ddr-ingesttools 0.3.0 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: fb22fa65d3323b243dcafc3b999f55d42894b130
-  data.tar.gz: f85324eb9f8efd80febb1de7c9d1a5f2f87898aa
+  metadata.gz: ef121cd2d5211caf9ba9b1beebbbc316d2b7214e
+  data.tar.gz: 922b0ca3eba98b5dc5d1bc07c126c842e1a093cc
 SHA512:
-  metadata.gz: 88bbc9f8d76a0b11bc0dec289c9bba3fc54efc340b6bfe9a0d7855108b0d040e88c385facda2d5d892a8aad02a18e9f64dc4d243bc547cce4987e7e9dd71df4e
-  data.tar.gz: 5e475ae95b4d2ca9c5c8dd4b1d9fa85efc34956f4aac5b9778f62f837b495dd55a5d6d683785526b002a23d131f8caefc6170e46dd485867d5b22e099bbedbae
+  metadata.gz: 89d1fc5adacf2986a47019e0cca3da6ba2a5cf8054d72a4c1a21b82805ef0a0711abc2569f5367878aa6e8b452d8e79482d63d4ac5fdb28deef459bb3175d820
+  data.tar.gz: 24ebbe3f71fce16d96e032ba6c283e0431ba7f87a3495144ab218908052837699ff0687430eceeffcac7c31a5ab75fda7a677759b685f0583c9257c290482493

data/.gitignore

@@ -1,3 +1,5 @@
+.byebug_history
 Gemfile.lock
 pkg
 tmp
+

data/bin/mint_manifest_arks.rb

@@ -0,0 +1,49 @@
+#!/usr/bin/env ruby
+
+require 'i18n'
+require 'ddr/ingesttools'
+require 'optparse'
+
+options = {}
+
+puts I18n.t('marquee')
+puts I18n.t('suite.name')
+puts I18n.t('manifest_ark_minter.name')
+puts I18n.t('marquee')
+
+# Parse command line arguments
+parser = OptionParser.new do |opts|
+  opts.banner = 'Usage: mint_manifest_arks.rb [options]'
+
+  opts.on('-c', '--config CONFIG_FILE', 'Path to configuration file') do |v|
+    options[:config] = v
+  end
+
+  opts.on('-m', '--manifest MANIFEST_FILE', 'Path to manifest file for which ARKs are to be minted') do |v|
+    options[:manifest] = v
+  end
+
+  opts.on('-o', '--output OUTPUT_FILE', 'Path to which updated manifest file should be written') do |v|
+    options[:output] = v
+  end
+end
+
+begin
+  parser.parse!
+  if options['config'].nil?
+    puts I18n.t('manifest_ark_minter.use_default_config_file',
+                default_config_file: Ddr::IngestTools::ManifestArkMinter::ManifestUpdater::DEFAULT_CONFIG_FILE)
+  end
+  mandatory = [ :manifest, :output ]
+  missing = mandatory.select{ |param| options[param].nil? }
+  unless missing.empty?
+    raise OptionParser::MissingArgument.new(missing.join(', '))
+  end
+rescue OptionParser::InvalidOption, OptionParser::MissingArgument
+  puts $!.to_s
+  puts parser
+  exit(false)
+end
+
+updater = Ddr::IngestTools::ManifestArkMinter::ManifestUpdater.new(options)
+updater.call

data/config/locales/en.yml

@@ -1,7 +1,7 @@
 en:
-  errors:
-    count: "%{count} error(s)"
-    checksum_mismatch: |
-      ***** Checksum mismatch:
-              %{c1} %{f1}
-              %{c2} %{f2}
+  marquee: ==================================================
+  suite:
+    name: DDR Ingest Tools
+  manifest_ark_minter:
+    name: Manifest ARK Minter
+    use_default_config_file: 'Will use default configuration file: %{default_config_file}'

data/ddr-ingesttools.gemspec

@@ -18,10 +18,11 @@ Gem::Specification.new do |spec|
   spec.test_files    = spec.files.grep(%r{^(test|spec|features)/})
   spec.require_paths = ["lib"]
 
-  spec.add_dependency "bagit", "~> 0.4"
-  spec.add_dependency "i18n", "~> 0.8"
+  spec.add_dependency "ezid-client", "~> 1.7"
+  spec.add_dependency "i18n", "~> 1.0"
 
   spec.add_development_dependency "bundler", "~> 1.14"
+  spec.add_development_dependency "byebug"
   spec.add_development_dependency "rake", "~> 12.0"
   spec.add_development_dependency "rspec", "~> 3.0"
 end

data/lib/ddr/ingesttools.rb

@@ -1,6 +1,5 @@
 require_relative 'ingesttools/version'
-require_relative 'ingesttools/dpc_folder_converter'
-require_relative 'ingesttools/checksum_file'
+require_relative 'ingesttools/manifest_ark_minter'
 
 require 'i18n'
 

data/lib/ddr/ingesttools/manifest_ark_minter.rb

@@ -0,0 +1,22 @@
+require_relative 'manifest_ark_minter/configuration'
+require_relative 'manifest_ark_minter/manifest_updater'
+require_relative 'manifest_ark_minter/manifest_parser'
+require_relative 'manifest_ark_minter/minter'
+
+module Ddr::IngestTools
+  module ManifestArkMinter
+
+    class << self
+      attr_writer :configuration
+
+      def configuration
+        @configuration ||= Configuration.new
+      end
+
+      def configure
+        yield(configuration)
+      end
+    end
+
+  end
+end

data/lib/ddr/ingesttools/manifest_ark_minter/configuration.rb

@@ -0,0 +1,13 @@
+module Ddr::IngestTools::ManifestArkMinter
+  class Configuration
+
+    attr_accessor :ezid_default_shoulder, :ezid_password, :ezid_user
+
+    def initialize
+      @ezid_default_shoulder = nil
+      @ezid_password = nil
+      @ezid_user = nil
+    end
+
+  end
+end

