gdor-indexer 0.4.0 → 0.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b7d3fef57a3dcc4df36487bead9f96a691ae385d
-  data.tar.gz: f724d1e3bfc0aee77798a10feb3c06939b29c30a
+  metadata.gz: b01625bc9ae63bd07de7edbee92754e2d5c3ce82
+  data.tar.gz: 9586786a5f096aeaf9a7cacac5aa4d07a846ac38
 SHA512:
-  metadata.gz: bbed1557cd19168a20ca5c26656186206ef80ea83817715fea27fb79218f45ece95f24186f9eeb4fdb17c87b2ccf147556398dcba2fe5bed02db6983951dbe7d
-  data.tar.gz: 578889a4be5005e8a6504313e46b082685a393937a46cd4fe4315756b8da1191ef7a383ec122b5586f6da6ab76ba3157ee325fedf9c87cf67edb24d1cacf0e67
+  metadata.gz: c82da4f86166c0864a0ad5ed969304b8ee8a357d07efa71738d7a7ce50976cb509568ff812b9ca3b28a44d514237a37ff69eb43a8a617cd28e6aadf2e6794378
+  data.tar.gz: 9bc2a66dbf0652206ca609b5658eebf0d80d32695c0ccf03cace3bf58506650dc212f7b49e09bf6ecfb50c2f6f1c5a5389a913345318e1bf3b212f74b0dcdad9

data/.rubocop_todo.yml

@@ -1,6 +1,6 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2016-01-06 15:25:10 -0800 using RuboCop version 0.34.2.
+# on 2016-01-07 14:25:59 -0800 using RuboCop version 0.34.2.
 # 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
@@ -24,9 +24,9 @@ Lint/UselessAssignment:
     - 'spec/unit/indexer_spec.rb'
     - 'spec/unit/solr_doc_builder_spec.rb'
 
-# Offense count: 21
+# Offense count: 22
 Metrics/AbcSize:
-  Max: 108
+  Max: 82
 
 # Offense count: 1
 # Configuration parameters: CountComments.
@@ -37,15 +37,15 @@ Metrics/ClassLength:
 Metrics/CyclomaticComplexity:
   Max: 9
 
-# Offense count: 307
+# Offense count: 309
 # Configuration parameters: AllowURI, URISchemes.
 Metrics/LineLength:
   Max: 258
 
-# Offense count: 15
+# Offense count: 16
 # Configuration parameters: CountComments.
 Metrics/MethodLength:
-  Max: 51
+  Max: 43
 
 # Offense count: 3
 Metrics/PerceivedComplexity:
@@ -70,6 +70,12 @@ RSpec/InstanceVariable:
     - 'spec/unit/public_xml_fields_spec.rb'
     - 'spec/unit/solr_doc_builder_spec.rb'
 
+# Offense count: 6
+# Cop supports --auto-correct.
+Style/AlignArray:
+  Exclude:
+    - 'lib/gdor/indexer/mods_fields.rb'
+
 # Offense count: 7
 # Cop supports --auto-correct.
 # Configuration parameters: EnforcedStyle, SupportedStyles.
@@ -100,7 +106,7 @@ Style/DoubleNegation:
   Exclude:
     - 'lib/gdor/indexer/solr_doc_hash.rb'
 
-# Offense count: 1
+# Offense count: 2
 # Cop supports --auto-correct.
 # Configuration parameters: EnforcedStyle, SupportedStyles.
 Style/EmptyLinesAroundBlockBody:
@@ -113,7 +119,7 @@ Style/GuardClause:
   Exclude:
     - 'lib/gdor/indexer.rb'
 
-# Offense count: 12
+# Offense count: 14
 # Cop supports --auto-correct.
 # Configuration parameters: EnforcedStyle, SupportedStyles.
 Style/MultilineOperationIndentation:

data/Rakefile

@@ -18,10 +18,10 @@ end
 #  sh 'irb -rubygems -I lib  -r ./frda_indexer.rb'
 # end
 
-task default: [:ci, :rubocop]
+task default: :ci
 
-desc 'run continuous integration suite (tests, coverage, docs)'
-task ci: [:rspec, :doc, :rubocop]
+desc 'run continuous integration suite'
+task ci: [:rspec, :rubocop]
 
 task spec: :rspec
 

data/gdor-indexer.gemspec

@@ -8,7 +8,7 @@ Gem::Specification.new do |spec|
   spec.version       = GDor::Indexer::VERSION
   spec.authors       = ['Naomi Dushay', 'Laney McGlohon', 'Chris Beer']
   spec.email         = ['cabeer@stanford.edu']
-  spec.summary       = 'Gryphondor Solr indexing logic'
+  spec.summary       = 'PURL doc =>  Solr hash logic'
   spec.homepage      = 'https://github.com/sul-dlss/gdor-indexer'
   spec.license       = 'Apache 2'
 
