darlingtonia 2.0.0 → 2.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ad3721dc07a3ded3cdd80bd95b0068f098e930a002b5f6e29eeb03a855d62731
-  data.tar.gz: c5b3895c77363da927d7952ff786fe6c63048054db96400cf809511ce5cba144
+  metadata.gz: 145bf822e40f94c82cf2a0c61213e964a8f5e780209703e9506fa6ce35466612
+  data.tar.gz: ed804197a3ccad1047da9b8c8127603e38814c03d799325bb9c317938de71c88
 SHA512:
-  metadata.gz: 0d02ffe72075c23e20aa4d9d222074aa50b669ef3cb7485739e5b9b5cb473b0ff94f9981d3896e2ce3c64b10218222fcd2b65083ec3fb6dafb823806f46492c0
-  data.tar.gz: 27451965e8f4d3493bb7083dc8e6f7f4ad0b189186812d50374b8a7627c4006551bf56c3586b70e62621d470a7b692bc824c2eacb7f6726acf4c3b7a951d3f49
+  metadata.gz: dbaec8b01d54cb94859e1630fec14b84082c981e25143646eb1aa58ab4dbb5e0bbebcb4954457799dccdac0d8cc20f78b315486f2f2d38bfa709bf209df70980
+  data.tar.gz: d5ce22c9e01ba293513d7398fa5af18bb5cfcfe1df67dc6313ebb70cce13c655dd8c1b0998042a1b823014337b2e2ae7980b288f5168a0d4d4caa5077d32701e

data/.gitignore

@@ -4,3 +4,12 @@ Gemfile.lock
 .byebug_history
 .fcrepo_wrapper_test.yml
 .solr_wrapper_test.yml
+darlingtonia_import.log
+*~
+\#*\#
+/.emacs.desktop
+/.emacs.desktop.lock
+*.elc
+auto-save-list
+tramp
+.\#*

data/.rubocop.yml

@@ -19,6 +19,10 @@ Metrics/BlockLength:
     - 'spec/**/*'
     - 'lib/darlingtonia/spec/**/*'
 
+Metrics/CyclomaticComplexity:
+  Exclude:
+    - lib/darlingtonia/hyrax_basic_metadata_mapper.rb
+
 Metrics/LineLength:
   Exclude:
     - 'lib/darlingtonia/hyrax_basic_metadata_mapper.rb'
@@ -47,3 +51,6 @@ RSpec/MultipleExpectations:
   Exclude:
     - 'spec/darlingtonia/hyrax_basic_metadata_mapper_spec.rb'
     - 'spec/integration/import_hyrax_csv.rb'
+
+Style/StructInheritance:
+  Enabled: false

data/CHANGELOG.md

@@ -1,9 +1,18 @@
+2.1.0 - Tue Feb 12, 2019
+
+* Map a variety of values for 'visibility' to their Hyrax approved equivalents
+* Remove date_uploaded from HyraxBasicMetadataMapper so it doesn't over-write the Hyrax provided timestamp
+* Import headers are now case and whitespace insensitive
+* Raise a more meaningful error when an expected file isn't found
+* Transform location URIs into a based_near hash so geonames URIs import properly into Hyrax
+* Add a default LogStream class so that Darlingtonia has logging by default instead of only sending output to STDOUT
+
 2.0.0 - Mon Feb 4, 2019
 
 Assume our base use case is to import into Hyrax.
   * Use HyraxRecordImporter as default
   * Use HyraxMapper as default
-  * Add a getting started guide to the docs directory 
+  * Add a getting started guide to the docs directory
 
 1.2.3 - Thu Jan 24, 2019
 ------------------------

data/darlingtonia.gemspec

@@ -20,13 +20,14 @@ Gem::Specification.new do |gem|
   gem.add_dependency 'active-fedora', '>= 11.5.2'
 
   gem.add_development_dependency 'yard',           '~> 0.9'
-  gem.add_development_dependency 'bixby',          '~> 0.3'
+  gem.add_development_dependency 'bixby',          '~> 1.0'
   gem.add_development_dependency 'hyrax-spec',     '~> 0.2'
   gem.add_development_dependency 'rspec',          '~> 3.6'
   gem.add_development_dependency 'coveralls',      '~> 0.8'
   gem.add_development_dependency 'solr_wrapper',   '~> 2.1'
   gem.add_development_dependency 'fcrepo_wrapper', '~> 0.9'
   gem.add_development_dependency 'byebug'
+  gem.add_development_dependency 'rake'
 
   gem.files         = `git ls-files`.split("\n")
   gem.test_files    = `git ls-files -- {spec}/*`.split("\n")