data/lib/ddr/ingesttools/manifest_ark_minter/manifest_parser.rb

@@ -0,0 +1,33 @@
+require 'csv'
+
+module Ddr::IngestTools::ManifestArkMinter
+  class ManifestParser
+
+    attr_reader :manifest_file_path
+
+    ARK_HEADER = 'ark'
+
+    def initialize(manifest_file_path)
+      @manifest_file_path = manifest_file_path
+    end
+
+    def as_csv_table
+      @csv_table ||= CSV.read(manifest_file_path, headers: true)
+    end
+
+    def arks_missing?
+      arks.any? { |value| value.compact.empty? }
+    end
+
+    def headers
+      as_csv_table.headers
+    end
+
+    private
+
+    def arks
+      as_csv_table.values_at(ARK_HEADER)
+    end
+
+  end
+end

data/lib/ddr/ingesttools/manifest_ark_minter/manifest_updater.rb

@@ -0,0 +1,80 @@
+require 'fileutils'
+require 'logger'
+require 'tempfile'
+
+module Ddr::IngestTools::ManifestArkMinter
+  class ManifestUpdater
+
+    attr_reader :config, :logger, :manifest, :output
+
+    DEFAULT_CONFIG_FILE = 'manifest_ark_minter_config.yml'
+
+    def initialize(config: DEFAULT_CONFIG_FILE, manifest:, output:, logger: nil)
+      @config = config
+      @manifest = manifest
+      @output = output
+      @logger = logger || Logger.new(STDOUT)
+    end
+
+    def call
+      configure
+      if needs_updating?
+        update
+      else
+        logger.info("Manifest file already has ARKs ... nothing to mint")
+      end
+    end
+
+    private
+
+    def configure
+      conf = YAML::load(IO.read(config))
+      Ddr::IngestTools::ManifestArkMinter.configure do |config|
+        config.ezid_default_shoulder = conf.fetch('ezid_default_shoulder')
+        config.ezid_password = conf.fetch('ezid_password')
+        config.ezid_user = conf.fetch('ezid_user')
+      end
+    end
+
+    def update
+      update_manifest_table
+      write_updated_manifest
+    end
+
+    def update_manifest_table
+      logger.info("Minting ARKs for manifest file")
+      mint_counter = 0
+      manifest_as_csv_table.each do |row|
+        unless row['ark']
+          row['ark'] = minter.mint
+          mint_counter += 1
+        end
+      end
+      logger.info("Minted #{mint_counter} ARK(s)")
+    end
+
+    def manifest_as_csv_table
+      @manifest_as_csv_table ||= parser.as_csv_table
+    end
+
+    def write_updated_manifest
+      File.open(output, 'w') do |f|
+        f.write(manifest_as_csv_table.to_csv)
+      end
+      logger.info("Updated manifest file is at #{output}")
+    end
+
+    def needs_updating?
+      parser.arks_missing?
+    end
+
+    def minter
+      @minter ||= Minter.new
+    end
+
+    def parser
+      @parser ||= ManifestParser.new(manifest)
+    end
+
+  end
+end

data/lib/ddr/ingesttools/manifest_ark_minter/minter.rb

@@ -0,0 +1,43 @@
+require 'ezid-client'
+
+module Ddr::IngestTools::ManifestArkMinter
+  class Minter
+
+    DEFAULT_EXPORT  = 'no'.freeze
+    DEFAULT_PROFILE = 'dc'.freeze
+    DEFAULT_STATUS  = Ezid::Status::RESERVED
+
+    def initialize
+      configure_ark
+      configure_client
+    end
+
+    def mint
+      Ezid::Identifier.mint
+    end
+
+    private
+
+    def configure_ark
+      Ezid::Identifier.defaults = {
+          export: DEFAULT_EXPORT,
+          profile: DEFAULT_PROFILE,
+          status: DEFAULT_STATUS
+      }
+    end
+
+    def configure_client
+      Ezid::Client.configure do |config|
+        config.default_shoulder = module_configuration.ezid_default_shoulder
+        config.password = module_configuration.ezid_password
+        config.user = module_configuration.ezid_user
+        config.logger = Logger.new(File::NULL)
+      end
+    end
+
+    def module_configuration
+      Ddr::IngestTools::ManifestArkMinter.configuration
+    end
+
+  end
+end

data/lib/ddr/ingesttools/version.rb

@@ -1,5 +1,5 @@
 module Ddr
   module IngestTools
-    VERSION = '0.3.0'
+    VERSION = '0.4.0'
   end
 end

data/spec/fixtures/rdr_importer/configs/default.yml

@@ -0,0 +1,3 @@
+ezid_default_shoulder: ark:/99999/fk4
+ezid_password: apitest
+ezid_user: apitest

data/spec/fixtures/rdr_importer/manifests/manifest_with_all_arks.csv

@@ -0,0 +1,4 @@
+ark,visibility,title,contributor,resource_type,license,file,file,file
+ark:/99999/fk4s76kg89,open,Test 1,"Smith, Sue",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data1.csv,data/data2.csv,docs/doc1.txt
+ark:/99999/fk4ng5vp6m,,Test 2,"Jones, Bill",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data3.csv,docs/doc2.txt
+ark:/99999/fk4hq54w3t,authenticated,Test 3,"Allen, Jane",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data1.csv

data/spec/fixtures/rdr_importer/manifests/manifest_with_no_arks.csv

@@ -0,0 +1,4 @@
+visibility,title,contributor,resource_type,license,file,file,file
+open,Test 1,"Smith, Sue",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data1.csv,data/data2.csv,docs/doc1.txt
+,Test 2,"Jones, Bill",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data3.csv,docs/doc2.txt
+authenticated,Test 3,"Allen, Jane",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data1.csv

data/spec/fixtures/rdr_importer/manifests/manifest_with_some_arks.csv

@@ -0,0 +1,4 @@
+ark,visibility,title,contributor,resource_type,license,file,file,file
+,open,Test 1,"Smith, Sue",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data1.csv,data/data2.csv,docs/doc1.txt
+ark:/99999/fk4ng5vp6m,,Test 2,"Jones, Bill",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data3.csv,docs/doc2.txt
+,authenticated,Test 3,"Allen, Jane",dataset,"http://creativecommons.org/publicdomain/zero/1.0/",data/data1.csv

