version: 4
 summary: Ruby implementation of digital object versioning toolkit used by the SULAIR

data/lib/moab/deposit_bag_validator.rb

@@ -1,328 +0,0 @@
-# frozen_string_literal: true
-
-module Moab
-  # Given a deposit bag, ensures the contents valid for becoming a StorageObjectVersion
-  # this is a Shameless Green implementation, combining code from:
-  #  - sdr-preservation-core/lib/sdr_ingest/validate_bag  <-- old preservation robots
-  #  - archive-utils/lib/bagit_bag  <-- gem only used by sdr-preservation-robots
-  #  - archive-utils/lib/file_fixity
-  #  - archive-utils/lib/fixity
-  # this code adds duplication to this gem (see github issue #119);
-  # for example, computing checksums is done
-  #  - deposit_bag_validator
-  #  - file_signature
-  class DepositBagValidator
-    BAG_DIR_NOT_FOUND = :bag_dir_not_found
-    CHECKSUM_MISMATCH = :checksum_mismatch
-    CHECKSUM_TYPE_UNRECOGNIZED = :checksum_type_unrecognized
-    INVALID_VERSION_XXX_XML = :invalid_versionXxx_xml
-    PAYLOAD_SIZE_MISMATCH = :payload_size_mismatch
-    REQUIRED_FILE_NOT_FOUND = :required_file_not_found
-    VERSION_MISMATCH_TO_MOAB = :version_mismatch_to_moab
-    VERSION_MISSING_FROM_FILE = :version_missing_from_file
-
-    ERROR_CODE_TO_MESSAGES = {
-      BAG_DIR_NOT_FOUND => "Deposit bag directory %{bag_dir} does not exist",
-      CHECKSUM_MISMATCH => "Failed %{manifest_type} verification. Differences: \n%{diffs}",
-      CHECKSUM_TYPE_UNRECOGNIZED => "Checksum type unrecognized: %{checksum_type}; file: %{filename}",
-      INVALID_VERSION_XXX_XML => "Unable to parse %{file_pathname}: %{err_info}",
-      PAYLOAD_SIZE_MISMATCH => "Failed payload size verification. Expected: %{bag_info_sizes}; found: %{generated_sizes}",
-      REQUIRED_FILE_NOT_FOUND => "Deposit bag required file %{file_pathname} not found",
-      VERSION_MISMATCH_TO_MOAB => "Version mismatch in %{file_pathname}: Moab expected %{new_version}; found %{file_version}",
-      VERSION_MISSING_FROM_FILE => "Version xml file %{version_file} missing data at %{xpath} containing version id"
-    }.freeze
-
-    REQUIRED_MANIFEST_CHECKSUM_TYPE = 'sha256'
-    RECOGNIZED_CHECKSUM_ALGORITHMS = %i[md5 sha1 sha256 sha384 sha512].freeze
-
-    TAGMANIFEST = 'tagmanifest'
-    MANIFEST = 'manifest'
-    DATA_DIR_BASENAME = 'data'
-    BAG_INFO_TXT_BASENAME = 'bag-info.txt'
-    VERSION_ADDITIONS_BASENAME = 'versionAdditions.xml'
-    VERSION_INVENTORY_BASENAME = 'versionInventory.xml'
-    VERSION_METADATA_PATH = "#{DATA_DIR_BASENAME}/metadata/versionMetadata.xml"
-
-    REQUIRED_BAG_FILES = [
-      DATA_DIR_BASENAME,
-      'bagit.txt',
-      BAG_INFO_TXT_BASENAME,
-      "#{MANIFEST}-#{REQUIRED_MANIFEST_CHECKSUM_TYPE}.txt",
-      "#{TAGMANIFEST}-#{REQUIRED_MANIFEST_CHECKSUM_TYPE}.txt",
-      VERSION_ADDITIONS_BASENAME,
-      VERSION_INVENTORY_BASENAME,
-      VERSION_METADATA_PATH
-    ].freeze
-
-    attr_reader :deposit_bag_pathname, :expected_new_version, :result_array
-
-    def initialize(storage_object)
-      @deposit_bag_pathname = storage_object.deposit_bag_pathname
-      @expected_new_version = storage_object.current_version_id + 1
-      @result_array = []
-    end
-
-    # returns Array of tiny error hashes, allowing multiple occurrences of a single error code
-    def validation_errors
-      return [single_error_hash(BAG_DIR_NOT_FOUND, bag_dir: deposit_bag_pathname)] unless deposit_bag_pathname.exist?
-      return result_array unless required_bag_files_exist?
-
-      verify_version
-      verify_tagmanifests
-      verify_payload_size
-      verify_payload_manifests