data/lib/darlingtonia.rb

@@ -33,6 +33,7 @@ module Darlingtonia
   end
   module_function :config
 
+  require 'darlingtonia/log_stream'
   ##
   # Module-wide options for `Darlingtonia`.
   class Configuration
@@ -44,8 +45,8 @@ module Darlingtonia
     attr_accessor :default_error_stream, :default_info_stream
 
     def initialize
-      self.default_error_stream = STDOUT
-      self.default_info_stream  = STDOUT
+      self.default_error_stream = Darlingtonia::LogStream.new
+      self.default_info_stream  = Darlingtonia::LogStream.new
     end
   end
 

data/lib/darlingtonia/hyrax_basic_metadata_mapper.rb

@@ -1,4 +1,5 @@
 # frozen_string_literal: true
+require 'uri'
 
 module Darlingtonia
   ##
@@ -37,31 +38,52 @@ module Darlingtonia
     # Hyrax should return a single value instead of
     # an Array of values.
     def depositor
-      metadata['depositor']
-    end
-
-    def date_uploaded
-      metadata['date_uploaded']
+      single_value('depositor')
     end
 
     def date_modified
-      metadata['date_modified']
+      single_value('date_modified')
     end
 
     def label
-      metadata['label']
+      single_value('label')
     end
 
     def relative_path
-      metadata['relative_path']
+      single_value('relative_path')
     end
 
     def import_url
-      metadata['import_url']
+      single_value('import_url')
     end
 
+    # We should accept visibility values that match the UI and transform them into
+    # the controlled vocabulary term expected by Hyrax
     def visibility
-      metadata['visibility']
+      case metadata[matching_header('visibility')]&.downcase&.gsub(/\s+/, "")
+      when 'public'
+        'open'
+      when 'open'
+        'open'
+      when 'registered'
+        'authenticated'
+      when "authenticated"
+        'authenticated'
+      when ::Hyrax::Institution.name&.downcase&.gsub(/\s+/, "")
+        'authenticated'
+      when ::Hyrax::Institution.name_full&.downcase&.gsub(/\s+/, "")
+        'authenticated'
+      when 'private'
+        'restricted'
+      when 'restricted'
+        'restricted'
+      else
+        'restricted' # This is the default if nothing else matches
+      end
+    end
+
+    def files
+      map_field('files')
     end
 
     ##
@@ -79,16 +101,38 @@ module Darlingtonia
     ##
     # @see MetadataMapper#map_field
     def map_field(name)
-      method_name = name
+      method_name = name.to_s
       method_name = CSV_HEADERS[name] if CSV_HEADERS.keys.include?(name)
-      Array(metadata[method_name.to_s]&.split(delimiter))
+      key = matching_header(method_name)
+      Array(metadata[key]&.split(delimiter))
     end
 
     protected
 
+      # Some fields should have single values instead
+      # of array values.
+      def single_value(field_name)
+        metadata[matching_header(field_name)]
+      end
+
+      # Lenient matching for headers.
+      # If the user has headers like:
+      #   'Title' or 'TITLE' or 'Title  '
+      # it should match the :title field.
+      def matching_header(field_name)
+        metadata.keys.find do |key|
+          next unless key
+          key.downcase.strip == field_name
+        end
+      end
+
       # Properties defined in Hyrax::CoreMetadata
+      # Note that date_uploaded is NOT set here, even though it is defined in
+      # Hyrax::CoreMetadata. Hyrax expects to set date_uploaded itself, and
+      # sending a metadata value for that field interferes with Hyrax expected
+      # behavior.
       def core_fields
-        [:depositor, :title, :date_uploaded, :date_modified]
+        [:depositor, :title, :date_modified]
       end
 
       # Properties defined in Hyrax::BasicMetadata

data/lib/darlingtonia/hyrax_record_importer.rb

@@ -1,26 +1,42 @@
 # frozen_string_literal: true
+
 module Darlingtonia
   class HyraxRecordImporter < RecordImporter
     # TODO: Get this from Hyrax config
     DEFAULT_CREATOR_KEY = 'batchuser@example.com'
 
-    ##
-    # @!attribute [rw] creator
-    #   @return [User]
-    attr_accessor :creator
+    # @!attribute [rw] depositor
+    # @return [User]
+    attr_accessor :depositor
+
+    # @!attribute [rw] collection_id
+    # @return [String] The fedora ID for a Collection.
     attr_accessor :collection_id
 