data/spec/manifest_ark_minter/manifest_parser_spec.rb

@@ -0,0 +1,38 @@
+require 'spec_helper'
+
+module Ddr::IngestTools::ManifestArkMinter
+
+  RSpec.describe ManifestParser do
+
+    subject { described_class.new(manifest_file) }
+
+    describe '#as_csv_table' do
+      let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_some_arks.csv') }
+      specify { expect(subject.as_csv_table).to be_a CSV::Table }
+    end
+
+    describe '#headers' do
+      let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_some_arks.csv') }
+      let(:expected_headers) { %w(ark visibility title contributor resource_type license file) }
+      it 'parses out the list of headers' do
+        expect(subject.headers).to include(*expected_headers)
+      end
+    end
+
+    describe '#arks_missing?' do
+      describe 'no arks assigned' do
+        let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_no_arks.csv') }
+        specify { expect(subject.arks_missing?).to be true }
+      end
+      describe 'some arks assigned' do
+        let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_some_arks.csv') }
+        specify { expect(subject.arks_missing?).to be true }
+      end
+      describe 'all arks assigned' do
+        let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_all_arks.csv') }
+        specify { expect(subject.arks_missing?).to be false }
+      end
+    end
+  end
+
+end

data/spec/manifest_ark_minter/manifest_updater_spec.rb

@@ -0,0 +1,46 @@
+require 'spec_helper'
+require 'tempfile'
+
+module Ddr::IngestTools::ManifestArkMinter
+
+  RSpec.describe ManifestUpdater do
+
+    subject { described_class.new(config: config_file, manifest: manifest_file, output: output_file) }
+
+    let(:config_file) { File.join('spec', 'fixtures', 'rdr_importer', 'configs', 'default.yml') }
+    let(:output_dir) { Dir.mktmpdir }
+    let(:output_file) { File.join(output_dir, 'output.csv') }
+
+    after { FileUtils.remove_dir output_dir }
+
+    describe 'manifest has ARKs for all rows' do
+      let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_all_arks.csv') }
+      it 'does not produce an output file' do
+        expect{ subject.call }.not_to change{ File.exist?(output_file) }
+      end
+    end
+
+    describe 'manifest has ARKs for some rows' do
+      let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_some_arks.csv') }
+      it 'mints ARKs for rows without them' do
+        expect_any_instance_of(Minter).to receive(:mint).exactly(2).times.and_call_original
+        subject.call
+        table = CSV.read(output_file, headers: true)
+        # expect(table['ark']).to all(match(/ark:\/99999\/fk4/))
+        expect(table['ark']).to match([ /ark:\/99999\/fk4/, 'ark:/99999/fk4ng5vp6m', /ark:\/99999\/fk4/ ])
+      end
+    end
+
+    describe 'manifest has ARKs for no rows' do
+      let(:manifest_file) { File.join('spec', 'fixtures', 'rdr_importer', 'manifests', 'manifest_with_no_arks.csv') }
+      it 'mints ARKs for every row' do
+        expect_any_instance_of(Minter).to receive(:mint).exactly(3).times.and_call_original
+        subject.call
+        table = CSV.read(output_file, headers: true)
+        expect(table['ark']).to all(match(/ark:\/99999\/fk4/))
+      end
+    end
+
+  end
+
+end

data/spec/manifest_ark_minter/minter_spec.rb

@@ -0,0 +1,50 @@
+require 'spec_helper'
+
+module Ddr::IngestTools::ManifestArkMinter
+
+  RSpec.describe Minter do
+
+    describe '#initialize' do
+      describe 'Ezid::Identifier defaults' do
+        before { described_class.new }
+        let(:ark_defaults) { { export: described_class::DEFAULT_EXPORT,
+                               profile: described_class::DEFAULT_PROFILE,
+                               status: described_class::DEFAULT_STATUS } }
+        it 'configures Ezid::Identifer defaults' do
+          expect(Ezid::Identifier.defaults).to match(ark_defaults)
+        end
+      end
+      describe 'Ezid::Client configuration' do
+        let(:configuration) { Configuration.new }
+        before do
+          configuration.ezid_default_shoulder = 'ark:/99999/fk4'
+          configuration.ezid_password = 'apitest'
+          configuration.ezid_user = 'apitest'
+          allow(Ddr::IngestTools::ManifestArkMinter).to receive(:configuration) { configuration }
+          described_class.new
+        end
+        it 'configures the Ezid::Client' do
+          expect(Ezid::Client.config.default_shoulder).to eq('ark:/99999/fk4')
+          expect(Ezid::Client.config.password).to eq('apitest')
+          expect(Ezid::Client.config.user).to eq('apitest')
+        end
+      end
+    end
+
+    describe '#mint' do
+      let(:configuration) { Configuration.new }
+      before do
+        configuration.ezid_default_shoulder = 'ark:/99999/fk4'
+        configuration.ezid_password = 'apitest'
+        configuration.ezid_user = 'apitest'
+        allow(Ddr::IngestTools::ManifestArkMinter).to receive(:configuration) { configuration }
+      end
+      it 'calls Ezid::Identifier to mint an ark' do
+        expect(Ezid::Identifier).to receive(:mint)
+        subject.mint
+      end
+    end
+
+  end
+
+end

data/spec/spec_helper.rb

@@ -1,3 +1,4 @@
+require 'byebug'
 require 'ddr/ingesttools'
 require 'i18n'
 