@@ -35,10 +35,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency 'rubocop-rspec'
   spec.add_development_dependency 'simplecov'
   spec.add_development_dependency 'equivalent-xml', '~> 0.5'
-  spec.add_development_dependency 'capybara'
-  spec.add_development_dependency 'poltergeist', '>= 1.5.0'
   spec.add_development_dependency 'vcr'
-  spec.add_development_dependency 'jettywrapper'
   spec.add_development_dependency 'webmock'
   spec.add_development_dependency 'pry-byebug'
 end

data/lib/gdor/indexer/mods_fields.rb

@@ -4,6 +4,7 @@ module GDor::Indexer::ModsFields
   # Create a Hash representing a Solr doc, with all MODS related fields populated.
   # @return [Hash] Hash representing the Solr document
   def doc_hash_from_mods
+    sort_str_w_approx_dates = smods_rec.pub_date_sortable_string(false)
     doc_hash = {
       # title fields
       title_245a_search: smods_rec.sw_short_title,
@@ -45,10 +46,10 @@ module GDor::Indexer::ModsFields
 
       # publication fields
       pub_search: smods_rec.place,
-      pub_date_sort: smods_rec.pub_date_sortable_string(false), # include approx dates
+      pub_date_sort: sort_str_w_approx_dates,
       # these are for single value facet display (in leiu of date slider (pub_year_tisim) and deprecated pub_date)
-      pub_year_no_approx_isi: smods_rec.pub_date_facet_single_value(true),
-      pub_year_w_approx_isi: smods_rec.pub_date_facet_single_value(false),
+      pub_year_no_approx_isi: smods_rec.pub_date_sortable_string(true),
+      pub_year_w_approx_isi: sort_str_w_approx_dates,
       # TODO:  remove pub_date after reindexing existing colls;  deprecated in favor of pub_year_xxx_approx_isi ...
       pub_date: smods_rec.pub_date_facet,
       # display fields
@@ -60,20 +61,7 @@ module GDor::Indexer::ModsFields
       all_search: smods_rec.text.gsub(/\s+/, ' ')
     }
 
-    # more pub date field processing
-    pub_date_sort_val = doc_hash[:pub_date_sort]
-    if is_positive_int? pub_date_sort_val
-      doc_hash[:pub_year_tisim] = pub_date_sort_val # for date slider
-      # remove leading zeros
-      doc_hash[:creation_year_isi] = remove_leading_zeros(doc_hash[:creation_year_isi]) if doc_hash[:creation_year_isi]
-      doc_hash[:publication_year_isi] = remove_leading_zeros(doc_hash[:publication_year_isi]) if doc_hash[:publication_year_isi]
-    else
-      # turn B.C. into -yyy for display fields
-      doc_hash[:creation_year_isi] = '-' + (1000 + doc_hash[:creation_year_isi].to_i).to_s if doc_hash[:creation_year_isi]
-      doc_hash[:publication_year_isi] = '-' + (1000 + doc_hash[:publication_year_isi].to_i).to_s if doc_hash[:publication_year_isi]
-    end
-
-    doc_hash
+    more_pub_date_goodness(doc_hash)
   end
 
   private
@@ -88,6 +76,34 @@ module GDor::Indexer::ModsFields
     vals
   end
 
+  # additional pub date field processing for Solr doc hash
+  # @param [Hash] Hash representing the Solr document
+  # @return [Hash] updated Hash representing the Solr document
+  def more_pub_date_goodness(doc_hash)
+    pub_date_sort_val = doc_hash[:pub_date_sort]
+    if is_positive_int? pub_date_sort_val
+      doc_hash[:pub_year_tisim] = pub_date_sort_val # for date slider
+      # remove leading zeros
+      [:pub_year_no_approx_isi,
+        :pub_year_w_approx_isi,
+        :creation_year_isi,
+        :publication_year_isi
+      ].each do |field_sym|
+        doc_hash[field_sym] = remove_leading_zeros(doc_hash[field_sym]) if doc_hash[field_sym]
+      end
+    else
+      # turn B.C. into -yyy for facet/display fields
+      [:pub_year_no_approx_isi,
+        :pub_year_w_approx_isi,
+        :creation_year_isi,
+        :publication_year_isi
+      ].each do |field_sym|
+        doc_hash[field_sym] = '-' + (1000 + doc_hash[field_sym].to_i).to_s if doc_hash[field_sym]
+      end
+    end
+    doc_hash
+  end
+
   # @return true if the string parses into an int, and if so, the int is >= 0
   def is_positive_int?(str)
     str.to_i >= 0

data/lib/gdor/indexer/version.rb

@@ -1,5 +1,5 @@
 module GDor
   class Indexer
-    VERSION = '0.4.0'
+    VERSION = '0.4.1'
   end
 end

data/spec/unit/indexer_spec.rb

@@ -120,14 +120,14 @@ describe GDor::Indexer do
       @indexer.index collection
     end
 