-    ##
-    # @param creator   [User]
+    # @param attributes [Hash] Attributes that come
+    #        from the UI or importer rather than from
+    #        the CSV/mapper.
+    # @example
+    #   attributes: { collection_id: '123',
+    #                 depositor_id: '456' }
     def initialize(error_stream: Darlingtonia.config.default_error_stream,
                    info_stream: Darlingtonia.config.default_info_stream,
-                   collection_id: nil)
-      self.creator = ::User.find_or_create_system_user(DEFAULT_CREATOR_KEY)
-      self.collection_id = collection_id
+                   attributes: {})
+      self.collection_id = attributes[:collection_id]
+      set_depositor(attributes[:depositor_id])
 
       super(error_stream: error_stream, info_stream: info_stream)
     end
 
+    # "depositor" is a required field for Hyrax.  If
+    # it hasn't been set, set it to the Hyrax default
+    # batch user.
+    def set_depositor(user_id)
+      user = User.find(user_id) if user_id
+      user ||= ::User.find_or_create_system_user(DEFAULT_CREATOR_KEY)
+      self.depositor = user
+    end
+
     ##
     # @param record [ImportRecord]
     #
@@ -42,14 +58,6 @@ module Darlingtonia
       Hyrax.config.curation_concerns.first
     end
 
-    # Depositor is a required field for Hyrax
-    # If it hasn't been set in the metadata, set it to the Hyrax default batch user
-    # Even if it has been set, for now, override that with the batch user.
-    # @param Darlingtonia::InputRecord
-    def set_depositor(record)
-      record.mapper.metadata["depositor"] = @creator.user_key
-    end
-
     # The path on disk where file attachments can be found
     def file_attachments_path
       ENV['IMPORT_PATH'] || '/opt/data'
@@ -61,15 +69,13 @@ module Darlingtonia
     # @param [Darlingtonia::InputRecord]
     # @return [Array] an array of Hyrax::UploadedFile ids
     def create_upload_files(record)
-      file_attachment_filenames = record.mapper.metadata["files"]
-      return [] if file_attachment_filenames.nil? || file_attachment_filenames.empty?
-
-      files_to_attach = file_attachment_filenames.split(record.mapper.delimiter)
+      files_to_attach = record.mapper.files
+      return [] if files_to_attach.nil? || files_to_attach.empty?
 
       uploaded_file_ids = []
       files_to_attach.each do |filename|
         file = File.open(find_file_path(filename))
-        uploaded_file = Hyrax::UploadedFile.create(user: @creator, file: file)
+        uploaded_file = Hyrax::UploadedFile.create(user: @depositor, file: file)
         uploaded_file_ids << uploaded_file.id
         file.close
       end
@@ -79,10 +85,51 @@ module Darlingtonia
     ##
     # Within the directory specified by ENV['IMPORT_PATH'], find the first
     # instance of a file matching the given filename.
+    # If there is no matching file, raise an exception.
     # @param [String] filename
     # @return [String] a full pathname to the found file
     def find_file_path(filename)
-      Dir.glob("#{ENV['IMPORT_PATH']}/**/#{filename}").first
+      filepath = Dir.glob("#{ENV['IMPORT_PATH']}/**/#{filename}").first
+      raise "Cannot find file #{filename}... Are you sure it has been uploaded and that the filename matches?" if filepath.nil?
+      filepath
+    end
+
+    ##
+    # When submitting location data (a.k.a. the "based near" attribute) via the UI,
+    # Hyrax expects to receive a `based_near_attributes` hash in a specific format.
+    # We need to take geonames urls as provided by the customer and transform them to
+    # mimic what the Hyrax UI would ordinarily produce. These will get turned into
+    # Hyrax::ControlledVocabularies::Location objects upon ingest.
+    # The expected hash looks like this:
+    #   "based_near_attributes"=>
+    #     {
+    #       "0"=> {
+    #               "id"=>"http://sws.geonames.org/5667009/", "_destroy"=>""
+    #             },
+    #       "1"=> {
+    #               "id"=>"http://sws.geonames.org/6252001/", "_destroy"=>""
+    #             },
+    #   }
+    # @return [Hash] a "based_near_attributes" hash as
+    def based_near_attributes(based_near)
+      original_geonames_uris = based_near
+      return if original_geonames_uris.empty?
+      based_near_attributes = {}
+      original_geonames_uris.each_with_index do |uri, i|
+        based_near_attributes[i.to_s] = { 'id' => uri_to_sws(uri), "_destroy" => "" }
+      end
+      based_near_attributes
+    end
+
+    #
+    # Take a user-facing geonames URI and return an sws URI, of the form Hyrax expects
+    # (e.g., "http://sws.geonames.org/6252001/")
+    # @param [String] uri
+    # @return [String] an sws style geonames uri
+    def uri_to_sws(uri)
+      uri = URI(uri)
+      geonames_number = uri.path.split('/')[1]
+      "http://sws.geonames.org/#{geonames_number}/"
     end
 
