pennmarc 0.0.1

data/spec/fixtures/marcxml/test.xml

@@ -0,0 +1,167 @@
+<record>
+  <leader>01908cam a2200469 a 4500</leader>
+  <controlfield tag="005">20220609191957.0</controlfield>
+  <controlfield tag="008">890224t19891989nyua          001 0 eng  </controlfield>
+  <controlfield tag="001">9910148543503681</controlfield>
+  <datafield tag="010" ind1=" " ind2=" ">
+    <subfield code="a">   88029825 </subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(OCoLC)ocm18521155</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(OCoLC)18521155</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(CStRLIN)PAUG89-B7246</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(CaOTULAS)185188489</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="9">AHA6856</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">1014854</subfield>
+  </datafield>
+  <datafield tag="035" ind1=" " ind2=" ">
+    <subfield code="a">(PU)1014854-penndb-Voyager</subfield>
+  </datafield>
+  <datafield tag="040" ind1=" " ind2=" ">
+    <subfield code="b">eng</subfield>
+    <subfield code="d">CSt-B</subfield>
+  </datafield>
+  <datafield tag="043" ind1=" " ind2=" ">
+    <subfield code="a">n-us---</subfield>
+  </datafield>
+  <datafield tag="050" ind1="0" ind2=" ">
+    <subfield code="a">KF6450</subfield>
+    <subfield code="b">.C59 1989</subfield>
+  </datafield>
+  <datafield tag="082" ind1="0" ind2=" ">
+    <subfield code="a">343.7306/8</subfield>
+    <subfield code="a">347.30368</subfield>
+    <subfield code="2">19</subfield>
+  </datafield>
+  <datafield tag="090" ind1=" " ind2=" ">
+    <subfield code="a">KF6450</subfield>
+    <subfield code="b">.C59 1989</subfield>
+    <subfield code="i">11/30/89 CTZ</subfield>
+  </datafield>
+  <datafield tag="245" ind1="0" ind2="4">
+    <subfield code="a">The Coopers &amp; Lybrand guide to business tax strategies and planning /</subfield>
+    <subfield code="c">by the partners of Coopers &amp; Lybrand.</subfield>
+  </datafield>
+  <datafield tag="264" ind1=" " ind2="1">
+    <subfield code="a">New York :</subfield>
+    <subfield code="b">Simon and Schuster,</subfield>
+    <subfield code="c">[1989]</subfield>
+  </datafield>
+  <datafield tag="264" ind1=" " ind2="4">
+    <subfield code="c">©1989</subfield>
+  </datafield>
+  <datafield tag="300" ind1=" " ind2=" ">
+    <subfield code="a">x, 198 pages :</subfield>
+    <subfield code="b">illustrations ;</subfield>
+    <subfield code="c">24 cm</subfield>
+  </datafield>
+  <datafield tag="336" ind1=" " ind2=" ">
+    <subfield code="a">text</subfield>
+    <subfield code="b">txt</subfield>
+    <subfield code="2">rdacontent</subfield>
+  </datafield>
+  <datafield tag="337" ind1=" " ind2=" ">
+    <subfield code="a">unmediated</subfield>
+    <subfield code="b">n</subfield>
+    <subfield code="2">rdamedia</subfield>
+  </datafield>
+  <datafield tag="338" ind1=" " ind2=" ">
+    <subfield code="a">volume</subfield>
+    <subfield code="b">nc</subfield>
+    <subfield code="2">rdacarrier</subfield>
+  </datafield>
+  <datafield tag="500" ind1=" " ind2=" ">
+    <subfield code="a">Edited by Jonathan J. Davies and others.</subfield>
+  </datafield>
+  <datafield tag="650" ind1=" " ind2="0">
+    <subfield code="a">Business enterprises</subfield>
+    <subfield code="x">Taxation</subfield>
+    <subfield code="z">United States.</subfield>