metadata

@@ -1,43 +1,43 @@
 --- !ruby/object:Gem::Specification
 name: ddr-ingesttools
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.4.0
 platform: ruby
 authors:
 - Jim Coble
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2017-05-05 00:00:00.000000000 Z
+date: 2018-03-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: bagit
+  name: ezid-client
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.4'
+        version: '1.7'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.4'
+        version: '1.7'
 - !ruby/object:Gem::Dependency
   name: i18n
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '1.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: '1.0'
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
@@ -52,6 +52,20 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '1.14'
+- !ruby/object:Gem::Dependency
+  name: byebug
+  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: rake
   requirement: !ruby/object:Gem::Requirement
@@ -84,7 +98,7 @@ description: A collection of Ruby tools supporting ingest into the Duke Digital
 email:
 - jim.coble@duke.edu
 executables:
-- convert_dpc_folder.rb
+- mint_manifest_arks.rb
 extensions: []
 extra_rdoc_files: []
 files:
@@ -96,21 +110,24 @@ files:
 - LICENSE.txt
 - README.md
 - Rakefile
-- bin/convert_dpc_folder.rb
+- bin/mint_manifest_arks.rb
 - config/locales/en.yml
 - ddr-ingesttools.gemspec
 - lib/ddr/ingesttools.rb
-- lib/ddr/ingesttools/checksum_file.rb
-- lib/ddr/ingesttools/dpc_folder_converter.rb
-- lib/ddr/ingesttools/dpc_folder_converter/converter.rb
+- lib/ddr/ingesttools/manifest_ark_minter.rb
+- lib/ddr/ingesttools/manifest_ark_minter/configuration.rb
+- lib/ddr/ingesttools/manifest_ark_minter/manifest_parser.rb
+- lib/ddr/ingesttools/manifest_ark_minter/manifest_updater.rb
+- lib/ddr/ingesttools/manifest_ark_minter/minter.rb
 - lib/ddr/ingesttools/version.rb
-- spec/fixtures/files/bad-checksums-sha1.txt
-- spec/fixtures/files/good-checksums-sha1.txt
-- spec/fixtures/files/manifest-sha1-collection-title_admin_set.txt
-- spec/fixtures/files/manifest-sha1.txt
-- spec/integration/dpc_folder_converter_spec.rb
+- spec/fixtures/rdr_importer/configs/default.yml
+- spec/fixtures/rdr_importer/manifests/manifest_with_all_arks.csv
+- spec/fixtures/rdr_importer/manifests/manifest_with_no_arks.csv
+- spec/fixtures/rdr_importer/manifests/manifest_with_some_arks.csv
+- spec/manifest_ark_minter/manifest_parser_spec.rb
+- spec/manifest_ark_minter/manifest_updater_spec.rb
+- spec/manifest_ark_minter/minter_spec.rb
 - spec/spec_helper.rb
-- spec/unit/checksum_file_spec.rb
 homepage: https://github.com/duke-libraries/ddr-ingesttools
 licenses:
 - BSD-3-Clause
@@ -131,15 +148,16 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.6.11
+rubygems_version: 2.6.14
 signing_key: 
 specification_version: 4
 summary: Ruby tools supporting ingest into the Duke Digital Repository.
 test_files:
-- spec/fixtures/files/bad-checksums-sha1.txt
-- spec/fixtures/files/good-checksums-sha1.txt
-- spec/fixtures/files/manifest-sha1-collection-title_admin_set.txt
-- spec/fixtures/files/manifest-sha1.txt
-- spec/integration/dpc_folder_converter_spec.rb
+- spec/fixtures/rdr_importer/configs/default.yml
+- spec/fixtures/rdr_importer/manifests/manifest_with_all_arks.csv
+- spec/fixtures/rdr_importer/manifests/manifest_with_no_arks.csv
+- spec/fixtures/rdr_importer/manifests/manifest_with_some_arks.csv
+- spec/manifest_ark_minter/manifest_parser_spec.rb
+- spec/manifest_ark_minter/manifest_updater_spec.rb
+- spec/manifest_ark_minter/minter_spec.rb
 - spec/spec_helper.rb
-- spec/unit/checksum_file_spec.rb

data/bin/convert_dpc_folder.rb

@@ -1,61 +0,0 @@
-#!/usr/bin/env ruby
-
-require 'ddr/ingesttools'
-require 'optparse'
-
-options = {}
-
-# Parse command line arguments
-parser = OptionParser.new do |opts|
-  opts.banner = 'Usage: convert_dpc_folder.rb [options]'
-
-  opts.on('-s', '--source SOURCE', 'Path to DPC Folder to be converted') do |v|
-    options[:source] = v
-  end
-
-  opts.on('-t', '--target TARGET', 'Path to folder where Standard Ingest Format is to be built') do |v|
-    options[:target] = v
-  end
-
-  opts.on('-i', '--item_id_length LENGTH', Integer, 'Number of characters to copy from the beginning of each file name',
-          'to use as the local ID of the item of which that file is a component') do |v|
-    options[:item_id_length] = v
-  end
-
-  opts.on('-c', '--checksums [CHECKSUM_FILE]', 'External checksum file') do |v|
-    options[:checksums] = v
-  end
-
-  opts.on('--[no-]copy_files', 'Copy files to target location instead of using a symlink') do |v|
-    options[:copy_files] = v
-  end
-
-  opts.on('--collection_title [TITLE]', 'Title for collection',
-          'required if intending to create a collection-creating Standard Ingest') do |v|
-    options[:collection_title] = v
-  end
-
-  opts.on('--admin_set [ADMIN_SET]', 'Admin set for collection',
-          'required if intending to create a collection-creating Standard Ingest') do |v|
-    options[:admin_set] = v
-  end
-
-end
-
-begin
-  parser.parse!
-  mandatory = [ :source, :target, :item_id_length ]
-  missing = mandatory.select{ |param| options[param].nil? }
-  unless missing.empty?
-    raise OptionParser::MissingArgument.new(missing.join(', '))
-  end
-rescue OptionParser::InvalidOption, OptionParser::MissingArgument
-  puts $!.to_s
-  puts parser
-  exit(false)
-end
-
-converter = Ddr::IngestTools::DpcFolderConverter::Converter.new(options)
-results = converter.call
-puts I18n.translate('errors.count', { count: results.errors.size })
-results.errors.each { |e| puts e }

data/lib/ddr/ingesttools/checksum_file.rb

@@ -1,28 +0,0 @@
-module Ddr::IngestTools
-  class ChecksumFile
-
-    attr_reader :digests
-
-    def initialize(checksum_filepath)
-      @digests = digest_hash(checksum_filepath)
-    end
-
-    def digest(filepath)
-      digests[filepath]
-    end
-
-    private
-
-    def digest_hash(checksum_filepath)
-      h = {}
-      File.open(checksum_filepath, 'r') do |file|
-        file.each_line do |line|
-          digest, path = line.chomp.split
-          h[path] = digest
-        end
-      end
-      h
-    end
-
-  end
-end