     private
@@ -92,16 +139,21 @@ module Darlingtonia
       def create_for(record:)
         info_stream << 'Creating record: ' \
                        "#{record.respond_to?(:title) ? record.title : record}."
-        set_depositor(record)
-        uploaded_files = { uploaded_files: create_upload_files(record) }
-        created    = import_type.new
+        additional_attrs = {
+          uploaded_files: create_upload_files(record),
+          depositor: @depositor.user_key
+        }
+        created = import_type.new
+
+        attrs = record.attributes.merge(additional_attrs)
+        attrs = attrs.merge(member_of_collections_attributes: { '0' => { id: collection_id } }) if collection_id
 
-        attributes = record.attributes.merge(uploaded_files)
-        attributes = attributes.merge(member_of_collection_ids: [collection_id]) if collection_id
+        based_near = attrs.delete(:based_near)
+        attrs = attrs.merge(based_near_attributes: based_near_attributes(based_near)) unless based_near.nil? || based_near.empty?
 
-        actor_env  = Hyrax::Actors::Environment.new(created,
-                                                    ::Ability.new(@creator),
-                                                    attributes)
+        actor_env = Hyrax::Actors::Environment.new(created,
+                                                   ::Ability.new(@depositor),
+                                                   attrs)
 
         if Hyrax::CurationConcern.actor.create(actor_env)
           info_stream << "Record created at: #{created.id}"

data/lib/darlingtonia/log_stream.rb

@@ -0,0 +1,43 @@
+# frozen_string_literal: true
+
+module Darlingtonia
+  class LogStream
+    ##
+    # @!attribute [rw] logger
+    #   @return [Logger]
+    # @!attribute [rw] severity
+    #   @return [Logger::Serverity]
+    attr_accessor :logger, :severity
+
+    def initialize(logger: nil, severity: nil)
+      self.logger   = logger   || Logger.new(build_filename)
+      self.severity = severity || Logger::INFO
+    end
+
+    def <<(msg)
+      logger.add(severity, msg)
+      STDOUT << msg
+    end
+
+    private
+
+      def build_filename
+        return ENV['IMPORT_LOG'] if ENV['IMPORT_LOG']
+        return rails_log_name if rails_log_name
+        return './darlingtonia_import.log' unless rails_log_name
+      end
+
+      def rails_log_name
+        case Rails.env
+        when 'production'
+          Rails.root.join('log', "csv_import.log").to_s
+        when 'development'
+          Rails.root.join('log', "dev_csv_import.log").to_s
+        when 'test'
+          Rails.root.join('log', "test_csv_import.log").to_s
+        end
+      rescue NameError
+        false
+      end
+  end
+end

data/lib/darlingtonia/spec/fakes/fake_parser.rb

@@ -14,11 +14,9 @@ class FakeParser < Darlingtonia::Parser
   end
 end
 
-# rubocop:disable RSpec/FilePath
 describe FakeParser do
   it_behaves_like 'a Darlingtonia::Parser' do
     subject(:parser)   { described_class.new }
     let(:record_count) { 3 }
   end
 end
-# rubocop:enable RSpec/FilePath

data/lib/darlingtonia/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Darlingtonia
-  VERSION = '2.0.0'
+  VERSION = '2.1.0'
 end

data/spec/darlingtonia/hyrax_basic_metadata_mapper_spec.rb

@@ -6,7 +6,7 @@ describe Darlingtonia::HyraxBasicMetadataMapper do
 
   # Properties defined in Hyrax::CoreMetadata
   let(:core_fields) do
-    [:depositor, :title, :date_uploaded, :date_modified]
+    [:depositor, :title, :date_modified]
   end
 