-    it 'indexs other resources as items' do
+    it 'indexes other resources as items' do
       expect(@indexer).to receive(:item_solr_document).with(resource)
       @indexer.index resource
     end
   end
 
   describe '#index_with_exception_handling' do
-    it 'capture,s log, and re-raise any exception thrown by the indexing process' do
+    it 'captures log and re-raises any exception thrown by the indexing process' do
       expect(@indexer).to receive(:index).with(resource).and_raise 'xyz'
       expect(@indexer.logger).to receive(:error)
       expect { @indexer.index_with_exception_handling(resource) }.to raise_error RuntimeError
@@ -433,7 +433,7 @@ describe GDor::Indexer do
   end
 
   # context "skip heartbeat" do
-  #   it "allows me to use a fake url for dor-fetcher-client" do
+  #   it "allows use of a fake url for dor-fetcher-client" do
   #     expect {GDor::Indexer.new(@config_yml_path)}.not_to raise_error
   #   end
   # end

data/spec/unit/mods_pub_fields_spec.rb

@@ -33,9 +33,51 @@ describe GDor::Indexer::ModsFields do
       end
     end
 
+    context 'pub_date (to know current behavior)' do
+      it 'calls Stanford::Mods::Record instance pub_date_sortable_string(false)' do
+        expect(sdb.smods_rec).to receive(:pub_date_facet)
+        sdb.doc_hash_from_mods[:pub_date]
+      end
+      it 'includes approx dates' do
+        m = mods_origin_info_start_str +
+              "<dateIssued qualifier='approximate'>1945</dateIssued>" +
+            mods_origin_info_end_str
+        sdb = sdb_for_mods(m)
+        expect(sdb.doc_hash_from_mods[:pub_date]).to eq('1945')
+      end
+      it 'takes single dateCreated' do
+        m = mods_origin_info_start_str +
+              "<dateCreated>1904</dateCreated>" +
+            mods_origin_info_end_str
+        sdb = sdb_for_mods(m)
+        expect(sdb.doc_hash_from_mods[:pub_date]).to eq('1904')
+      end
+      it_behaves_like 'expected', :pub_date, 'blah blah 1945 blah', '1945'
+      it_behaves_like 'expected', :pub_date, '1945', '1945'
+      it_behaves_like 'expected', :pub_date, '945', '945'
+      it_behaves_like 'expected', :pub_date, '66', nil
+      it_behaves_like 'expected', :pub_date, '5', nil
+      it_behaves_like 'expected', :pub_date, '0', nil
+      it_behaves_like 'expected', :pub_date, '-4', nil
+      it_behaves_like 'expected', :pub_date, '-15', nil
+      it_behaves_like 'expected', :pub_date, '-666', '666' # WRONG
+      it_behaves_like 'expected', :pub_date, '16--', nil
+      it_behaves_like 'expected', :pub_date, '8--', nil
+      it_behaves_like 'expected', :pub_date, '19th century', '19th century'
+      it_behaves_like 'expected', :pub_date, '9th century', '9th century'
+      it_behaves_like 'expected', :pub_date, '300 B.C.', '300 B.C.'
+      it_behaves_like 'expected', :pub_date, 'Text dated June 4, 1594; miniatures added by 1596', '1594'
+      it_behaves_like 'expected', :pub_date, 'Aug. 3rd, 1886', '1886'
+      it_behaves_like 'expected', :pub_date, 'Aug. 3rd, [18]86?', '1886'
+      it_behaves_like 'expected', :pub_date, 'early 1890s', '1890'
+      it_behaves_like 'expected', :pub_date, '1865-6', '1865'
+
+    end
+
     context 'pub_date_sort' do
       it 'calls Stanford::Mods::Record instance pub_date_sortable_string(false)' do
         expect(sdb.smods_rec).to receive(:pub_date_sortable_string).with(false)
+        allow(sdb.smods_rec).to receive(:pub_date_sortable_string).with(true) # for pub_year_no_approx_isi
         sdb.doc_hash_from_mods[:pub_date_sort]
       end
       it 'includes approx dates' do
@@ -78,30 +120,36 @@ describe GDor::Indexer::ModsFields do
       end
       it 'pub_year_no_approx_isi calls Stanford::Mods::Record instance pub_date_facet_single_value(true)' do
         sdb = sdb_for_mods(mods)
-        expect(sdb.smods_rec).to receive(:pub_date_facet_single_value).with(true).and_call_original
-        allow(sdb.smods_rec).to receive(:pub_date_facet_single_value).with(false) # for other flavor
+        expect(sdb.smods_rec).to receive(:pub_date_sortable_string).with(true).and_call_original
+        allow(sdb.smods_rec).to receive(:pub_date_sortable_string).with(false) # for other flavor
         expect(sdb.doc_hash_from_mods[:pub_year_no_approx_isi]).to eq '2000'
       end
       it 'pub_year_w_approx_isi calls Stanford::Mods::Record instance pub_date_facet_single_value(false)' do