data/lib/ddr/ingesttools/dpc_folder_converter.rb

@@ -1,35 +0,0 @@
-require_relative 'dpc_folder_converter/converter'
-
-module Ddr::IngestTools
-  module DpcFolderConverter
-    #Configuration defaults
-    @config = {
-        included_extensions: [ '.jpg', '.mov', '.mp3', '.mp4', '.pdf', '.tif', '.tiff', '.wav' ],
-        csv_options: { :encoding=>"UTF-8", :col_sep=>"\t", :headers=>true, :write_headers=>true,
-                       :header_converters=>:symbol }
-    }
-
-    @valid_config_keys = @config.keys
-
-    # Configure through hash
-    def self.configure(opts = {})
-      opts.each {|k,v| @config[k.to_sym] = v if @valid_config_keys.include?(k.to_sym)}
-    end
-
-    # Configure through yaml file
-    def self.configure_with(path_to_yaml_file)
-      begin
-        config = YAML::load(IO.read(path_to_yaml_file))
-      rescue Errno::ENOENT
-        log(:warning, "YAML configuration file couldn't be found. Using defaults."); return
-      rescue Psych::SyntaxError
-        log(:warning, "YAML configuration file contains invalid syntax. Using defaults."); return
-      end
-      configure(config)
-    end
-
-    def self.config
-      @config
-    end
-  end
-end

data/lib/ddr/ingesttools/dpc_folder_converter/converter.rb

@@ -1,151 +0,0 @@
-require 'bagit'
-require 'csv'
-require 'fileutils'
-require 'find'
-
-module Ddr::IngestTools::DpcFolderConverter
-    class Converter
-
-      INTERMEDIATE_FILES_DIRNAME = 'intermediate_files'
-      DPC_TARGETS_DIRNAME = 'targets'
-      SIF_TARGETS_DIRNAME = 'dpc_targets'
-      SIF_METADATA_FILENAME = 'metadata.txt'
-      SIF_MANIFEST_SHA1_FILENAME = 'manifest-sha1.txt'
-
-      Results = Struct.new(:file_map, :errors)
-
-      attr_reader :source, :target, :data_dir, :item_id_length, :checksums, :copy_files, :collection_title,
-                  :admin_set, :metadata_headers
-      attr_accessor :errors, :file_map, :local_id_metadata, :results
-
-      def initialize(source:, target:, item_id_length:, checksums: nil, copy_files: false, collection_title: nil,
-                      admin_set: nil)
-        @source = source
-        @target = target
-        @item_id_length = item_id_length
-        @checksums = checksums
-        @copy_files = copy_files
-        @collection_title = collection_title
-        @admin_set = admin_set
-        @metadata_headers = [ 'path', 'local_id' ]
-        @metadata_headers << 'title' unless collection_title.nil?
-        @metadata_headers << 'admin_set' unless admin_set.nil?
-      end
-
-      def call
-        setup
-        scan_files(source)
-        output_metadata
-        bagitup
-        validate_checksums if checksums
-        Results.new(file_map, errors)
-      end
-
-      private
-
-      def setup
-        @data_dir = File.join(target, 'data')
-        @errors = []
-        @file_map = {}
-        @local_id_metadata = {}
-        FileUtils.mkdir_p data_dir
-      end
-
-      def included_extensions
-        Ddr::IngestTools::DpcFolderConverter.config[:included_extensions]
-      end
-
-      def scan_files(dirpath, file_handler='handle_component'.to_sym)
-        Dir.foreach(dirpath).each do |entry|
-          next if [ '.', '..' ].include?(entry)
-          path = File.join(dirpath, entry)
-          if File.directory?(path)
-            if entry == DPC_TARGETS_DIRNAME
-              scan_files(path, :handle_target)
-            elsif entry == INTERMEDIATE_FILES_DIRNAME
-              scan_files(path, :handle_intermediate_file)
-            else
-              scan_files(path, file_handler)
-            end
-          else
-            if included_extensions.include?(File.extname(entry))
-              self.send(file_handler, path)
-            end
-          end
-        end
-      end
-
-      def handle_component(file)
-        base = File.basename(file, File.extname(file))
-        item_id = item_id_length == 0 ? base : base[0, item_id_length]
-        FileUtils.mkdir_p(File.join(data_dir, item_id))
-        local_id_metadata[item_id] = item_id
-        handle_file(file, item_id)
-        local_id_metadata[File.join(item_id, File.basename(file))] = base
-      end
-
-      def handle_intermediate_file(file)
-        FileUtils.mkdir_p(File.join(data_dir, INTERMEDIATE_FILES_DIRNAME))
-        handle_file(file, INTERMEDIATE_FILES_DIRNAME)
-      end
-
-      def handle_target(file)
-        base = File.basename(file, File.extname(file))
-        FileUtils.mkdir_p(File.join(data_dir, SIF_TARGETS_DIRNAME))
-        handle_file(file, SIF_TARGETS_DIRNAME)
-        local_id_metadata[File.join(SIF_TARGETS_DIRNAME, File.basename(file))] = base
-      end
-
-      def handle_file(file, folder_name)
-        if copy_files
-          FileUtils.cp file, File.join(data_dir, folder_name)
-        else
-          FileUtils.ln_s file, File.join(data_dir, folder_name)
-        end
-        file_map[file] = File.join(data_dir, folder_name, File.basename(file))
-      end
-
-      def output_metadata
-        metadata_rows = []
-        case
-          when collection_title && admin_set
-            metadata_rows << CSV::Row.new(metadata_headers, [ nil, nil, collection_title, admin_set ])
-          when collection_title
-            metadata_rows << CSV::Row.new(metadata_headers, [ nil, nil, collection_title ])
-          when admin_set
-            metadata_rows << CSV::Row.new(metadata_headers, [ nil, nil, admin_set ])
-        end
-        local_id_metadata.each_pair do |k,v|
-          row_elements = [ k, v ]
-          row_elements << nil if collection_title
-          row_elements << nil if admin_set
-          metadata_rows << CSV::Row.new(metadata_headers, row_elements)
-        end
-        File.open(File.join(data_dir, SIF_METADATA_FILENAME), 'w') do |file|
-          file.puts(metadata_headers.join(Ddr::IngestTools::DpcFolderConverter.config[:csv_options][:col_sep]))
-          metadata_rows.each do |row|
-            file.puts(row.to_csv(Ddr::IngestTools::DpcFolderConverter.config[:csv_options]))
-          end
-        end
-      end
-
-      def bagitup
-        bag = BagIt::Bag.new(target)
-        bag.manifest!
-      end
-
-      def validate_checksums
-        external_checksums = Ddr::IngestTools::ChecksumFile.new(checksums)
-        sif_manifest = Ddr::IngestTools::ChecksumFile.new(File.join(target, SIF_MANIFEST_SHA1_FILENAME))
-        file_map.each do |source_path, target_path|
-          external_checksum = external_checksums.digest(source_path)
-          manifest_path = target_path.sub("#{target}/", '')
-          sif_checksum = sif_manifest.digest(manifest_path)
-          unless external_checksum == sif_checksum
-            errors << I18n.translate('errors.checksum_mismatch', { c1: external_checksum, f1: source_path,
-                                                                   c2: sif_checksum, f2: target_path })
-          end
-        end
-      end
-    end
-end