   # Properties defined in Hyrax::BasicMetadata
@@ -49,7 +49,6 @@ describe Darlingtonia::HyraxBasicMetadataMapper do
 
     it 'returns single values for single-value fields' do
       expect(mapper.depositor).to eq 'someone@example.org'
-      expect(mapper.date_uploaded).to eq nil
       expect(mapper.date_modified).to eq nil
       expect(mapper.label).to eq nil
       expect(mapper.relative_path).to eq nil
@@ -83,4 +82,109 @@ describe Darlingtonia::HyraxBasicMetadataMapper do
       expect(mapper.delimiter).to eq 'ಠ_ಠ'
     end
   end
+
+  describe 'lenient headers' do
+    context 'headers with capital letters' do
+      before { mapper.metadata = metadata }
+      let(:metadata) do
+        { 'Title' => 'A Title',
+          'Related URL' => 'http://example.com',
+          'Abstract or Summary' => 'desc1|~|desc2',
+          'visiBILITY' => 'open',
+          'Depositor' => 'someone@example.org',
+          'DATE_modified' => 'mod date',
+          'laBel' => 'label',
+          'relative_PATH' => 'rel path',
+          'import_URL' => 'imp url' }
+      end
+
+      it 'matches the correct fields' do
+        expect(mapper.title).to eq ['A Title']
+        expect(mapper.related_url).to eq ['http://example.com']
+        expect(mapper.description).to eq ['desc1', 'desc2']
+        expect(mapper.creator).to eq []
+        expect(mapper.visibility).to eq 'open'
+        expect(mapper.depositor).to eq 'someone@example.org'
+        expect(mapper.date_modified).to eq 'mod date'
+        expect(mapper.label).to eq 'label'
+        expect(mapper.relative_path).to eq 'rel path'
+        expect(mapper.import_url).to eq 'imp url'
+      end
+    end
+
+    context 'headers with sloppy whitespace' do
+      before { mapper.metadata = metadata }
+      let(:metadata) do
+        { ' Title ' => 'A Title',
+          " Related URL \n " => 'http://example.com',
+          ' visiBILITY ' => 'open' }
+      end
+
+      it 'matches the correct fields' do
+        expect(mapper.title).to eq ['A Title']
+        expect(mapper.related_url).to eq ['http://example.com']
+        expect(mapper.visibility).to eq 'open'
+      end
+    end
+
+    context 'Visibility values in the CSV should match the Edit UI' do
+      load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
+      include_context 'with a work type'
+      context 'public is a synonym for open' do
+        before { mapper.metadata = metadata }
+        let(:metadata) do
+          { ' Title ' => 'A Title',
+            " Related URL \n " => 'http://example.com',
+            ' visiBILITY ' => 'PubLIC' }
+        end
+
+        it 'transforms public to open regardless of capitalization' do
+          expect(mapper.title).to eq ['A Title']
+          expect(mapper.related_url).to eq ['http://example.com']
+          expect(mapper.visibility).to eq 'open'
+        end
+      end
+      context 'institution name is a synonym for authenticated' do
+        before { mapper.metadata = metadata }
+        let(:metadata) do
+          { ' Title ' => 'A Title',
+            " Related URL \n " => 'http://example.com',
+            ' visiBILITY ' => 'my_institution' }
+        end
+
+        it 'transforms institution name to authenticated regardless of capitalization' do
+          expect(mapper.title).to eq ['A Title']
+          expect(mapper.related_url).to eq ['http://example.com']
+          expect(mapper.visibility).to eq 'authenticated'
+        end
+      end
+      context 'full institution name is a synonym for authenticated' do
+        before { mapper.metadata = metadata }
+        let(:metadata) do
+          { ' Title ' => 'A Title',
+            " Related URL \n " => 'http://example.com',
+            ' visiBILITY ' => 'my full institution name' }
+        end
+
+        it 'transforms full institution name to authenticated regardless of capitalization' do
+          expect(mapper.title).to eq ['A Title']
+          expect(mapper.related_url).to eq ['http://example.com']
+          expect(mapper.visibility).to eq 'authenticated'
+        end
+      end
+    end
+
+    # When someone accidentally has too many commas in the CSV rows
+    context 'headers with a nil' do
+      before { mapper.metadata = metadata }
+      let(:metadata) do
+        { ' Title ' => 'A Title',
+          nil => nil }
+      end
+
+      it 'doesn\'t raise an error for missing fields' do
+        expect(mapper.depositor).to eq nil
+      end
+    end
+  end
 end

data/spec/darlingtonia/hyrax_record_importer_spec.rb

@@ -6,19 +6,19 @@ describe Darlingtonia::HyraxRecordImporter, :clean do
     described_class.new(error_stream: error_stream, info_stream: info_stream)
   end
 
+  load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
+  include_context 'with a work type'
+
   let(:error_stream) { [] }
   let(:info_stream)  { [] }
   let(:record)       { Darlingtonia::InputRecord.from(metadata: metadata) }
 
   context 'collection id' do
     subject(:importer) do
-      described_class.new(collection_id: collection_id)
+      described_class.new(attributes: { collection_id: collection_id })
     end
     let(:collection_id) { '123' }
 
