pennmarc 1.0.1 → 1.0.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e574f1fae4f59435df3bf467da310ae7e426d82c2fe54ae2321ad326c05e5e99
-  data.tar.gz: de545e5a150b26c73e814e36b167bb688d267a97205cf4cdbab15dfd5c3225a8
+  metadata.gz: 33527bf43532170690b9b591097fb03f854130b3383debfafe931417d0394f52
+  data.tar.gz: a18a5be08df6d7b74d2aaa61275cdecdedec6caa4b5e8b2b81c99e91411ce093
 SHA512:
-  metadata.gz: 3c3fc981f5999eaca710a54fa3d247f2d01cc8183e13212151afab0cbb5b37f93531cbd264b22d1d4f30ef730b800efe38039b4bfc42a44cd28778c06f0cde3f
-  data.tar.gz: 729f0f80fd9d9ade7d1ae1085610cd85b8ab31beb1245efcaee246a18c01238abd5adc18dcf995f624448d518f4e36222a71c972d00c537c6367229f7d8d35c6
+  metadata.gz: 627a4da06351037f520bc02b0a9fd61ce6cfdec35c563f2b29e3c1c01b4ad76766f155630cafe680b010783ea4f4c285df2d499fd784b7ee4f3cb4948cd421dd
+  data.tar.gz: c966233bb00009a14babc5bc92f0399e75640e886f7c32d176fea8983813aca8936d0e8fab98576a797afaf0b60c5f24268c2ae4e2d8cca31900e590db0a9bc6

data/.rubocop.yml

@@ -2,3 +2,6 @@ inherit_from: .rubocop_todo.yml
 
 inherit_gem:
   upennlib-rubocop: upennlib_rubocop_defaults.yml
+
+Rails:
+  Enabled: false

data/.rubocop_todo.yml

@@ -1,11 +1,35 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config --auto-gen-only-exclude --exclude-limit 10000`
-# on 2023-08-14 17:52:48 UTC using RuboCop version 1.51.0.
+# on 2023-08-25 13:55:25 UTC using RuboCop version 1.51.0.
 # 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
 # versions of RuboCop, may require this file to be generated again.
 
+# Offense count: 2
+# This cop supports safe autocorrection (--autocorrect).
+# Configuration parameters: EnforcedStyle, IndentationWidth.
+# SupportedStyles: with_first_argument, with_fixed_indentation
+Layout/ArgumentAlignment:
+  Exclude:
+    - 'spec/lib/pennmarc/helpers/series_spec.rb'
+
+# Offense count: 1
+# This cop supports safe autocorrection (--autocorrect).
+# Configuration parameters: EnforcedStyle.
+# SupportedStyles: empty_lines, no_empty_lines
+Layout/EmptyLinesAroundBlockBody:
+  Exclude:
+    - 'spec/lib/pennmarc/helpers/identifer_spec.rb'
+
+# Offense count: 1
+# This cop supports safe autocorrection (--autocorrect).
+# Configuration parameters: EnforcedStyle.
+# SupportedStyles: final_newline, final_blank_line
+Layout/TrailingEmptyLines:
+  Exclude:
+    - 'spec/lib/pennmarc/helpers/identifer_spec.rb'
+
 # Offense count: 22
 # Configuration parameters: AllowedMethods, AllowedPatterns, CountRepeatedAttributes, Max.
 Metrics/AbcSize:
@@ -86,7 +110,7 @@ Metrics/PerceivedComplexity:
     - 'lib/pennmarc/helpers/title.rb'
     - 'lib/pennmarc/util.rb'
 
-# Offense count: 2
+# Offense count: 4
 # This cop supports safe autocorrection (--autocorrect).
 # Configuration parameters: EnforcedStyle, BlockForwardingName.
 # SupportedStyles: anonymous, explicit
@@ -104,7 +128,7 @@ Naming/PredicateName:
   Exclude:
     - 'lib/pennmarc/helpers/relation.rb'
 
-# Offense count: 2
+# Offense count: 1
 # Configuration parameters: EnforcedStyle, CheckMethodNames, CheckSymbols, AllowedIdentifiers, AllowedPatterns.
 # SupportedStyles: snake_case, normalcase, non_integer
 # AllowedIdentifiers: capture3, iso8601, rfc1123_date, rfc822, rfc2822, rfc3339, x86_64
@@ -135,9 +159,9 @@ RSpec/NestedGroups:
     - 'spec/lib/pennmarc/helpers/format_spec.rb'
 