         sdb = sdb_for_mods(mods)
-        expect(sdb.smods_rec).to receive(:pub_date_facet_single_value).with(false).and_call_original
-        allow(sdb.smods_rec).to receive(:pub_date_facet_single_value).with(true) # for other flavor
+        expect(sdb.smods_rec).to receive(:pub_date_sortable_string).with(false).and_call_original
+        allow(sdb.smods_rec).to receive(:pub_date_sortable_string).with(true) # for other flavor
         expect(sdb.doc_hash_from_mods[:pub_year_w_approx_isi]).to eq '1500'
       end
       RSpec.shared_examples "single pub year facet" do |field_sym|
+        it_behaves_like 'expected', field_sym, 'blah blah 1945 blah', '1945'
         it_behaves_like 'expected', field_sym, '1945', '1945'
         it_behaves_like 'expected', field_sym, '945', '945'
         it_behaves_like 'expected', field_sym, '66', '66'
         it_behaves_like 'expected', field_sym, '5', '5'
         it_behaves_like 'expected', field_sym, '0', '0'
-        it_behaves_like 'expected', field_sym, '-4', '4 B.C.'
-        it_behaves_like 'expected', field_sym, '-15', '15 B.C.'
-        it_behaves_like 'expected', field_sym, '-666', '666 B.C.'
-        it_behaves_like 'expected', field_sym, '16--', '17th century'
-        it_behaves_like 'expected', field_sym, '8--', '9th century'
-        it_behaves_like 'expected', field_sym, '19th century', '19th century'
-        it_behaves_like 'expected', field_sym, '9th century', '9th century'
-        it_behaves_like 'expected', field_sym, '300 B.C.', '300 B.C.'
+        it_behaves_like 'expected', field_sym, '-4', '-4'
+        it_behaves_like 'expected', field_sym, '-15', '-15'
+        it_behaves_like 'expected', field_sym, '-666', '-666'
+        it_behaves_like 'expected', field_sym, '16--', '1600'
+        it_behaves_like 'expected', field_sym, '8--', '800'
+        it_behaves_like 'expected', field_sym, '19th century', '1800'
+        it_behaves_like 'expected', field_sym, '9th century', '800'
+        it_behaves_like 'expected', field_sym, '300 B.C.', '-300'
+        it_behaves_like 'expected', field_sym, 'Text dated June 4, 1594; miniatures added by 1596', '1594'
+        it_behaves_like 'expected', field_sym, 'Aug. 3rd, 1886', '1886'
+        it_behaves_like 'expected', field_sym, 'Aug. 3rd, [18]86?', '1886'
+        it_behaves_like 'expected', field_sym, 'early 1890s', '1890'
+        it_behaves_like 'expected', field_sym, '1865-6', '1865'
       end
       it_behaves_like "single pub year facet", :pub_year_no_approx_isi
       it_behaves_like "single pub year facet", :pub_year_w_approx_isi
@@ -113,6 +161,7 @@ describe GDor::Indexer::ModsFields do
       # FIXME:  it should be using a method approp for date slider values, not single value
       it 'pub_year_tisim calls Stanford::Mods::Record instance pub_date_sortable_string(false)' do
         expect(sdb.smods_rec).to receive(:pub_date_sortable_string).with(false)
+        allow(sdb.smods_rec).to receive(:pub_date_sortable_string).with(true) # for pub_year_no_approx_isi
         sdb.doc_hash_from_mods[:pub_year_tisim]
       end
       it 'includes approx dates' do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gdor-indexer
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.4.1
 platform: ruby
 authors:
 - Naomi Dushay
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-01-06 00:00:00.000000000 Z
+date: 2016-01-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: harvestdor-indexer
@@ -264,34 +264,6 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.5'
-- !ruby/object:Gem::Dependency
-  name: capybara
-  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: poltergeist
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.5.0
-  type: :development
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: 1.5.0
 - !ruby/object:Gem::Dependency
   name: vcr
   requirement: !ruby/object:Gem::Requirement
@@ -306,20 +278,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
-- !ruby/object:Gem::Dependency
-  name: jettywrapper
-  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: webmock
   requirement: !ruby/object:Gem::Requirement
@@ -408,7 +366,7 @@ rubyforge_project:
 rubygems_version: 2.4.6
 signing_key: 
 specification_version: 4
-summary: Gryphondor Solr indexing logic
+summary: PURL doc =>  Solr hash logic
 test_files:
 - spec/config/walters_integration_spec.yml
 - spec/spec_helper.rb