-    load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
-    include_context 'with a work type'
-
     it 'can have a collection id' do
       expect(importer.collection_id).to eq collection_id
     end
@@ -32,8 +32,6 @@ describe Darlingtonia::HyraxRecordImporter, :clean do
         'visibility' => 'open'
       }
     end
-    load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
-    include_context 'with a work type'
 
     it 'creates a work for record' do
       expect { importer.import(record: record) }
@@ -54,8 +52,6 @@ describe Darlingtonia::HyraxRecordImporter, :clean do
         'files' => 'darlingtonia.png|~|cat.png'
       }
     end
-    load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
-    include_context 'with a work type'
     it 'finds a file even if it is in a subdirectory' do
       expect(importer.find_file_path('cat.png')).to eq "#{ENV['IMPORT_PATH']}/animals/cat.png"
     end
@@ -71,33 +67,89 @@ describe Darlingtonia::HyraxRecordImporter, :clean do
     end
   end
 
-  context 'with and without a depositor value' do
-    context 'when there is no depositor set' do
-      let(:metadata) do
-        {
-          'title' => 'A Title',
-          'language' => 'English',
-          'visibility' => 'open'
-        }
+  context 'with attached files, alternate capitalization and whitespace in "files" header' do
+    before do
+      ENV['IMPORT_PATH'] = File.expand_path('../fixtures/images', File.dirname(__FILE__))
+    end
+    let(:metadata) do
+      {
+        'title' => 'A Title',
+        'visibility' => 'open',
+        ' Files' => 'darlingtonia.png|~|cat.png'
+      }
+    end
+
+    it 'makes an uploaded file object for each file attachment' do
+      expect { importer.import(record: record) }
+        .to change { Hyrax::UploadedFile.count }
+        .by 2
+    end
+  end
+
+  context 'with missing files' do
+    before do
+      ENV['IMPORT_PATH'] = File.expand_path('../fixtures/images', File.dirname(__FILE__))
+    end
+    it 'raises an exception' do
+      expect { importer.find_file_path('foo.png') }.to raise_exception(RuntimeError)
+    end
+  end
+
+  describe '#set_depositor' do
+    let(:metadata) { { 'title' => 'A Title' } }
+
+    context 'when no depositor is set' do
+      it 'sets the Hyrax default batch user' do
+        expect(importer.depositor.user_key).to eq 'batchuser@example.com'
       end
-      it 'adds the batch user as the depositor' do
-        importer.set_depositor(record)
-        expect(record.mapper.metadata["depositor"]).to eq "batchuser@example.com"
+    end
+
+    context 'when depositor is passed to initializer' do
+      subject(:importer) { described_class.new(error_stream: error_stream, info_stream: info_stream, attributes: { depositor_id: user.id }) }
+
+      let(:user) { ::User.new(id: '123', user_key: 'special_user@example.com') }
+      before { allow(::User).to receive(:find).and_return(user) }
+
+      it 'sets it to the passed-in depositor' do
+        expect(importer.depositor.user_key).to eq 'special_user@example.com'
       end
     end
-    context 'when there is a depositor set' do
+
+    context 'when depositor is set in metadata' do
       let(:metadata) do
-        {
-          'title' => 'A Title',
-          'language' => 'English',
-          'visibility' => 'open',
-          'depositor' => 'my@custom.depositor'
-        }
+        { 'title' => 'A Title',
+          'Depositor' => 'metadata_user@example.com' }
       end
-      it 'resets the batch user as the depositor' do
-        importer.set_depositor(record)
-        expect(record.mapper.metadata["depositor"]).to eq "batchuser@example.com"
+
+      it 'sets the Hyrax default batch user' do
+        expect(importer.depositor.user_key).to eq 'batchuser@example.com'
+        # TODO: expect(importer.depositor.user_key).to eq 'metadata_user@example.com'
+        # The metadata depositor should probably override any passed-in or default depositor.
       end
     end
   end
+  # When submitting location data (a.k.a., the "based near" attribute) via the UI,
+  # Hyrax expects to receive a `based_near_attributes` hash in a specific format.
+  # We need to take geonames urls as provided by the customer and transform them to
+  # mimic what the Hyrax UI would ordinarily produce. These will get turned into
+  # Hyrax::ControlledVocabularies::Location objects upon ingest.
+  context 'with location uris' do
+    let(:based_near) { ['http://www.geonames.org/5667009/montana.html', 'http://www.geonames.org/6252001/united-states.html'] }
+    let(:expected_bn_hash) do
+      {
+        "0" => {
+          "id" => "http://sws.geonames.org/5667009/", "_destroy" => ""
+        },
+        "1" => {
+          "id" => "http://sws.geonames.org/6252001/", "_destroy" => ""
+        }
+      }
+    end
+    it "gets a sws uri from a geonames uri" do
+      expect(importer.uri_to_sws("http://www.geonames.org/6252001/united-states.html")).to eq "http://sws.geonames.org/6252001/"
+    end
+    it 'transforms an array of geonames uris into the expected based_near_attributes hash' do
+      expect(importer.based_near_attributes(based_near)).to eq expected_bn_hash
+    end
+  end
 end