 # Offense count: 2
-# This cop supports unsafe autocorrection (--autocorrect-all).
-# Configuration parameters: Include.
-# Include: app/**/*.rb, config/**/*.rb, db/**/*.rb, lib/**/*.rb
-Rails/Output:
+# This cop supports safe autocorrection (--autocorrect).
+# Configuration parameters: Max, AllowHeredoc, AllowURI, URISchemes, IgnoreCopDirectives, AllowedPatterns.
+# URISchemes: http, https
+Layout/LineLength:
   Exclude:
-    - 'lib/pennmarc/helpers/date.rb'
+    - 'spec/lib/pennmarc/helpers/creator_spec.rb'

data/Gemfile.lock

@@ -21,6 +21,8 @@ GEM
     minitest (5.18.0)
     nokogiri (1.15.2-arm64-darwin)
       racc (~> 1.4)
+    nokogiri (1.15.2-x64-mingw-ucrt)
+      racc (~> 1.4)
     nokogiri (1.15.2-x86_64-linux)
       racc (~> 1.4)
     parallel (1.23.0)
@@ -101,6 +103,7 @@ GEM
 PLATFORMS
   arm64-darwin-21
   arm64-darwin-22
+  x64-mingw-ucrt
   x86_64-linux
 
 DEPENDENCIES

data/lib/pennmarc/helpers/identifier.rb

@@ -33,14 +33,12 @@ module PennMARC
       #
       # @param [MARC::Record] record
       # @return [Array<String>]
-      # @todo look into z subfield for 020 field, should we show cancelled isbn?
       def isbn_show(record)
         isbn_values = record.fields('020').filter_map do |field|
-          joined_isbn = join_subfields(field, &subfield_in?(%w[a z]))
+          joined_isbn = join_subfields(field, &subfield_in?(%w[a]))
           joined_isbn.presence
         end
-        isbn_values += linked_alternate(record, '020', &subfield_in?(%w[a z]))
-        isbn_values
+        isbn_values + linked_alternate(record, '020', &subfield_in?(%w[a]))
       end
 
       # Get ISSN values for display from the {https://www.oclc.org/bibformats/en/0xx/022.html 022 field} and related
@@ -50,11 +48,10 @@ module PennMARC
       # @return [Array<String>]
       def issn_show(record)
         issn_values = record.fields('022').filter_map do |field|
-          joined_issn = join_subfields(field, &subfield_in?(%w[a z]))
+          joined_issn = join_subfields(field, &subfield_in?(%w[a]))
           joined_issn.presence
         end
-        issn_values += linked_alternate(record, '022', &subfield_in?(%w[a z]))
-        issn_values
+        issn_values + linked_alternate(record, '022', &subfield_in?(%w[a]))
       end
 
       # Get numeric OCLC ID of first {https://www.oclc.org/bibformats/en/0xx/035.html 035 field}
@@ -85,17 +82,22 @@ module PennMARC
 
       # Get publisher issued identifiers from fields {https://www.oclc.org/bibformats/en/0xx/024.html 024},
       # {https://www.oclc.org/bibformats/en/0xx/024.html 028}, and related
-      # {https://www.oclc.org/bibformats/en/8xx/880.html 880 field}.
+      # {https://www.oclc.org/bibformats/en/8xx/880.html 880 field}. We do not return DOI values stored in 024 ǂ2,
+      # see {PennMARC::Identifier.doi_show} for parsing DOI values.
       #
       # @param [MARC::Record] record
       # @return [Array<string>]
       def publisher_number_show(record)
-        publisher_numbers = record.fields(%w[024 028]).filter_map do |field|
-          joined_identifiers = join_subfields(field, &subfield_not_in?(%w[5 6]))
-          joined_identifiers.presence
+        record.fields(%w[024 028 880]).filter_map do |field|
+          next if field.tag == '880' && subfield_value_not_in?(field, '6', %w[024 028])
+
+          # do not return doi values from 024 ǂ2
+          if field.tag == '024' && subfield_value_in?(field, '2', %w[doi])
+            join_subfields(field, &subfield_not_in?(%w[a 2 5 6])).presence
+          else
+            join_subfields(field, &subfield_not_in?(%w[5 6])).presence
+          end
         end
-        publisher_numbers += linked_alternate(record, %w[024 028], &subfield_not_in?(%w[5 6]))
-        publisher_numbers
       end
 
       # Get publisher issued identifiers for searching of a record. Values extracted from fields
@@ -119,6 +121,21 @@ module PennMARC
         end
       end
 