-      result_array # attr that accumulates any errors encountered along the way
-    end
-
-    private
-
-    def bag_dir_exists?
-      deposit_bag_pathname.exist?
-    end
-
-    # assumes this is called when result_array is empty, as subsequent checks will use these required files
-    def required_bag_files_exist?
-      REQUIRED_BAG_FILES.each do |filename|
-        pathname = deposit_bag_pathname.join(filename)
-        result_array << single_error_hash(REQUIRED_FILE_NOT_FOUND, file_pathname: pathname) unless pathname.exist?
-      end
-      result_array.empty?
-    end
-
-    def verify_version
-      version_md_pathname = deposit_bag_pathname.join(VERSION_METADATA_PATH)
-      version_from_file = last_version_id_from_version_md_xml(version_md_pathname)
-      verify_version_from_xml_file(version_md_pathname, version_from_file) if version_from_file
-
-      version_additions_pathname = deposit_bag_pathname.join(VERSION_ADDITIONS_BASENAME)
-      version_from_file = version_id_from_version_manifest_xml(version_additions_pathname)
-      verify_version_from_xml_file(version_additions_pathname, version_from_file) if version_from_file
-
-      version_inventory_pathname = deposit_bag_pathname.join(VERSION_INVENTORY_BASENAME)
-      version_from_file = version_id_from_version_manifest_xml(version_inventory_pathname)
-      verify_version_from_xml_file(version_inventory_pathname, version_from_file) if version_from_file
-    end
-
-    def last_version_id_from_version_md_xml(version_md_pathname)
-      last_version_id_from_xml(version_md_pathname, '/versionMetadata/version/@versionId')
-    end
-
-    def version_id_from_version_manifest_xml(version_manifest_xml_pathname)
-      last_version_id_from_xml(version_manifest_xml_pathname, '/fileInventory/@versionId')
-    end
-
-    def last_version_id_from_xml(pathname, xpath)
-      doc = Nokogiri::XML(File.open(pathname.to_s), &:strict)
-      version_id = doc.xpath(xpath).last.text unless doc.xpath(xpath).empty?
-      return version_id.to_i if version_id
-
-      err_data = {
-        version_file: pathname,
-        xpath: xpath
-      }
-      result_array << single_error_hash(VERSION_MISSING_FROM_FILE, err_data) unless version_id
-      nil
-    rescue StandardError => e
-      err_data = {
-        file_pathname: pathname,
-        err_info: "#{e}\n#{e.backtrace}"
-      }
-      result_array << single_error_hash(INVALID_VERSION_XXX_XML, err_data)
-      nil
-    end
-
-    def verify_version_from_xml_file(file_pathname, found)
-      return if found == expected_new_version
-
-      err_data = {
-        file_pathname: file_pathname,
-        new_version: expected_new_version,
-        file_version: found
-      }
-      result_array << single_error_hash(VERSION_MISMATCH_TO_MOAB, err_data)
-    end
-
-    # adds to result_array if tagmanifest checksums don't match generated checksums
-    def verify_tagmanifests
-      tagmanifests_checksums_hash = checksums_hash_from_manifest_files(TAGMANIFEST)
-      types_to_generate = checksum_types_from_manifest_checksums_hash(tagmanifests_checksums_hash)
-      generated_checksums_hash = generate_tagmanifest_checksums_hash(types_to_generate)
-      verify_manifest_checksums(TAGMANIFEST, tagmanifests_checksums_hash, generated_checksums_hash)
-    end
-
-    # adds to result_array if manifest checksums don't match generated checksums
-    def verify_payload_manifests
-      manifests_checksums_hash = checksums_hash_from_manifest_files(MANIFEST)
-      types_to_generate = checksum_types_from_manifest_checksums_hash(manifests_checksums_hash)