data/spec/fixtures/files/bad-checksums-sha1.txt

@@ -1,11 +0,0 @@
-3201454891f1d63a1493f393e061491e2d65ffcd  SOURCE_DIRECTORY/Thumbs.db
-59ec01f979a76b968bc579e5cd0ceb3bcf3e629f  SOURCE_DIRECTORY/abc001001.tif
-d0a2f2482783ae3c83d06f3cdeaa1a306cc043ad  SOURCE_DIRECTORY/abc001002.tif
-38ee72ab417192589f3a54ef1016131c7d7e9e4e  SOURCE_DIRECTORY/abc002001.tif
-1da7765e3ca71ed76395bc56dae69d64732beff8  SOURCE_DIRECTORY/checksums.txt
-c227abc095d3b758ab1c1c1c9e922494b6b6e0b0  SOURCE_DIRECTORY/g/abc003001.wav
-541af8df7d2c631382bd0ec189476894d0323df5  SOURCE_DIRECTORY/g/abc003002.wav
-49d77d27bf82ec3dafa1967490594883f5e8b432  SOURCE_DIRECTORY/intermediate_files/abc001001.jpg
-260b3c2d20a1726de96671d29f73ba09d13b61ba  SOURCE_DIRECTORY/intermediate_files/abc002001.jpg
-a08c4d5a76d1b8735587be6ffcba66a9baf475c4  SOURCE_DIRECTORY/targets/T001.tif
-40f9993d06945544fa57af88f7c3e9102ecc69e3  SOURCE_DIRECTORY/targets/T002.tif

data/spec/fixtures/files/good-checksums-sha1.txt

@@ -1,11 +0,0 @@
-3201454891f1d63a1493f393e061491e2d65ffcd  SOURCE_DIRECTORY/Thumbs.db
-59ec01f979a76b968bc579e5cd0ceb3bcf3e629f  SOURCE_DIRECTORY/abc001001.tif
-d0a2f2482783ae3c38d06f3cdeaa1a306cc043ad  SOURCE_DIRECTORY/abc001002.tif
-38ee72ab417192589f3a54ef1016131c7d7e9e4e  SOURCE_DIRECTORY/abc002001.tif
-1da7765e3ca71ed76395bc56dae69d64732beff8  SOURCE_DIRECTORY/checksums.txt
-c227abc095d3b758051c1c1c9e922494b6b6e0b0  SOURCE_DIRECTORY/g/abc003001.wav
-541af8df7d2c631382bd0ec189476894d0323df5  SOURCE_DIRECTORY/g/abc003002.wav
-49d77d27bf82ec3dafa1967490594883f5e8b432  SOURCE_DIRECTORY/intermediate_files/abc001001.jpg
-260b3c2d20a7126de96671d29f73ba09d13b61ba  SOURCE_DIRECTORY/intermediate_files/abc002001.jpg
-a08c4d5a76d1b8734487be6ffcba66a9baf475c4  SOURCE_DIRECTORY/targets/T001.tif
-40f9993d06945544fa57af88f7c3e9102ecc69e3  SOURCE_DIRECTORY/targets/T002.tif

data/spec/fixtures/files/manifest-sha1-collection-title_admin_set.txt

@@ -1,10 +0,0 @@
-59ec01f979a76b968bc579e5cd0ceb3bcf3e629f data/abc001/abc001001.tif
-d0a2f2482783ae3c38d06f3cdeaa1a306cc043ad data/abc001/abc001002.tif
-38ee72ab417192589f3a54ef1016131c7d7e9e4e data/abc002/abc002001.tif
-c227abc095d3b758051c1c1c9e922494b6b6e0b0 data/abc003/abc003001.wav
-541af8df7d2c631382bd0ec189476894d0323df5 data/abc003/abc003002.wav
-a08c4d5a76d1b8734487be6ffcba66a9baf475c4 data/dpc_targets/T001.tif
-40f9993d06945544fa57af88f7c3e9102ecc69e3 data/dpc_targets/T002.tif
-49d77d27bf82ec3dafa1967490594883f5e8b432 data/intermediate_files/abc001001.jpg
-260b3c2d20a7126de96671d29f73ba09d13b61ba data/intermediate_files/abc002001.jpg
-21f041cdd694f7755ed84b8cd2668214a43bad6c data/metadata.txt

data/spec/fixtures/files/manifest-sha1.txt

@@ -1,10 +0,0 @@
-59ec01f979a76b968bc579e5cd0ceb3bcf3e629f data/abc001/abc001001.tif
-d0a2f2482783ae3c38d06f3cdeaa1a306cc043ad data/abc001/abc001002.tif
-38ee72ab417192589f3a54ef1016131c7d7e9e4e data/abc002/abc002001.tif
-c227abc095d3b758051c1c1c9e922494b6b6e0b0 data/abc003/abc003001.wav
-541af8df7d2c631382bd0ec189476894d0323df5 data/abc003/abc003002.wav
-a08c4d5a76d1b8734487be6ffcba66a9baf475c4 data/dpc_targets/T001.tif
-40f9993d06945544fa57af88f7c3e9102ecc69e3 data/dpc_targets/T002.tif
-49d77d27bf82ec3dafa1967490594883f5e8b432 data/intermediate_files/abc001001.jpg
-260b3c2d20a7126de96671d29f73ba09d13b61ba data/intermediate_files/abc002001.jpg
-913699468893882d1dec463f3df1a405c7f32784 data/metadata.txt