+    <subfield code="0">http://id.loc.gov/authorities/subjects/sh2009118044</subfield>
+  </datafield>
+  <datafield tag="650" ind1=" " ind2="7">
+    <subfield code="a">Business enterprises</subfield>
+    <subfield code="x">Taxation.</subfield>
+    <subfield code="2">fast</subfield>
+    <subfield code="0">http://id.worldcat.org/fast/842634</subfield>
+  </datafield>
+  <datafield tag="651" ind1=" " ind2="7">
+    <subfield code="a">United States.</subfield>
+    <subfield code="2">fast</subfield>
+    <subfield code="0">http://id.worldcat.org/fast/1204155</subfield>
+  </datafield>
+  <datafield tag="650" ind1=" " ind2="0">
+    <subfield code="a">Tax planning</subfield>
+    <subfield code="z">United States.</subfield>
+    <subfield code="0">http://id.loc.gov/authorities/subjects/sh2008112546</subfield>
+  </datafield>
+  <datafield tag="650" ind1=" " ind2="7">
+    <subfield code="a">Tax planning.</subfield>
+    <subfield code="2">fast</subfield>
+    <subfield code="0">http://id.worldcat.org/fast/1143815</subfield>
+  </datafield>
+  <datafield tag="700" ind1="1" ind2=" ">
+    <subfield code="a">Davies, Jonathan J.</subfield>
+    <subfield code="0">http://id.loc.gov/authorities/names/n82126416</subfield>
+  </datafield>
+  <datafield tag="710" ind1="2" ind2=" ">
+    <subfield code="a">Coopers &amp; Lybrand.</subfield>
+    <subfield code="0">http://id.loc.gov/authorities/names/n79063025</subfield>
+  </datafield>
+  <datafield tag="740" ind1="0" ind2=" ">
+    <subfield code="a">Guide to business tax strategies and planning.</subfield>
+  </datafield>
+  <datafield tag="740" ind1="0" ind2=" ">
+    <subfield code="a">Business tax strategies and planning</subfield>
+  </datafield>
+  <datafield tag="740" ind1="0" ind2=" ">
+    <subfield code="a">Tax strategies and planning.</subfield>
+  </datafield>
+  <datafield tag="902" ind1=" " ind2=" ">
+    <subfield code="a">MARCIVE 2022</subfield>
+  </datafield>
+  <datafield tag="950" ind1=" " ind2=" ">
+    <subfield code="l">LIPP</subfield>
+    <subfield code="i">11/30/89 C</subfield>
+  </datafield>
+  <datafield tag="955" ind1=" " ind2=" ">
+    <subfield code="l">LIPP</subfield>
+    <subfield code="c">1</subfield>
+    <subfield code="q">89-B7246-1</subfield>
+    <subfield code="r">[01043 6272]</subfield>
+    <subfield code="i">11/30/89 C</subfield>
+  </datafield>
+  <datafield tag="hld" ind1="0" ind2="0">
+    <subfield code="b">Libra</subfield>
+    <subfield code="c">stor</subfield>
+    <subfield code="h">KF6450</subfield>
+    <subfield code="i">.C59 1989</subfield>
+    <subfield code="x">transferred from lipp</subfield>
+    <subfield code="8">22418068440003681</subfield>
+  </datafield>
+  <datafield tag="itm" ind1=" " ind2=" ">
+    <subfield code="r">22418068440003681</subfield>
+    <subfield code="b">1</subfield>
+    <subfield code="h">0</subfield>
+    <subfield code="g">stor</subfield>
+    <subfield code="9">31198058025532</subfield>
+    <subfield code="s">R27M22S09T04</subfield>
+    <subfield code="e">stor</subfield>
+    <subfield code="8">23418068430003681</subfield>
+    <subfield code="a">book/seria</subfield>
+    <subfield code="q">2017-06-06 09:53:20 US/Eastern</subfield>
+    <subfield code="i">KF6450 .C59 1989</subfield>
+    <subfield code="d">Libra</subfield>
+    <subfield code="f">Libra</subfield>
+  </datafield>
+</record>