-      generated_checksums_hash = generate_payload_checksums(types_to_generate)
-      verify_manifest_checksums(MANIFEST, manifests_checksums_hash, generated_checksums_hash)
-    end
-
-    # construct hash based on manifest_type-alg.txt files in bag home dir
-    # key: file_name, relative to base_path, value: hash of checksum alg => checksum value
-    def checksums_hash_from_manifest_files(manifest_type)
-      checksums_hash = {}
-      deposit_bag_pathname.children.each do |child_pathname|
-        if child_pathname.file?
-          child_fname = child_pathname.basename.to_s
-          match_result = child_fname.match("^#{manifest_type}-(.*).txt")
-          if match_result
-            checksum_type = match_result.captures.first.to_sym
-            if RECOGNIZED_CHECKSUM_ALGORITHMS.include?(checksum_type)
-              child_pathname.readlines.each do |line|
-                line.chomp!.strip!
-                checksum, file_name = line.split(/[\s*]+/, 2)
-                file_checksums = checksums_hash[file_name] || {}
-                file_checksums[checksum_type] = checksum
-                checksums_hash[file_name] = file_checksums
-              end
-            else
-              result_array << single_error_hash(CHECKSUM_TYPE_UNRECOGNIZED, checksum_type: checksum_type, filename: child_pathname)
-            end
-          end
-        end
-      end
-      checksums_hash
-    end
-
-    # generate hash of checksums by file name for bag home dir files
-    def generate_tagmanifest_checksums_hash(types_to_generate)
-      # all names in the bag home dir except those starting with 'tagmanifest'
-      home_dir_pathnames = deposit_bag_pathname.children.reject { |file| file.basename.to_s.start_with?(TAGMANIFEST) }
-      hash_with_full_pathnames = generate_checksums_hash(home_dir_pathnames, types_to_generate)
-      # return hash keys as basenames only
-      hash_with_full_pathnames.transform_keys { |k| Pathname.new(k).basename.to_s }
-    end
-
-    # generate hash of checksums by file name for bag data dir files
-    def generate_payload_checksums(types_to_generate)
-      data_pathnames = deposit_bag_pathname.join(DATA_DIR_BASENAME).find
-      hash_with_full_pathnames = generate_checksums_hash(data_pathnames, types_to_generate)
-      # return hash keys beginning with 'data/'
-      hash_with_full_pathnames.transform_keys { |k| Pathname.new(k).relative_path_from(deposit_bag_pathname).to_s }
-    end
-
-    def generate_checksums_hash(pathnames, types_to_generate)
-      file_checksums_hash = {}
-      pathnames.each do |pathname|
-        file_checksums_hash[pathname.to_s] = generated_checksums(pathname, types_to_generate) if pathname.file?
-      end
-      file_checksums_hash
-    end
-
-    def generated_checksums(pathname, types_to_generate)
-      my_digester_hash = digester_hash(types_to_generate)
-      pathname.open('r') do |stream|
-        while (buffer = stream.read(8192))
-          my_digester_hash.each_value { |digest| digest.update(buffer) }
-        end
-      end
-      file_checksums = {}
-      my_digester_hash.each do |checksum_type, digest|
-        file_checksums[checksum_type] = digest.hexdigest
-      end
-      file_checksums
-    end
-
-    def digester_hash(types_to_generate = DEFAULT_CHECKSUM_TYPES)
-      types_to_generate.each_with_object({}) do |checksum_type, digester_hash|
-        case checksum_type
-        when :md5
-          digester_hash[checksum_type] = Digest::MD5.new
-        when :sha1
-          digester_hash[checksum_type] = Digest::SHA1.new
-        when :sha256
-          digester_hash[checksum_type] = Digest::SHA2.new(256)