+      # Retrieve DOI values stored in {https://www.oclc.org/bibformats/en/0xx/024.html 024}.
+      # Penn MARC records give the first indicator a value of '7' and ǂ2 a value of 'doi' to denote that ǂa is a doi.
+      # @param [MARC::Record] record
+      # @return [Array<String>]
+      def doi_show(record)
+        record.fields('024').filter_map do |field|
+          # skip unless indicator1 is '7'
+          next unless field.indicator1.in?(%w[7])
+          # skip unless ǂ2 is the string literal 'doi'
+          next unless subfield_value_in?(field, '2', %w[doi])
+
+          join_subfields(field, &subfield_in?(%w[a]))
+        end
+      end
+
       private
 
       # Determine if subfield 'a' is an OCLC id.
@@ -126,11 +143,11 @@ module PennMARC
       # @param [MARC::Subfield]
       # @return [TrueClass, FalseClass]
       def subfield_a_is_oclc?(subfield)
-        subfield.code == 'a' && subfield.value =~ /^\(OCoLC\).*/
+        subfield.code == 'a' && (subfield.value =~ /^\(OCoLC\).*/).present?
       end
 
       # Normalize isbn value using {https://github.com/billdueber/library_stdnums library_stdnums gem}.
-      # Converts ISBN10 (ten-digit) to validated ISBN13 (thriteen-digit) and returns both values. If passed
+      # Converts ISBN10 (ten-digit) to validated ISBN13 (thirteen-digit) and returns both values. If passed
       # ISBN13 parameter, only returns validated ISBN13 value.
       #
       #  @param [String] isbn

data/lib/pennmarc/helpers/note.rb

@@ -11,8 +11,6 @@ module PennMARC
       # {https://www.oclc.org/bibformats/en/5xx/550.html 550}, {https://www.oclc.org/bibformats/en/5xx/580.html 580},
       # {https://www.oclc.org/bibformats/en/5xx/586.html 586}, {https://www.oclc.org/bibformats/en/5xx/588.html 588},
       # and their linked alternates.
-      # @todo legacy implementation used conditional to separate join logic for 588 field. However, this doesn't seem
-      #   necessary because 588 only has subfields 'a', '5', '6', and '8'. Do we need to look into this?
       # @param [MARC::Record] record
       # @return [Array<String>]
       def notes_show(record)

data/lib/pennmarc/helpers/series.rb

@@ -4,7 +4,6 @@ module PennMARC
   # Do Series and series-related field processing. Many of these fields are added entries that are justified by
   # corresponding series statements (usually 490). These fields provide information about the published series in which
   # a book, encoded finding aid, or other published work has appeared
-  # @todo We may want to include 410 in the display tags, since it is included in references below.
   class Series < Helper
     class << self
       # 800 - Series Added Entry-Personal Name - https://www.loc.gov/marc/bibliographic/bd800.html
@@ -16,7 +15,7 @@ module PennMARC
       # 411 - Series Statement/Added Entry Meeting Name - https://www.loc.gov/marc/bibliographic/bd411.html
       # 440 - Series Statement/Added Entry-Title - https://www.loc.gov/marc/bibliographic/bd440.html
       # 490 - Series Statement - https://www.loc.gov/marc/bibliographic/bd490.html
-      DISPLAY_TAGS = %w[800 810 811 830 400 411 440 490].freeze
+      DISPLAY_TAGS = %w[800 810 811 830 400 410 411 440 490].freeze
 
       # Fields for display that pertain to series information.
       # @param [MARC::Record] record

data/lib/pennmarc/helpers/subject.rb

@@ -168,7 +168,8 @@ module PennMARC
       def format_term(type:, term:)
         return unless type.in? %i[facet display]
 
-        normalize_single_subfield(term[:parts].first) if term[:count] == 1
+        # attempt to handle poorly coded record
+        normalize_single_subfield(term[:parts].first) if term[:count] == 1 && term[:parts].first.present?
 
         case type.to_sym
         when :facet

data/lib/pennmarc/helpers/title.rb

@@ -89,7 +89,7 @@ module PennMARC
                  end
         raw_title = join_subfields(title_field, &subfield_in?(['a'])) # get title from subfield a
         value = if offset.between?(1, 9)
-                  { prefix: raw_title[0..offset - 1].strip, filing: raw_title[offset..].strip }
+                  { prefix: raw_title[0..offset - 1]&.strip, filing: raw_title[offset..]&.strip }
                 elsif raw_title.present?
                   handle_bracket_prefix raw_title
                 else

data/lib/pennmarc/version.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+
+module PennMARC
+  VERSION = '1.0.2'
+end

data/pennmarc.gemspec

@@ -1,8 +1,12 @@
 # frozen_string_literal: true
 
+lib = File.expand_path('lib', __dir__)
+$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
+require 'pennmarc/version'
+
 Gem::Specification.new do |s|
   s.name        = 'pennmarc'
-  s.version     = '1.0.1'
+  s.version     = PennMARC::VERSION
   s.summary     = 'Penn Libraries Catalog MARC parsing wisdom for cross-project usage'
   s.description = 'This gem provides methods for parsing a Penn Libraries MARCXML record into string, array and date
                    objects for use in discovery or preservation applications.'

