gdor-indexer 0.7.1 → 0.8.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 37ead667bc97ed2a13e444026fe72ebb79f661fb
-  data.tar.gz: 418391aec671bf772462548a9aa20cf23ed2ea7c
+  metadata.gz: d3e6a69d961cda79c503df64ba9f6e96acd97a91
+  data.tar.gz: 4409fdd1c2379e5277a021d332e4be96178f8ec7
 SHA512:
-  metadata.gz: 68c0d4599bd65d199d84e9664f5c2c6f09896fb9273d2832f9e71cd58072bf3aef32ccf5a52e6ecdfc23315b200a4a464ec8f0248687df414d66f0267d89d9ff
-  data.tar.gz: 6a64eaab7d20112d9b6b59fd5d0673f2f8855c91f1d5be95cf6c64d6e3ec83eefdcb106e9ef012a906f5a6f8d2211c1b766123ce66df6860bf28bc65a2a10b7b
+  metadata.gz: 0bb42c00986ea0fcbcaa2ef3765467da74badd29328f53e19b2337a8f0e643cc8cecbb0b95240a01abed0ad2c97984fc85bf7e8aa09ee20735d95b13c512de78
+  data.tar.gz: f72607a1e4eba439825b9bb2a649271483b329b60d91aa4deb4a902074bc8b78c70ad28af026dd139d92533bcbc0f756f94330fe776758dc1d68cecd28023e0e

data/.rubocop_todo.yml

@@ -2,7 +2,7 @@ require: rubocop-rspec
 
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2016-08-30 14:13:02 -0700 using RuboCop version 0.42.0.
+# on 2016-09-22 06:51:42 -0700 using RuboCop version 0.43.0.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
@@ -20,7 +20,7 @@ Metrics/AbcSize:
 # Offense count: 1
 # Configuration parameters: CountComments.
 Metrics/ClassLength:
-  Max: 242
+  Max: 240
 
 # Offense count: 5
 Metrics/CyclomaticComplexity:
@@ -32,7 +32,7 @@ Metrics/CyclomaticComplexity:
 Metrics/LineLength:
   Max: 258
 
-# Offense count: 15
+# Offense count: 14
 # Configuration parameters: CountComments.
 Metrics/MethodLength:
   Max: 43
@@ -70,7 +70,7 @@ RSpec/FilePath:
     - 'spec/unit/solr_doc_builder_spec.rb'
     - 'spec/unit/solr_doc_hash_spec.rb'
 
-# Offense count: 163
+# Offense count: 158
 # Configuration parameters: AssignmentOnly.
 RSpec/InstanceVariable:
   Exclude:
@@ -119,14 +119,6 @@ RSpec/VerifiedDoubles:
   Exclude:
     - 'spec/unit/indexer_spec.rb'
 
-# Offense count: 7
-# Cop supports --auto-correct.
-# Configuration parameters: EnforcedStyle, SupportedStyles.
-# SupportedStyles: braces, no_braces, context_dependent
-Style/BracesAroundHashParameters:
-  Exclude:
-    - 'spec/unit/indexer_spec.rb'
-
 # Offense count: 4
 # Configuration parameters: EnforcedStyle, SupportedStyles.
 # SupportedStyles: nested, compact
@@ -137,20 +129,14 @@ Style/ClassAndModuleChildren:
     - 'lib/gdor/indexer/solr_doc_builder.rb'
     - 'lib/gdor/indexer/solr_doc_hash.rb'
 
-# Offense count: 22
-# Cop supports --auto-correct.
-# Configuration parameters: EnforcedStyle, SupportedStyles, IndentationWidth.
-# SupportedStyles: aligned, indented
-Style/MultilineOperationIndentation:
-  Exclude:
-    - 'Rakefile'
-    - 'spec/unit/mods_pub_fields_spec.rb'
-
 # Offense count: 2
-# Cop supports --auto-correct.
-# Configuration parameters: EnforcedStyle, SupportedStyles, AllowInnerSlashes.
-# SupportedStyles: slashes, percent_r, mixed
-Style/RegexpLiteral:
+# Configuration parameters: MinBodyLength.
+Style/GuardClause:
   Exclude:
-    - 'spec/unit/indexer_spec.rb'
-    - 'spec/unit/mods_subject_fields_spec.rb'
+    - 'lib/gdor/indexer.rb'
+
+# Offense count: 4
+# Configuration parameters: SupportedStyles.
+# SupportedStyles: snake_case, normalcase, non_integer
+Style/VariableNumber:
+  Enabled: false