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

@@ -0,0 +1,27 @@
+# frozen_string_literal: true
+
+describe 'PennMARC::Citation' do
+  include MarcSpecHelpers
+
+  let(:helper) { PennMARC::Citation }
+
+  describe '.cited_in_show' do
+    let(:record) do
+      marc_record fields: [marc_field(tag: '510', subfields: { a: 'Perkins Archive' }),
+                           marc_field(tag: '880', subfields: { '6': '510', a: 'パーキンスのアーカイブ' })]
+    end
+
+    it 'returns expected citation values' do
+      expect(helper.cited_in_show(record)).to contain_exactly('Perkins Archive', 'パーキンスのアーカイブ')
+    end
+  end
+
+  describe '.cite_as_show' do
+    let(:record) { marc_record fields: [marc_field(tag: '524', subfields: {a: 'Perkins Historical Archive, Box 2'})] }
+
+    it 'returns expected citation values' do
+      expect(helper.cite_as_show(record)).to contain_exactly('Perkins Historical Archive, Box 2')
+    end
+  end
+end
+

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

@@ -0,0 +1,183 @@
+# frozen_string_literal: true
+
+describe 'PennMARC::Creator' do
+  include MarcSpecHelpers
+
+  let(:helper) { PennMARC::Creator }
+  let(:mapping) { { aut: 'Author' } }
+
+  describe '.search' do
+    let(:record) { marc_record fields: fields }
+
+    context 'for a single author record' do
+      let(:fields) do
+        [marc_field(tag: '100', subfields: { a: 'Surname, Name', '0': 'http://cool.uri/12345',
+                                             e: 'author', d: '1900-2000' }),
+         marc_field(tag: '880', subfields: { a: 'Surname, Alternative', '6': '100' })]
+      end
+
+      it 'contains the expected search field values for a single author work' do
+        expect(helper.search(record, mapping)).to eq ['Name Surname http://cool.uri/12345 author 1900-2000.',
+                                                      'Surname, Name http://cool.uri/12345 author 1900-2000.',
+                                                      'Alternative Surname']
+      end
+    end
+
+    context 'for a corporate author record' do
+      let(:fields) do
+        [marc_field(tag: '110', subfields: { a: 'Group of People', b: 'Annual Meeting', '4': 'aut' }),
+         marc_field(tag: '880', subfields: { '6': '110', a: 'Alt. Group Name', b: 'Alt. Annual Meeting' })]
+      end
+
+      it 'contains the expected search field values for a corporate author work' do
+        expect(helper.search(record, mapping)).to eq ['Group of People Annual Meeting Author.',
+                                                      'Alt. Group Name Alt. Annual Meeting']
+      end
+    end
+  end
+
+  describe '.values' do
+    let(:record) { marc_record fields: fields }
+
+    context 'for a single author record' do
+      let(:fields) do
+        [marc_field(tag: '100', subfields: { a: 'Author', c: 'Fancy', d: 'active 24th century AD', '4': 'aut' }),
+         marc_field(tag: '880', subfields: { '6': '100', a: 'Alt Author', c: 'Alt Fanciness' })]
+      end
+
+      it 'returns values for the author, including mapped relator code from ǂ4' do
+        values = helper.values(record, mapping)
+        expect(values).to contain_exactly 'Author Fancy active 24th century AD, Author.'
+        expect(values.join.downcase).not_to include 'alt'
+      end
+    end
+
+    context 'for a corporate author record' do
+      let(:fields) do
+        [marc_field(tag: '110', subfields: { a: 'Annual Report', b: 'Leader', e: 'author', '4': 'aut' })]
+      end
+
+      it 'returns values for the corporate author, including mapped relator code from ǂ4' do
+        expect(helper.values(record, mapping)).to contain_exactly 'Annual Report Leader author, Author.'
+      end
+    end
+  end
+
+  describe '.show' do
+    let(:record) { marc_record fields: fields }
+
+    context 'for a single author record' do
+      let(:fields) do
+        [marc_field(tag: '100', subfields: { a: 'Surname, Name', '0': 'http://cool.uri/12345', d: '1900-2000',
+                                             e: 'author', '4': 'http://cool.uri/vocabulary/relators/aut' }),
+         marc_field(tag: '880', subfields: { a: 'Surname, Alternative', '6': '100' })]
+      end
+      it 'returns single author values with no URIs anywhere' do
+        values = helper.show(record)
+        expect(values).to contain_exactly 'Surname, Name 1900-2000', 'Surname, Alternative'
+        expect(values.join.downcase).not_to include 'http'
+      end
+    end
+
+    context 'for a corporate author record' do
+      let(:fields) do