-        when :sha384
-          digesters[checksum_type] = Digest::SHA2.new(384)
-        when :sha512
-          digesters[checksum_type] = Digest::SHA2.new(512)
-        else
-          result_array << single_error_hash(CHECKSUM_TYPE_UNRECOGNIZED, checksum_type: checksum_type, filename: nil)
-        end
-        digester_hash
-      end
-    end
-
-    def verify_manifest_checksums(manifest_type, manifests_checksum_hash, generated_checksum_hash)
-      diff_hash = {}
-      # NOTE: this is intentionally | instead of ||
-      (manifests_checksum_hash.keys | generated_checksum_hash.keys).each do |file_name|
-        manifest_checksums = manifests_checksum_hash[file_name] || {}
-        generated_checksums = generated_checksum_hash[file_name] || {}
-        if manifest_checksums != generated_checksums
-          cdh = checksums_diff_hash(manifest_checksums, generated_checksums, manifest_type, 'generated')
-          diff_hash[file_name] = cdh if cdh
-        end
-      end
-      return if diff_hash.empty?
-
-      err_data = {
-        manifest_type: manifest_type,
-        diffs: diff_hash
-      }
-      result_array << single_error_hash(CHECKSUM_MISMATCH, err_data)
-    end
-
-    def checksums_diff_hash(left_checksums, right_checksums, left_label, right_label)
-      diff_hash = {}
-      # NOTE: these are intentionally & and | instead of && and ||
-      checksum_types_to_compare = (left_checksums.keys & right_checksums.keys)
-      checksum_types_to_compare = (left_checksums.keys | right_checksums.keys) if checksum_types_to_compare.empty?
-      checksum_types_to_compare.each do |type|
-        left_checksum = left_checksums[type]
-        right_checksum = right_checksums[type]
-        diff_hash[type] = { left_label => left_checksum, right_label => right_checksum } if left_checksum != right_checksum
-      end
-      diff_hash.empty? ? nil : diff_hash
-    end
-
-    def verify_payload_size
-      sizes_from_bag_info_file = bag_info_payload_size
-      generated_sizes = generated_payload_size
-      return if sizes_from_bag_info_file == generated_sizes
-
-      err_data = {
-        bag_info_sizes: sizes_from_bag_info_file,
-        generated_sizes: generated_sizes
-      }
-      result_array << single_error_hash(PAYLOAD_SIZE_MISMATCH, err_data)
-    end
-
-    def bag_info_payload_size
-      bag_info_txt_pathname = deposit_bag_pathname.join(BAG_INFO_TXT_BASENAME)
-      bag_info_txt_pathname.readlines.each do |line|
-        line.chomp!.strip!
-        key, value = line.split(':', 2)
-        if key.strip == 'Payload-Oxum'
-          num_bytes, num_files = value.strip.split('.') if value
-          return { bytes: num_bytes.to_i, files: num_files.to_i }
-        end
-      end
-    end
-
-    def generated_payload_size
-      payload_pathname = deposit_bag_pathname.join(DATA_DIR_BASENAME)
-      payload_pathname.find.select(&:file?).each_with_object(bytes: 0, files: 0) do |file, hash|
-        hash[:bytes] += file.size
-        hash[:files] += 1
-        hash
-      end
-    end
-
-    # checksums_hash: { fname => {:md5=>"xxx", :sha1=>"yyy"}, fname => ... }
-    def checksum_types_from_manifest_checksums_hash(checksums_hash)
-      types = []
-      checksums_hash.each_value { |v| v.each_key { |k| types << k unless types.include?(k) } }
-      types
-    end
-
-    def single_error_hash(error_code, err_data_hash)
-      { error_code => error_code_msg(error_code, err_data_hash) }
-    end
-
-    def error_code_msg(error_code, err_data_hash)
-      ERROR_CODE_TO_MESSAGES[error_code] % err_data_hash
-    end
-  end
-end