data/Rakefile

@@ -46,7 +46,7 @@ begin
 
   YARD::Rake::YardocTask.new(:doc) do |yt|
     yt.files = Dir.glob(File.join(project_root, 'lib', '**', '*.rb')) +
-      [File.join(project_root, 'README.rdoc')]
+               [File.join(project_root, 'README.rdoc')]
     yt.options = ['--output-dir', doc_dest_dir, '--readme', 'README.rdoc', '--title', 'Gryphondor Indexer Documentation']
   end
 rescue LoadError

data/gdor-indexer.gemspec

@@ -18,7 +18,7 @@ Gem::Specification.new do |spec|
   spec.require_paths = ['lib']
 
   spec.add_dependency 'harvestdor-indexer'
-  spec.add_dependency 'stanford-mods', '~> 2.2.1'
+  spec.add_dependency 'stanford-mods', '~> 2.2', '>= 2.2.1'
   spec.add_dependency 'nokogiri'
   spec.add_dependency 'rsolr'
   spec.add_dependency 'activesupport'

data/lib/gdor/indexer.rb

@@ -128,9 +128,10 @@ module GDor
       raise e
     end
 
-    # create Solr doc for the druid and add it to Solr, unless it is on the blacklist.
-    #  NOTE: don't forget to send commit to Solr, either once at end (already in harvest_and_index), or for each add, or ...
+    # Create Solr doc for the druid and add it to Solr, unless it is on the blacklist.
     # @param [Harvestdor::Indexer::Resource] resource an item record (a member of a collection)
+    # @return [Hash]
+    # @note don't forget to send commit to Solr, either once at end (already in harvest_and_index), or for each add, or ...
     def item_solr_document(resource)
       sdb = GDor::Indexer::SolrDocBuilder.new(resource, logger)
 
@@ -154,7 +155,7 @@ module GDor
     end
 
     # Create Solr document for the collection druid suitable for SearchWorks
-    #  and write the result to the SearchWorks Solr Index
+    # and write the result to the SearchWorks Solr Index
     # @param [Harvestdor::Indexer::Resource] resource a collection record
     # @return [Hash]
     def collection_solr_document(resource)
@@ -179,7 +180,7 @@ module GDor
       doc_hash.to_h
     end
 
-    # add coll level data to this solr doc and/or cache collection level information
+    # Add collection level data to this solr doc and/or cache collection level information
     # @param [Hash] doc_hash representing the Solr document (for an item)
     # @param [Array<Harvestdor::Indexer::Resource>] collections the collections the item is a member of
     def add_coll_info(doc_hash, collections)

data/lib/gdor/indexer/solr_doc_builder.rb

@@ -43,7 +43,7 @@ class GDor::Indexer::SolrDocBuilder
   # if not found, we look for
   #  identityMetadata/otherId[@name='barcode']
   #   if found, we look for catkey in MODS
-  #     mods/recordInfo/recordIdentifier[@source="SIRSI"
+  #     mods/recordInfo/recordIdentifier[@source="SIRSI"]
   #     and if found, remove the leading a
   # otherwise, nil
   def catkey

data/lib/gdor/indexer/version.rb

@@ -1,5 +1,5 @@
 module GDor
   class Indexer
-    VERSION = '0.7.1'.freeze
+    VERSION = '0.8.0'.freeze
   end
 end

data/spec/unit/indexer_spec.rb

@@ -228,12 +228,12 @@ describe GDor::Indexer do
         expect(doc_hash).to include format_main_ssim: 'Archive/Manuscript'
       end
       it 'other values present' do
-        allow_any_instance_of(GDor::Indexer::SolrDocBuilder).to receive(:doc_hash).and_return(GDor::Indexer::SolrDocHash.new({ format_main_ssim: %w(Image Video) }))
+        allow_any_instance_of(GDor::Indexer::SolrDocBuilder).to receive(:doc_hash).and_return(GDor::Indexer::SolrDocHash.new(format_main_ssim: %w(Image Video)))
         doc_hash = @indexer.collection_solr_document collection
         expect(doc_hash).to include format_main_ssim: ['Image', 'Video', 'Archive/Manuscript']
       end
       it 'already has values Archive/Manuscript' do
-        allow_any_instance_of(GDor::Indexer::SolrDocBuilder).to receive(:doc_hash).and_return(GDor::Indexer::SolrDocHash.new({ format_main_ssim: 'Archive/Manuscript' }))
+        allow_any_instance_of(GDor::Indexer::SolrDocBuilder).to receive(:doc_hash).and_return(GDor::Indexer::SolrDocHash.new(format_main_ssim: 'Archive/Manuscript'))
         doc_hash = @indexer.collection_solr_document collection
         expect(doc_hash).to include format_main_ssim: ['Archive/Manuscript']
       end
@@ -246,9 +246,7 @@ describe GDor::Indexer do
   end #  index_coll_obj_per_config
 