data/spec/integration/dpc_folder_converter_spec.rb

@@ -1,201 +0,0 @@
-module Ddr::IngestTools::DpcFolderConverter
-
-  RSpec.describe Converter do
-
-    shared_examples 'a conversion to standard ingest format' do
-      subject { described_class.new(converter_args) }
-      it 'produces the correct standard ingest format directory' do
-        results = subject.call
-        # Target directory contains all the expected files and only the expected files
-        expect(Array(Find.find(target_directory))).to match_array(expected_files)
-        # Target content files are same as source content files
-        expect(FileUtils.compare_file(File.join(data_directory, 'abc001', 'abc001001.tif'),
-                                      File.join(source_directory, 'abc001001.tif'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'abc001', 'abc001002.tif'),
-                                      File.join(source_directory, 'abc001002.tif'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'abc002', 'abc002001.tif'),
-                                      File.join(source_directory, 'abc002001.tif'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'abc003', 'abc003001.wav'),
-                                      File.join(source_directory, 'g', 'abc003001.wav'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'abc003', 'abc003002.wav'),
-                                      File.join(source_directory, 'g', 'abc003002.wav'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'dpc_targets', 'T001.tif'),
-                                      File.join(source_directory, 'targets', 'T001.tif'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'dpc_targets', 'T002.tif'),
-                                      File.join(source_directory, 'targets', 'T002.tif'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'intermediate_files', 'abc001001.jpg'),
-                                      File.join(source_directory, 'intermediate_files', 'abc001001.jpg'))).to be true
-        expect(FileUtils.compare_file(File.join(data_directory, 'intermediate_files', 'abc002001.jpg'),
-                                      File.join(source_directory, 'intermediate_files', 'abc002001.jpg'))).to be true
-        # Generated metadata file contains the expected contents
-        metadata_lines = File.readlines(File.join(data_directory, 'metadata.txt')).map(&:chomp)
-        expect(metadata_lines).to match_array(expected_metadata)
-        # Generated manifest contains the expected contents (ignoring line order)
-        generated_manifest = File.readlines(File.join(File.join(target_directory, 'manifest-sha1.txt'))).sort
-        expect(generated_manifest).to match_array(expected_manifest)
-        # Conversion process produces the expected errors
-        expect(results.errors).to match_array(checksum_errors)
-      end
-    end
-
-    let(:source_directory) { Dir.mktmpdir('dpc') }
-    let(:target_directory) { Dir.mktmpdir('sif') }
-    let(:data_directory) { File.join(target_directory, 'data') }
-    let(:item_id_length) { 6 }
-    let(:checksums_directory) { Dir.mktmpdir('checksums') }
-    let(:checksums) { File.join(checksums_directory, 'checksums-sha1.txt') }
-    let(:converter_args) { { source: source_directory, target: target_directory, item_id_length: item_id_length } }
-    let(:expected_files) { [
-        target_directory,
-        File.join(target_directory, 'bag-info.txt'),
-        File.join(target_directory, 'bagit.txt'),
-        data_directory,
-        File.join(data_directory, 'abc001'),
-        File.join(data_directory, 'abc001', 'abc001001.tif'),
-        File.join(data_directory, 'abc001', 'abc001002.tif'),
-        File.join(data_directory, 'abc002'),
-        File.join(data_directory, 'abc002', 'abc002001.tif'),
-        File.join(data_directory, 'abc003', 'abc003001.wav'),
-        File.join(data_directory, 'abc003'),
-        File.join(data_directory, 'abc003', 'abc003002.wav'),
-        File.join(data_directory, 'dpc_targets'),
-        File.join(data_directory, 'dpc_targets', 'T001.tif'),
-        File.join(data_directory, 'dpc_targets', 'T002.tif'),
-        File.join(data_directory, 'intermediate_files'),
-        File.join(data_directory, 'intermediate_files', 'abc001001.jpg'),
-        File.join(data_directory, 'intermediate_files', 'abc002001.jpg'),
-        File.join(data_directory, 'metadata.txt'),
-        File.join(target_directory, 'manifest-md5.txt'),
-        File.join(target_directory, 'manifest-sha1.txt'),
-        File.join(target_directory, 'tagmanifest-md5.txt'),
-        File.join(target_directory, 'tagmanifest-sha1.txt')
-    ] }
-    let(:expected_metadata) { [
-        "path\tlocal_id",
-        "abc001\tabc001",
-        "abc002\tabc002",
-        "abc003\tabc003",
-        "abc001/abc001001.tif\tabc001001",
-        "abc001/abc001002.tif\tabc001002",
-        "abc002/abc002001.tif\tabc002001",
-        "abc003/abc003001.wav\tabc003001",
-        "abc003/abc003002.wav\tabc003002",
-        "dpc_targets/T001.tif\tT001",
-        "dpc_targets/T002.tif\tT002"
-    ] }
-    let(:expected_manifest) do
-      File.readlines(File.join('spec', 'fixtures', 'files', 'manifest-sha1.txt')).sort
-    end
-
-    before do
-      File.open(File.join(source_directory, 'Thumbs.db'), 'w') { |f| f.write('Thumbs') }
-      File.open(File.join(source_directory, 'abc001001.tif'), 'w') { |f| f.write('abc001001') }
-      File.open(File.join(source_directory, 'abc001002.tif'), 'w') { |f| f.write('abc001002') }
-      File.open(File.join(source_directory, 'abc002001.tif'), 'w') { |f| f.write('abc002001') }
-      File.open(File.join(source_directory, 'checksums.txt'), 'w') { |f| f.write('checksums') }
-      Dir.mkdir(File.join(source_directory,'g'))
-      File.open(File.join(source_directory, 'g', 'abc003001.wav'), 'w') { |f| f.write('abc003001') }
-      File.open(File.join(source_directory, 'g', 'abc003002.wav'), 'w') { |f| f.write('abc003002') }
-      Dir.mkdir(File.join(source_directory, 'intermediate_files'))