+        [marc_field(tag: '110', subfields: { a: 'Group of People', b: 'Annual Meeting', '4': 'aut' }),
+         marc_field(tag: '880', subfields: { '6': '110', a: 'Alt. Group Name', b: 'Alt. Annual Meeting' })]
+      end
+      it 'returns corporate author values with no URIs anywhere' do
+        values = helper.show(record)
+        expect(values).to contain_exactly 'Alt. Group Name Alt. Annual Meeting', 'Group of People Annual Meeting'
+        expect(values.join.downcase).not_to include 'http'
+      end
+    end
+  end
+
+  describe '.sort' do
+    let(:record) do
+      marc_record fields: [
+        marc_field(tag: '100', subfields: { a: 'Sort, Value,', d: '1900-2000', e: 'Composer',
+                                            '0': 'http://cool.uri/12345', '4': 'aut' })
+      ]
+    end
+
+    it 'returns single value with no content from ǂ1, ǂ4, ǂ6, ǂ8 or ǂe' do
+      expect(helper.sort(record)).to eq 'Sort, Value, 1900-2000 http://cool.uri/12345'
+      expect(helper.sort(record)).not_to include 'aut'
+    end
+  end
+
+  describe '.facet' do
+    let(:record) { marc_record fields: fields }
+    let(:values) { helper.facet(record) }
+
+    context 'for a single author record' do
+      let(:fields) do
+        [marc_field(tag: '100', subfields: { a: 'Author, Great', d: '1900-2000' }),
+         marc_field(tag: '700', subfields: { a: 'Co-Author, Great', d: '1900-2000' }),
+         marc_field(tag: '800', subfields: { a: 'Author, Added' })]
+      end
+
+      it 'includes corporate author and creator values from allowed subfields' do
+        expect(values).to contain_exactly 'Author, Added', 'Author, Great 1900-2000', 'Co-Author, Great 1900-2000'
+      end
+    end
+    context 'for a corporate author record' do
+      let(:fields) do
+        [marc_field(tag: '110', subfields: { a: 'Group of People', b: 'Annual Meeting' }),
+         marc_field(tag: '710', subfields: { a: 'A Publisher', e: 'publisher' }),
+         marc_field(tag: '710', subfields: { a: 'A Sponsor', e: 'sponsoring body' }),
+         marc_field(tag: '810', subfields: { a: 'A Series Entity', t: 'Some Series' })]
+      end
+
+      it 'includes corporate author and creator values from allowed subfields' do
+        expect(values).to contain_exactly 'A Publisher', 'A Series Entity', 'A Sponsor',
+                                          'Group of People Annual Meeting'
+      end
+    end
+    context 'for a meeting author record' do
+      let(:fields) do
+        [marc_field(tag: '111', subfields: { a: 'Conference on Things', c: 'Earth' }),
+         marc_field(tag: '711', subfields: { a: 'Thing Institute', j: 'sponsoring body' }),
+         marc_field(tag: '811', subfields: { a: 'Series of Things', c: 'Earth' })]
+      end
+
+      it 'includes corporate author and creator values from allowed subfields' do
+        expect(values).to contain_exactly 'Conference on Things Earth', 'Series of Things Earth', 'Thing Institute'
+      end
+    end
+  end
+
+  describe '.conference_show' do
+    let(:record) do
+      marc_record fields: [
+        marc_field(tag: '111', subfields: { a: 'MARC History Symposium', '4': 'aut' }),
+        marc_field(tag: '880', subfields: { a: 'Alt. MARC History Symposium', '6': '111' })
+      ]
+    end
+
+    it 'returns conference name information for display, ignoring any linked 880 fields' do
+      expect(helper.conference_show(record, mapping)).to eq ['MARC History Symposium, Author.']
+    end
+  end
+
+  describe '.conference_detail_show' do
+    let(:record) do
+      marc_record fields: [
+        marc_field(tag: '111', subfields: { a: 'MARC History Symposium', c: 'Moscow' }),
+        marc_field(tag: '711', subfields: { a: 'Russian Library Conference', j: 'author' }),
+        marc_field(tag: '711', indicator2: '1', subfields: { a: 'Ignored Entry', j: 'author' }),
+        marc_field(tag: '880', subfields: { a: 'Proceedings', '6': '111' }),
+        marc_field(tag: '880', subfields: { a: 'Not Included', i: 'something', '6': '111' })
+      ]
+    end
+
+    it 'returns detailed conference name information for display, including linked 880 fields without ǂi, and ignoring
+        any 111 or 711 with a defined indicator 2 value' do
+      expect(helper.conference_detail_show(record)).to eq ['MARC History Symposium Moscow',
+                                                           'Russian Library Conference author', 'Proceedings']
+    end
+  end
+
+  xdescribe '.conference_search'
+  xdescribe '.search_aux'
+end