   context '#add_coll_info and supporting methods' do
-    before do
-      @coll_druids_array = [collection]
-    end
+    let(:coll_druids_array) { [collection] }
     let(:doc_hash) { GDor::Indexer::SolrDocHash.new({}) }
 
     it 'adds no collection field values to doc_hash if there are none' do
@@ -260,7 +258,7 @@ describe GDor::Indexer do
 
     context 'collection field' do
       it 'is added field to doc hash' do
-        @indexer.add_coll_info(doc_hash, @coll_druids_array)
+        @indexer.add_coll_info(doc_hash, coll_druids_array)
         expect(doc_hash[:collection]).to match_array [@coll_druid_from_test_config]
       end
       it 'adds two values to doc_hash when object belongs to two collections' do
@@ -293,18 +291,18 @@ describe GDor::Indexer do
       end
       it 'gets single item display_type for single collection (and no dups)' do
         allow(@indexer).to receive(:identity_md_obj_label)
-        doc_hash = GDor::Indexer::SolrDocHash.new({ display_type: 'image' })
-        @indexer.add_coll_info(doc_hash, @coll_druids_array)
-        doc_hash = GDor::Indexer::SolrDocHash.new({ display_type: 'image' })
-        @indexer.add_coll_info(doc_hash, @coll_druids_array)
+        doc_hash = GDor::Indexer::SolrDocHash.new(display_type: 'image')
+        @indexer.add_coll_info(doc_hash, coll_druids_array)
+        doc_hash = GDor::Indexer::SolrDocHash.new(display_type: 'image')
+        @indexer.add_coll_info(doc_hash, coll_druids_array)
         expect(@indexer.coll_display_types_from_items(collection)).to match_array ['image']
       end
       it 'gets multiple formats from multiple items for single collection' do
         allow(@indexer).to receive(:identity_md_obj_label)
-        doc_hash = GDor::Indexer::SolrDocHash.new({ display_type: 'image' })
-        @indexer.add_coll_info(doc_hash, @coll_druids_array)
-        doc_hash = GDor::Indexer::SolrDocHash.new({ display_type: 'file' })
-        @indexer.add_coll_info(doc_hash, @coll_druids_array)
+        doc_hash = GDor::Indexer::SolrDocHash.new(display_type: 'image')
+        @indexer.add_coll_info(doc_hash, coll_druids_array)
+        doc_hash = GDor::Indexer::SolrDocHash.new(display_type: 'file')
+        @indexer.add_coll_info(doc_hash, coll_druids_array)
         expect(@indexer.coll_display_types_from_items(collection)).to match_array %w(image file)
       end
     end # coll_display_types_from_items
@@ -360,7 +358,7 @@ describe GDor::Indexer do
     end
 
     it 'email includes reference to full log' do
-      expect(subject).to match(/full log is at gdor_indexer\/shared\/spec\/test_logs\/testcoll\.log/)
+      expect(subject).to match(%r{full log is at gdor_indexer/shared/spec/test_logs/testcoll\.log})
     end
   end
 
@@ -375,7 +373,6 @@ describe GDor::Indexer do
       expect(@indexer).to receive(:send_email) do |_to, opts|
         expect(opts[:subject]).to match(/is finished/)
       end
-
       @indexer.email_results
     end
 
@@ -383,7 +380,6 @@ describe GDor::Indexer do
       expect(@indexer).to receive(:send_email) do |to, _opts|
         expect(to).to eq @indexer.config.notification
       end
-
       @indexer.email_results
     end
 
@@ -391,7 +387,6 @@ describe GDor::Indexer do
       expect(@indexer).to receive(:send_email) do |_to, opts|
         expect(opts[:body]).to eq 'Report Body'
       end
-
       @indexer.email_results
     end
   end
@@ -410,7 +405,7 @@ describe GDor::Indexer do
         expect(mail.from).to match_array ['rspec']
         expect(mail.body).to eq 'Body'
       end
-      @indexer.send_email 'notification-list@example.com', { from: 'rspec', subject: 'Subject', body: 'Body' }
+      @indexer.send_email 'notification-list@example.com', from: 'rspec', subject: 'Subject', body: 'Body'
     end
   end
 