data/lib/moab/version_metadata.rb

@@ -1,34 +0,0 @@
-# frozen_string_literal: true
-
-module Moab
-  # The descriptive information about a digital object's collection of versions
-  #
-  # ====Data Model
-  # * <b>{VersionMetadata} = descriptive information about a digital object's versions</b>
-  #   * {VersionMetadataEntry} [1..*] = attributes of a digital object version
-  #     * {VersionMetadataEvent} [1..*] = object version lifecycle events with timestamps
-  #
-  # @example {include:file:spec/fixtures/data/jq937jp0017/v3/metadata/versionMetadata.xml}
-  # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
-  #   All rights reserved.  See {file:LICENSE.rdoc} for details.
-  class VersionMetadata < Serializer::Manifest
-    include HappyMapper
-
-    # The name of the XML element used to serialize this objects data
-    tag 'versionMetadata'
-
-    # (see Serializable#initialize)
-    def initialize(opts = {})
-      @versions = []
-      super(opts)
-    end
-
-    # @attribute
-    # @return [String] The digital object identifier
-    attribute :digital_object_id, String, tag: 'objectId'
-
-    # @attribute
-    # @return [Array<VersionMetadataEntry>] An array of version metadata entries, one per version
-    has_many :versions, VersionMetadataEntry, tag: 'version'
-  end
-end

data/lib/moab/version_metadata_event.rb

@@ -1,42 +0,0 @@
-# frozen_string_literal: true
-
-module Moab
-  # A container element to record object version lifecycle events with timestamps
-  #
-  # ====Data Model
-  # * {VersionMetadata} = descriptive information about a digital object's versions
-  #   * {VersionMetadataEntry} [1..*] = attributes of a digital object version
-  #     * <b>{VersionMetadataEvent} [1..*] = object version lifecycle events with timestamps</b>
-  #
-  # @see VersionMetadata
-  # @see VersionMetadataEntry
-  # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
-  #   All rights reserved.  See {file:LICENSE.rdoc} for details.
-  class VersionMetadataEvent < Serializer::Serializable
-    include HappyMapper
-
-    # The name of the XML element used to serialize this objects data
-    tag 'event'
-
-    # (see Serializable#initialize)
-    def initialize(opts = {})
-      super(opts)
-    end
-
-    # @attribute
-    # @return [String] The type of event
-    attribute :type, String
-
-    # @attribute
-    # @return [String] The date and time of an event
-    attribute :datetime, String
-
-    def datetime=(event_datetime)
-      @datetime = Moab::UtcTime.input(event_datetime)
-    end
-
-    def datetime
-      Moab::UtcTime.output(@datetime)
-    end
-  end
-end

data/lib/stanford/active_fedora_object.rb

@@ -1,30 +0,0 @@
-# frozen_string_literal: true
-
-module Stanford
-  # Utility Class for extracting content or other information from a Fedora Instance
-  #
-  # ====Data Model
-  # * {DorMetadata} = utility methods for interfacing with Stanford metadata files (esp contentMetadata)
-  #   * {ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparisons
-  #   * <b>{ActiveFedoraObject} [1..*] = utility for extracting content or other information from a Fedora Instance</b>
-  #
-  # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
-  #   All rights reserved.  See {file:LICENSE.rdoc} for details.
-  class ActiveFedoraObject
-    # @param fedora_object [Object]  The Active Fedora representation of the Fedora Object
-    # @return [Stanford::ActiveFedoraObject] Create a u
-    def initialize(fedora_object)
-      @fedora_object = fedora_object
-    end
-
-    # @return [Object]  The Active Fedora representation of the Fedora Object
-    attr_accessor :fedora_object
-
-    # @api external
-    # @param ds_id [String] The datastream identifier
-    # @return [String] The content of the specified datastream
-    def get_datastream_content(ds_id)
-      @fedora_object.datastreams[ds_id].content
-    end
-  end
-end