-      File.open(File.join(source_directory, 'intermediate_files', 'abc001001.jpg'), 'w') { |f| f.write('abc001001 jpg')}
-      File.open(File.join(source_directory, 'intermediate_files', 'abc002001.jpg'), 'w') { |f| f.write('abc002001 jpg')}
-      Dir.mkdir(File.join(source_directory, 'targets'))
-      File.open(File.join(source_directory, 'targets', 'T001.tif'), 'w') { |f| f.write('T001') }
-      File.open(File.join(source_directory, 'targets', 'T002.tif'), 'w') { |f| f.write('T002') }
-    end
-
-    describe 'external checksum files' do
-      describe 'external checksum file' do
-        before do
-          File.open(checksums, 'w') do |f|
-            f << File.open(checksum_file_template).read.gsub('SOURCE_DIRECTORY', source_directory)
-          end
-          converter_args[:checksums] = checksums
-        end
-        describe 'mismatch' do
-          let(:checksum_file_template) { File.join('spec', 'fixtures', 'files', 'bad-checksums-sha1.txt') }
-          let(:checksum_errors) {
-            [ I18n.translate('errors.checksum_mismatch', { c1: 'd0a2f2482783ae3c83d06f3cdeaa1a306cc043ad',
-                                                           f1: File.join(source_directory, 'abc001002.tif'),
-                                                           c2: 'd0a2f2482783ae3c38d06f3cdeaa1a306cc043ad',
-                                                           f2: File.join(target_directory, 'data/abc001/abc001002.tif') }),
-              I18n.translate('errors.checksum_mismatch', { c1: 'c227abc095d3b758ab1c1c1c9e922494b6b6e0b0',
-                                                           f1: File.join(source_directory, 'g/abc003001.wav'),
-                                                           c2: 'c227abc095d3b758051c1c1c9e922494b6b6e0b0',
-                                                           f2: File.join(target_directory, 'data/abc003/abc003001.wav') }),
-              I18n.translate('errors.checksum_mismatch', { c1: '260b3c2d20a1726de96671d29f73ba09d13b61ba',
-                                                           f1: File.join(source_directory, 'intermediate_files/abc002001.jpg'),
-                                                           c2: '260b3c2d20a7126de96671d29f73ba09d13b61ba',
-                                                           f2: File.join(target_directory, 'data/intermediate_files/abc002001.jpg') }),
-              I18n.translate('errors.checksum_mismatch', { c1: 'a08c4d5a76d1b8735587be6ffcba66a9baf475c4',
-                                                           f1: File.join(source_directory, 'targets/T001.tif'),
-                                                           c2: 'a08c4d5a76d1b8734487be6ffcba66a9baf475c4',
-                                                           f2: File.join(target_directory, 'data/dpc_targets/T001.tif') })
-            ]
-          }
-          describe 'files are copied' do
-            before { converter_args[:copy_files] = true }
-            it_behaves_like 'a conversion to standard ingest format'
-          end
-          describe 'files are not copied' do
-            before { converter_args[:copy_files] = false }
-            it_behaves_like 'a conversion to standard ingest format'
-          end
-        end
-        describe 'no mismatch' do
-          let(:checksum_file_template) { File.join('spec', 'fixtures', 'files', 'good-checksums-sha1.txt') }
-          let(:checksum_errors) { [] }
-          describe 'files are copied' do
-            before { converter_args[:copy_files] = true }
-            it_behaves_like 'a conversion to standard ingest format'
-          end
-          describe 'files are not copied' do
-            before { converter_args[:copy_files] = false }
-            it_behaves_like 'a conversion to standard ingest format'
-          end
-        end
-      end
-
-      describe 'no external checksum file' do
-        let(:checksum_errors) { [] }
-        describe 'files are copied' do
-          before { converter_args[:copy_files] = true }
-          it_behaves_like 'a conversion to standard ingest format'
-        end
-        describe 'files are not copied' do
-          before { converter_args[:copy_files] = false }
-          it_behaves_like 'a conversion to standard ingest format'
-        end
-      end
-    end
-
-    describe 'collection titles and admin sets' do
-      let(:checksum_errors) { [] }
-      describe 'collection title and admin set provided' do
-        let(:expected_metadata) { [
-            "path\tlocal_id\ttitle\tadmin_set",
-            "\t\tTest Collection\tfoo",
-            "abc001\tabc001\t\t",
-            "abc002\tabc002\t\t",
-            "abc003\tabc003\t\t",
-            "abc001/abc001001.tif\tabc001001\t\t",
-            "abc001/abc001002.tif\tabc001002\t\t",
-            "abc002/abc002001.tif\tabc002001\t\t",
-            "abc003/abc003001.wav\tabc003001\t\t",
-            "abc003/abc003002.wav\tabc003002\t\t",
-            "dpc_targets/T001.tif\tT001\t\t",
-            "dpc_targets/T002.tif\tT002\t\t"
-        ] }
-        let(:expected_manifest) do
-          File.readlines(File.join('spec', 'fixtures', 'files', 'manifest-sha1-collection-title_admin_set.txt')).sort
-        end
-        before do
-          converter_args[:collection_title] = 'Test Collection'
-          converter_args[:admin_set] = 'foo'
-        end
-        it_behaves_like 'a conversion to standard ingest format'
-      end
-    end
-
-  end
-end

data/spec/unit/checksum_file_spec.rb

@@ -1,17 +0,0 @@
-module Ddr::IngestTools
-
-  RSpec.describe ChecksumFile do
-
-    subject { described_class.new(checksum_filepath) }
-
-    let(:checksum_filepath) { File.join('spec', 'fixtures', 'files', 'manifest-sha1.txt') }
-
-    describe 'digest' do
-      it 'provides the requested digest' do
-        expect(subject.digest('data/abc001/abc001002.tif')).to eq('d0a2f2482783ae3c38d06f3cdeaa1a306cc043ad')
-        expect(subject.digest('not/in/checksum/file.txt')).to be nil
-      end
-    end
-  end
-
-end