@@ -422,7 +417,6 @@ describe GDor::Indexer do
     it 'can be set as an option' do
       solr_client = double
       @indexer = described_class.new(solr_client: solr_client)
-
       expect(@indexer.solr_client).to eq solr_client
     end
   end

data/spec/unit/mods_fields_spec.rb

@@ -311,6 +311,5 @@ describe GDor::Indexer::ModsFields do
   end # author fields
 
   # subject fields moved to mods_subject_fields_spec.rb
-
   # publication fields moved to mods_pub_fields_spec.rb
 end

data/spec/unit/mods_pub_fields_spec.rb

@@ -23,7 +23,7 @@ describe GDor::Indexer::ModsFields do
     RSpec.shared_examples 'expected (dateIssued)' do |solr_field_sym, mods_field_val, exp_val|
       it "#{exp_val} for #{mods_field_val}" do
         m = mods_origin_info_start_str +
-              "<dateIssued>#{mods_field_val}</dateIssued>" +
+            "<dateIssued>#{mods_field_val}</dateIssued>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[solr_field_sym]).to eq exp_val
@@ -62,14 +62,14 @@ describe GDor::Indexer::ModsFields do
       end
       it 'includes approx dates' do
         m = mods_origin_info_start_str +
-              "<dateIssued qualifier='approximate'>1945</dateIssued>" +
+            "<dateIssued qualifier='approximate'>1945</dateIssued>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_isi]).to eq 1945
       end
       it 'will use dateCreated' do
         m = mods_origin_info_start_str +
-              "<dateCreated>1904</dateCreated>" +
+            "<dateCreated>1904</dateCreated>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_isi]).to eq 1904
@@ -85,14 +85,14 @@ describe GDor::Indexer::ModsFields do
       end
       it 'includes approx dates' do
         m = mods_origin_info_start_str +
-              "<dateIssued qualifier='approximate'>1945</dateIssued>" +
+            "<dateIssued qualifier='approximate'>1945</dateIssued>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_date_sort]).to eq('1945')
       end
       it 'takes single dateCreated' do
         m = mods_origin_info_start_str +
-              "<dateCreated>1904</dateCreated>" +
+            "<dateCreated>1904</dateCreated>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_date_sort]).to eq('1904')
@@ -117,16 +117,16 @@ describe GDor::Indexer::ModsFields do
     context 'pub_year_no_approx_isi' do
       it 'does not include approx dates' do
         m = mods_origin_info_start_str +
-              "<dateIssued qualifier='approximate'>1945</dateIssued>" +
-          mods_origin_info_end_str
+            "<dateIssued qualifier='approximate'>1945</dateIssued>" +
+            mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_no_approx_isi]).to eq nil
       end
       it 'calls Stanford::Mods::Record instance pub_year_int(true)' do
         mods = mods_origin_info_start_str +
-                "<dateIssued qualifier=\"approximate\">1500</dateIssued>
-                <dateIssued>2000</dateIssued>" +
-              mods_origin_info_end_str
+               "<dateIssued qualifier=\"approximate\">1500</dateIssued>
+               <dateIssued>2000</dateIssued>" +
+               mods_origin_info_end_str
         sdb = sdb_for_mods(mods)
         expect(sdb.smods_rec).to receive(:pub_year_int).with(true).and_call_original
         allow(sdb.smods_rec).to receive(:pub_year_int).with(false) # for other flavor
@@ -134,7 +134,7 @@ describe GDor::Indexer::ModsFields do
       end
       it 'will use dateCreated' do
         m = mods_origin_info_start_str +
-              "<dateCreated>1904</dateCreated>" +
+            "<dateCreated>1904</dateCreated>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_no_approx_isi]).to eq 1904
@@ -146,16 +146,16 @@ describe GDor::Indexer::ModsFields do
     context 'pub_year_w_approx_isi' do
       it 'includes approx dates' do
         m = mods_origin_info_start_str +
-              "<dateIssued qualifier='approximate'>1945</dateIssued>" +
-          mods_origin_info_end_str
+            "<dateIssued qualifier='approximate'>1945</dateIssued>" +
+            mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_w_approx_isi]).to eq 1945
       end
       it 'calls Stanford::Mods::Record instance pub_year_int(false)' do
         mods = mods_origin_info_start_str +
-                "<dateIssued qualifier=\"approximate\">1500</dateIssued>
-                <dateIssued>2000</dateIssued>" +
-              mods_origin_info_end_str
+               "<dateIssued qualifier=\"approximate\">1500</dateIssued>
+               <dateIssued>2000</dateIssued>" +
+               mods_origin_info_end_str
         sdb = sdb_for_mods(mods)
         expect(sdb.smods_rec).to receive(:pub_year_int).with(false).and_call_original
         allow(sdb.smods_rec).to receive(:pub_year_int).with(true) # for other flavor
@@ -163,7 +163,7 @@ describe GDor::Indexer::ModsFields do
       end
       it 'will use dateCreated' do
         m = mods_origin_info_start_str +
-              "<dateCreated>1904</dateCreated>" +
+            "<dateCreated>1904</dateCreated>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_w_approx_isi]).to eq 1904
@@ -183,14 +183,14 @@ describe GDor::Indexer::ModsFields do
       end
       it 'includes approx dates' do
         m = mods_origin_info_start_str +
-              "<dateIssued qualifier='approximate'>1945</dateIssued>" +
-          mods_origin_info_end_str
+            "<dateIssued qualifier='approximate'>1945</dateIssued>" +
+            mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_tisim]).to eq 1945
       end
       it 'will use dateCreated' do
         m = mods_origin_info_start_str +
-              "<dateCreated>1904</dateCreated>" +
+            "<dateCreated>1904</dateCreated>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.doc_hash_from_mods[:pub_year_tisim]).to eq 1904
@@ -208,8 +208,8 @@ describe GDor::Indexer::ModsFields do
     context 'creation_year_isi' do
       it 'creation_year_isi calls Stanford::Mods::Record year_int for dateCreated elements' do
         m = mods_origin_info_start_str +
-              "<dateCreated qualifier='approximate'>1500</dateCreated>
-              <dateIssued qualifier='approximate'>2000</dateIssued>" +
+            "<dateCreated qualifier='approximate'>1500</dateCreated>
+            <dateIssued qualifier='approximate'>2000</dateIssued>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.smods_rec).to receive(:year_int).at_least(2).times.and_call_original
@@ -218,7 +218,7 @@ describe GDor::Indexer::ModsFields do
       RSpec.shared_examples 'expected (dateCreated)' do |mods_field_val, exp_val|
         it "#{exp_val} for #{mods_field_val}" do
           m = mods_origin_info_start_str +
-                "<dateCreated>#{mods_field_val}</dateCreated>" +
+              "<dateCreated>#{mods_field_val}</dateCreated>" +
               mods_origin_info_end_str
           sdb = sdb_for_mods(m)
           expect(sdb.doc_hash_from_mods[:creation_year_isi]).to eq exp_val
@@ -248,8 +248,8 @@ describe GDor::Indexer::ModsFields do
     context 'publication_year_isi' do
       it 'publication_year_isi calls Stanford::Mods::Record year_int for dateIssued elements' do
         m = mods_origin_info_start_str +
-              "<dateCreated qualifier='approximate'>1500</dateCreated>
-              <dateIssued qualifier='approximate'>2000</dateIssued>" +
+            "<dateCreated qualifier='approximate'>1500</dateCreated>
+            <dateIssued qualifier='approximate'>2000</dateIssued>" +
             mods_origin_info_end_str
         sdb = sdb_for_mods(m)
         expect(sdb.smods_rec).to receive(:year_int).at_least(2).times.and_call_original