data/lib/stanford/dor_metadata.rb

@@ -1,44 +0,0 @@
-# frozen_string_literal: true
-
-module Stanford
-  # Stanford-specific utility methods for interfacing with DOR metadata files
-  #
-  # ====Data Model
-  # * <b>{DorMetadata} = utility methods for interfacing with Stanford metadata files (esp contentMetadata)</b>
-  #   * {ContentInventory} [1..1] = utilities for transforming contentMetadata to versionInventory and doing comparisons
-  #   * {ActiveFedoraObject} [1..*] = utility for extracting content or other information from a Fedora Instance
-  #
-  # @note Copyright (c) 2012 by The Board of Trustees of the Leland Stanford Junior University.
-  #   All rights reserved.  See {file:LICENSE.rdoc} for details.
-  class DorMetadata
-    # @return [String] The digital object identifier (druid)
-    attr_accessor :digital_object_id
-
-    # @return [Integer] \@versionId = The ordinal version number
-    attr_accessor :version_id
-
-    # @param digital_object_id [String] The digital object identifier
-    # @param version_id [Integer] The ordinal version number
-    # @return [Stanford::DorMetadata]
-    def initialize(digital_object_id, version_id = nil)
-      @digital_object_id = digital_object_id
-      @version_id = version_id
-    end
-
-    # @api internal
-    # @param directory [String] The location of the directory to be inventoried
-    # @param version_id (see #initialize)
-    # @return [FileInventory]  Inventory of the files under the specified directory
-    def inventory_from_directory(directory, version_id = nil)
-      version_id ||= @version_id
-      version_inventory = Moab::FileInventory.new(type: 'version', digital_object_id: @digital_object_id,
-                                                  version_id: version_id)
-      content_metadata = IO.read(File.join(directory, 'contentMetadata.xml'))
-      content_group = Stanford::ContentInventory.new.group_from_cm(content_metadata, 'preserve')
-      version_inventory.groups << content_group
-      metadata_group = Moab::FileGroup.new(group_id: 'metadata').group_from_directory(directory)
-      version_inventory.groups << metadata_group
-      version_inventory
-    end
-  end
-end

data/lib/stanford/moab_storage_directory.rb

@@ -1,38 +0,0 @@
-# frozen_string_literal: true
-
-require 'find'
-
-module Stanford
-  ##
-  # methods for dealing with a directory which stores Moab objects
-  class MoabStorageDirectory
-    DRUID_TREE_REGEXP = '[[:lower:]]{2}/\\d{3}/[[:lower:]]{2}/\\d{4}'
-    DRUID_REGEXP = '[[:lower:]]{2}\\d{3}[[:lower:]]{2}\\d{4}'
-
-    def self.find_moab_paths(storage_dir)
-      Find.find(storage_dir) do |path|
-        Find.prune unless File.directory?(path) # don't bother with a matching on files, we only care about directories
-        path_match_data = storage_dir_regexp(storage_dir).match(path)
-        if path_match_data
-          yield path_match_data[1], path, path_match_data # yield the druid, the full path, and the MatchData object
-          Find.prune # we don't care about what's in the moab dir, we just want the paths that look like moabs
-        end
-      end
-    end
-
-    def self.list_moab_druids(storage_dir)
-      druids = []
-      find_moab_paths(storage_dir) { |druid, _path, _path_match_data| druids << druid }
-      druids
-    end
-
-    private_class_method def self.storage_dir_regexps
-      @storage_dir_regexps ||= {}
-    end
-
-    # this regexp caching makes things many times faster (e.g. went from ~2200 s to crawl disk11, down to ~300 s)
-    private_class_method def self.storage_dir_regexp(storage_dir)
-      storage_dir_regexps[storage_dir] ||= Regexp.new("^#{storage_dir}/#{DRUID_TREE_REGEXP}/(#{DRUID_REGEXP})$")
-    end
-  end
-end