data/spec/lib/pennmarc/helpers/identifer_spec.rb

@@ -16,13 +16,14 @@ describe 'PennMARC::Identifier' do
   describe '.isxn_search' do
     let(:record) do
       marc_record fields: [
-        marc_field(tag: '020', subfields: { a: '9781594205071', z: '1594205078' }),
-        marc_field(tag: '022', subfields: { a: '0008-6533', l: '0300-7162', z: '0008-6533' })
+        marc_field(tag: '020', subfields: { a: '9781594205071', z: '1555975275' }),
+        marc_field(tag: '022', subfields: { a: '0008-6533', l: '0300-7162', z: '0799-5946 ' })
       ]
     end
 
     it 'returns expected search values' do
-      expect(helper.isxn_search(record)).to contain_exactly('9781594205071', '1594205078', '0300-7162', '0008-6533')
+      expect(helper.isxn_search(record)).to contain_exactly('9781594205071', '1555975275', '9781555975272',
+                                                            '1594205078', '0300-7162', '0008-6533', '0799-5946 ')
     end
 
     it 'converts ISBN10 values to ISBN13' do
@@ -40,9 +41,9 @@ describe 'PennMARC::Identifier' do
       ]
     end
 
-    it 'returns expected show values' do
-      expect(helper.isbn_show(record)).to contain_exactly('9781594205071 1594205078', '0805073698 9780735222786',
-                                                          '0735222789 9780805073690')
+    it 'returns expected ǂa values' do
+      expect(helper.isbn_show(record)).to contain_exactly('9781594205071', '0805073698',
+                                                          '0735222789')
     end
   end
 
@@ -55,9 +56,9 @@ describe 'PennMARC::Identifier' do
       ]
     end
 
-    it 'returns expected show values' do
-      expect(helper.issn_show(record)).to contain_exactly('0008-6533 0008-6533', '2470-6302 1534-6714',
-                                                          '1080-6512 2213-4360')
+    it 'returns ǂa values' do
+      expect(helper.issn_show(record)).to contain_exactly('0008-6533', '2470-6302',
+                                                          '1080-6512')
     end
   end
 
@@ -79,9 +80,11 @@ describe 'PennMARC::Identifier' do
     let(:record) do
       marc_record fields: [
         marc_field(tag: '024', subfields: { a: '602537854325', '6': '880-01' }),
+        marc_field(tag: '024', subfields: { a: '10.18574/9781479842865', '2': 'doi' }),
         marc_field(tag: '028', subfields: { a: 'B002086600', b: 'Island Def Jam Music Group', '6': '880-01' }),
         marc_field(tag: '880', subfields: { a: '523458735206', '6': '024' }),
-        marc_field(tag: '880', subfields: { a: '006680200B', b: 'Island', '6': '028' })
+        marc_field(tag: '880', subfields: { a: '006680200B', b: 'Island', '6': '028' }),
+        marc_field(tag: '880', subfields: { a: '006680200B', b: 'Island', '6': '021' })
       ]
     end
 
@@ -90,18 +93,26 @@ describe 'PennMARC::Identifier' do
                                                                       'B002086600 Island Def Jam Music Group',
                                                                       '523458735206', '006680200B Island')
     end
+
+    it 'does not return DOI values' do
+      expect(helper.publisher_number_show(record)).not_to include('10.18574/9781479842865')
+      expect(helper.publisher_number_show(record)).not_to include('doi')
+
+    end
   end
 
   describe '.publisher_number_search' do
     let(:record) do
       marc_record fields: [
-        marc_field(tag: '024', subfields: { a: '602537854325' }),
+        marc_field(tag: '024', subfields: { a: '602537854325', b: 'exclude' }),
+        marc_field(tag: '024', subfields: { a: '10.18574/9781479842865', '2': 'doi' }),
         marc_field(tag: '028', subfields: { a: 'B002086600', b: 'Island Def Jam Music Group' })
       ]
     end
 