data/spec/unit/mods_subject_fields_spec.rb

@@ -105,7 +105,7 @@ describe GDor::Indexer::ModsFields do
         it "logs an info message when it encounters a geographicCode encoding it doesn't translate" do
           m = "<mods #{ns_decl}><subject><geographicCode authority='iso3166'>ca</geographicCode></subject></mods>"
           sdb = sdb_for_mods(m)
-          expect(sdb.smods_rec.sw_logger).to receive(:info).with(/#{fake_druid} has subject geographicCode element with untranslated encoding \(iso3166\): <geographicCode authority=.*>ca<\/geographicCode>/).at_least(1).times
+          expect(sdb.smods_rec.sw_logger).to receive(:info).with(%r{#{fake_druid} has subject geographicCode element with untranslated encoding \(iso3166\): <geographicCode authority=.*>ca</geographicCode>}).at_least(1).times
           sdb.doc_hash_from_mods
         end
       end # geographic_search

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gdor-indexer
 version: !ruby/object:Gem::Version
-  version: 0.7.1
+  version: 0.8.0
 platform: ruby
 authors:
 - Naomi Dushay
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-08-30 00:00:00.000000000 Z
+date: 2016-09-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: harvestdor-indexer
@@ -31,6 +31,9 @@ dependencies:
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.2'
+    - - ">="
       - !ruby/object:Gem::Version
         version: 2.2.1
   type: :runtime
@@ -38,6 +41,9 @@ dependencies:
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
+      - !ruby/object:Gem::Version
+        version: '2.2'
+    - - ">="
       - !ruby/object:Gem::Version
         version: 2.2.1
 - !ruby/object:Gem::Dependency