data/spec/darlingtonia/importer_spec.rb

@@ -3,6 +3,9 @@
 require 'spec_helper'
 
 describe Darlingtonia::Importer do
+  load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
+  include_context 'with a work type'
+
   subject(:importer) { described_class.new(parser: parser) }
   let(:parser)       { FakeParser.new(file: input) }
   let(:input)        { [{ 'title' => '1' }, { 'title' => '2' }, { 'title' => '3' }] }

data/spec/darlingtonia/input_record_spec.rb

@@ -3,6 +3,8 @@
 require 'spec_helper'
 
 describe Darlingtonia::InputRecord do
+  load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
+  include_context 'with a work type'
   subject(:record) { described_class.from(metadata: metadata) }
 
   let(:metadata) do

data/spec/darlingtonia_spec.rb

@@ -5,21 +5,15 @@ require 'spec_helper'
 describe Darlingtonia do
   describe '#config' do
     it 'can set a default error stream' do
-      stream = []
-
-      expect { described_class.config { |c| c.default_error_stream = stream } }
+      expect { described_class.config { |c| c.default_error_stream = STDOUT } }
         .to change { described_class.config.default_error_stream }
-        .from(STDOUT)
-        .to(stream)
+        .to(STDOUT)
     end
 
     it 'can set a default info stream' do
-      stream = []
-
-      expect { described_class.config { |c| c.default_info_stream = stream } }
+      expect { described_class.config { |c| c.default_info_stream = STDOUT } }
         .to change { described_class.config.default_info_stream }
-        .from(STDOUT)
-        .to(stream)
+        .to(STDOUT)
     end
   end
 end

data/spec/fixtures/hyrax/example.csv

@@ -1,3 +1,3 @@
-title,depositor,date_uploaded,date_modified,label,relative_path,import_url,resource type,creator,contributor,abstract or summary,keyword,license,rights statement,publisher,date created,subject,language,identifier,location,related url,bibliographic_citation,source
-Work 1 Title,user@example.com,2018-12-21,2018-01-01,Work 1 Label,tmp/files,https://example.com,Work 1 Type,Work 1 creator,Work 1 contrib,Desc 1,Key 1,Lic 1,RS 1,Pub 1,2018-06-06,Subj 1,English|~|Japanese,Ident 1,Based 1,https://example.com/related,Bib 1,Source 1
-Work 2 Title,,1970-12-21,,Work 2 Label,,,Work 2 Type,,,Desc 2,,,,Pub 2,,Subj 2
+title,depositor,date_modified,label,relative_path,import_url,resource type,creator,contributor,abstract or summary,keyword,license,rights statement,publisher,date created,subject,language,identifier,location,related url,bibliographic_citation,source
+Work 1 Title,user@example.com,2018-01-01,Work 1 Label,tmp/files,https://example.com,Work 1 Type,Work 1 creator,Work 1 contrib,Desc 1,Key 1,Lic 1,RS 1,Pub 1,2018-06-06,Subj 1,English|~|Japanese,Ident 1,Based 1,https://example.com/related,Bib 1,Source 1
+Work 2 Title,1970-12-21,,Work 2 Label,,,Work 2 Type,,,Desc 2,,,,Pub 2,,Subj 2

data/spec/integration/import_hyrax_csv.rb

@@ -2,17 +2,13 @@
 require 'spec_helper'
 
 describe 'importing a CSV with Hyrax defaults', :clean do
-  subject(:importer) { Darlingtonia::Importer.new(parser: parser) }
+  subject(:importer) { Darlingtonia::Importer.new(parser: parser, record_importer: record_importer) }
   let(:parser) { Darlingtonia::CsvParser.new(file: csv_file) }
+  let(:record_importer) { Darlingtonia::HyraxRecordImporter.new }
 
   let(:csv_file) { File.open('spec/fixtures/hyrax/example.csv') }
   after { csv_file.close }
 