-    it 'returns expected search values' do
-      expect(helper.publisher_number_search(record)).to contain_exactly('602537854325', 'B002086600')
+    it 'returns publisher numbers from 024/028 ǂa and DOI values in 024 ǂ2' do
+      expect(helper.publisher_number_search(record)).to contain_exactly('10.18574/9781479842865', '602537854325',
+                                                                        'B002086600')
     end
   end
 
@@ -117,4 +128,22 @@ describe 'PennMARC::Identifier' do
       expect(helper.fingerprint_show(record)).to contain_exactly('dete nkck vess lodo Anno Domini MDCXXXVI 3')
     end
   end
-end
+
+  describe '.doi_show' do
+    let(:record) do
+      marc_record fields: [
+        marc_field(tag: '024', indicator1: '7', subfields: { a: '10.1038/sdata.2016.18 ', '2': 'doi' }),
+        marc_field(tag: '024', indicator1: '7', subfields: { a: '10.18574/9781479842865', '2': 'doi' }),
+        marc_field(tag: '024', indicator1: '7',
+                   subfields: { a: '10.1016.12.31/nature.S0735-1097(98)2000/12?/31/34:7-7', '2': 'doi' }),
+        marc_field(tag: '024', indicator1: '7', subfields: { a: 'excluded', '2': 'non doi' }),
+        marc_field(tag: '024', indicator1: '0', subfields: { a: 'excluded', '2': 'doi' })
+      ]
+    end
+
+    it 'returns valid DOI values' do
+      expect(helper.doi_show(record)).to contain_exactly('10.1016.12.31/nature.S0735-1097(98)2000/12?/31/34:7-7',
+                                                         '10.1038/sdata.2016.18', '10.18574/9781479842865')
+    end
+  end
+end

data/spec/lib/pennmarc/helpers/series_spec.rb

@@ -6,7 +6,8 @@ describe 'PennMARC::Series' do
   let(:helper) { PennMARC::Series }
   let(:mapping) { { aut: 'Author' } }
   let(:record) do
-    marc_record fields: [marc_field(tag: '490', subfields: { a: 'Teachings of the feathered pillow' }),
+    marc_record fields: [marc_field(tag: '410', subfields: { a: 'Evil Giant Megacorp' }),
+                         marc_field(tag: '490', subfields: { a: 'Teachings of the feathered pillow' }),
                          marc_field(tag: '880', subfields: { '6': '490', a: 'Учения пернатой подушки' }),
                          marc_field(tag: '800', subfields: { a: 'Bean Bagatolvski', d: '1997-', v: 'bk. 1' }),
                          marc_field(tag: '780', subfields: { a: 'National Comfort Association' }),
@@ -20,7 +21,7 @@ describe 'PennMARC::Series' do
     it 'returns the series' do
       expect(helper.show(record, relator_map: mapping)).to contain_exactly('Bean Bagatolvski 1997- bk. 1',
                                                               'Teachings of the feathered pillow',
-                                                              'Учения пернатой подушки')
+                                                              'Учения пернатой подушки', 'Evil Giant Megacorp')
     end
   end
 
@@ -32,7 +33,7 @@ describe 'PennMARC::Series' do
 
   describe '.search' do
     it 'returns the search values' do
-      expect(helper.search(record)).to contain_exactly('Bean Bagatolvski 1997- bk. 1')
+      expect(helper.search(record)).to contain_exactly('Bean Bagatolvski 1997- bk. 1', 'Evil Giant Megacorp')
     end
 
     it 'returns an empty array if no values are found' do

metadata

@@ -1,16 +1,16 @@
 --- !ruby/object:Gem::Specification
 name: pennmarc
 version: !ruby/object:Gem::Version
-  version: 1.0.1
+  version: 1.0.2
 platform: ruby
 authors:
 - Mike Kanning
 - Amrey Mathurin
 - Patrick Perkins
-autorequire:
+autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-08-15 00:00:00.000000000 Z
+date: 2023-10-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
@@ -112,6 +112,7 @@ files:
 - lib/pennmarc/mappings/relator.yml
 - lib/pennmarc/parser.rb
 - lib/pennmarc/util.rb
+- lib/pennmarc/version.rb
 - pennmarc.gemspec
 - spec/fixtures/marcxml/test.xml
 - spec/lib/pennmarc/helpers/citation_spec.rb
@@ -139,7 +140,7 @@ licenses:
 - MIT
 metadata:
   rubygems_mfa_required: 'true'
-post_install_message:
+post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
@@ -155,7 +156,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubygems_version: 3.4.10
-signing_key:
+signing_key: 
 specification_version: 4
 summary: Penn Libraries Catalog MARC parsing wisdom for cross-project usage
 test_files: []