-  before do
-    # Force it to use the Hyrax mapper instead of the default mapper
-    allow(Darlingtonia::HashMapper).to receive(:new).and_return(Darlingtonia::HyraxBasicMetadataMapper.new)
-  end
-
   load File.expand_path("../../support/shared_contexts/with_work_type.rb", __FILE__)
   include_context 'with a work type'
 
@@ -24,8 +20,7 @@ describe 'importing a CSV with Hyrax defaults', :clean do
     work2 = works.find { |w| w.title == ['Work 2 Title'] }
 
     # First Record
-    expect(work1.depositor).to eq 'user@example.com'
-    expect(work1.date_uploaded).to eq '2018-12-21'
+    expect(work1.depositor).to eq 'batchuser@example.com'
     expect(work1.date_modified).to eq '2018-01-01'
     expect(work1.label).to eq 'Work 1 Label'
     expect(work1.relative_path).to eq 'tmp/files'
@@ -51,8 +46,7 @@ describe 'importing a CSV with Hyrax defaults', :clean do
     expect(work1.source).to eq ['Source 1']
 
     # Second Record
-    expect(work2.depositor).to be_nil
-    expect(work2.date_uploaded).to eq '1970-12-21'
+    expect(work2.depositor).to eq 'batchuser@example.com'
     expect(work2.date_modified).to be_nil
     expect(work2.label).to eq 'Work 2 Label'
     expect(work2.relative_path).to be_nil

data/spec/support/shared_contexts/with_work_type.rb

@@ -8,16 +8,22 @@ shared_context 'with a work type' do
 
     class Work < ActiveFedora::Base
       attr_accessor :visibility
+      attr_accessor :based_near_attributes
       include ::Hyrax::CoreMetadata
       include ::Hyrax::BasicMetadata
     end
 
-    class User
+    class User < Struct.new(:id, :user_key)
+      def initialize(inputs = {})
+        self.id = inputs[:id]
+        self.user_key = inputs[:user_key] || batch_user_key
+      end
+
       def self.find_or_create_system_user(_email)
         User.new
       end
 
-      def user_key
+      def batch_user_key
         'batchuser@example.com'
       end
     end
@@ -37,6 +43,16 @@ shared_context 'with a work type' do
         end
       end
 
+      class Institution
+        def self.name
+          'my_institution'
+        end
+
+        def self.name_full
+          'my full institution name'
+        end
+      end
+
       class UploadedFile < ActiveFedora::Base
         def self.create(*)
           h = Hyrax::UploadedFile.new
@@ -47,14 +63,20 @@ shared_context 'with a work type' do
 
       module Actors
         class Environment
-          def initialize(new_object, ability, attributes); end
+          attr_reader :new_object, :attributes
+          def initialize(new_object, _ability, attributes)
+            @new_object = new_object
+            @attributes = attributes
+          end
         end
       end
 
       class Actor
-        def create(_actor_env)
-          Work.create
-          true
+        def create(actor_env)
+          attrs = actor_env.attributes
+          attrs.delete(:uploaded_files)
+          actor_env.new_object.attributes = attrs
+          actor_env.new_object.save
         end
       end
 
@@ -68,6 +90,8 @@ shared_context 'with a work type' do
 
   after do
     Object.send(:remove_const, :Hyrax) if defined?(Hyrax)
-    Object.send(:remove_const, :Work)  if defined?(Work)
+    Object.send(:remove_const, :Work) if defined?(Work)
+    Object.send(:remove_const, :User) if defined?(User)
+    Object.send(:remove_const, :Ability) if defined?(Ability)
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: darlingtonia
 version: !ruby/object:Gem::Version
-  version: 2.0.0
+  version: 2.1.0
 platform: ruby
 authors:
 - Data Curation Experts
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-02-04 00:00:00.000000000 Z
+date: 2019-02-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: active-fedora
@@ -44,14 +44,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.3'
+        version: '1.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.3'
+        version: '1.0'
 - !ruby/object:Gem::Dependency
   name: hyrax-spec
   requirement: !ruby/object:Gem::Requirement
@@ -136,6 +136,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: rake
+  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: 
 email:
 - administrator@curationexperts.com
@@ -160,6 +174,7 @@ files:
 - lib/darlingtonia/hyrax_record_importer.rb
 - lib/darlingtonia/importer.rb
 - lib/darlingtonia/input_record.rb
+- lib/darlingtonia/log_stream.rb
 - lib/darlingtonia/metadata_mapper.rb
 - lib/darlingtonia/parser.rb
 - lib/darlingtonia/parsers/csv_parser.rb