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

@@ -0,0 +1,60 @@
+# frozen_string_literal: true
+
+describe 'PennMARC::Database' do
+  include MarcSpecHelpers
+
+  let(:helper) { PennMARC::Database }
+  let(:record) do
+    marc_record(fields: [
+                  marc_field(tag: '943', subfields: { a: 'Humanities', '2': 'penncoi' }),
+                  marc_field(tag: '943', subfields: { a: 'Social Sciences', b: 'Linguistics', '2': 'penncoi' }),
+                  marc_field(tag: '944',
+                             subfields: { a: 'Database & Article Index',
+                                          b: 'Dictionaries and Thesauri (language based)' }),
+                  marc_field(tag: '944', subfields: { a: 'Database & Article Index', b: ['Reference and Handbooks'] })
+                ])
+  end
+  let(:record_uncurated_db) do
+    marc_record(fields: [
+                  marc_field(tag: '943', subfields: { a: 'Social Sciences', b: 'Linguistics', '2': 'penncoi' }),
+                  marc_field(tag: '944', subfields: { a: 'Uncurated Database', b: 'Reference and Handbooks' })
+                ])
+  end
+
+  describe '.type' do
+    it 'returns database types' do
+      expect(helper.type(record)).to contain_exactly('Dictionaries and Thesauri (language based)',
+                                                     'Reference and Handbooks')
+    end
+
+    context 'with uncurated database' do
+      it 'returns empty array' do
+        expect(helper.type(record_uncurated_db)).to be_empty
+      end
+    end
+  end
+
+  describe '.db_category' do
+    it 'returns database categories' do
+      expect(helper.db_category(record)).to contain_exactly('Humanities', 'Social Sciences')
+    end
+
+    context 'with uncurated database' do
+      it 'returns empty array' do
+        expect(helper.db_category(record_uncurated_db)).to be_empty
+      end
+    end
+  end
+
+  describe '.db_subcategory' do
+    it 'returns database subcategories' do
+      expect(helper.db_subcategory(record)).to contain_exactly('Social Sciences--Linguistics')
+    end
+
+    context 'with uncurated database' do
+      it 'returns empty array' do
+        expect(helper.db_subcategory(record_uncurated_db)).to be_empty
+      end
+    end
+  end
+end

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

@@ -0,0 +1,105 @@
+# frozen_string_literal: true
+
+describe 'PennMARC::Date' do
+  include MarcSpecHelpers
+
+  let(:helper) { PennMARC::Date }
+  let(:record) { marc_record fields: fields }
+
+  describe '.publication' do
+    let(:fields) { [marc_control_field(tag: '008', value: '130827s2010 nyu o 000 1 eng d')] }
+
+    it 'returns publication date' do
+      expect(helper.publication(record)).to eq(DateTime.new(2010))
+    end
+
+    it 'returns a year value' do
+      expect(helper.publication(record).year).to eq(2010)
+    end
+
+    context 'with invalid date' do
+      let(:fields) { [marc_control_field(tag: '008', value: 'invalid date')] }
+
+      it 'returns nil' do
+        expect(helper.publication(record)).to be_nil
+      end
+    end
+  end
+
+  describe '.added' do
+    context "with date formatted '%Y-%m-%d' " do
+      let(:fields) { [marc_field(tag: 'itm', subfields: { q: '2023-06-28' })] }
+
+      it 'returns expected value' do
+        expect(helper.added(record)).to eq(DateTime.strptime('2023-06-28', '%Y-%m-%d'))
+      end
+
+      it 'returns a year value' do
+        expect(helper.added(record).year).to eq(2023)
+      end
+    end
+
+    context "with date formatted '%Y-%m-%d %H:%M:%S'" do
+      let(:fields) { [marc_field(tag: 'itm', subfields: { q: '2023-06-29 11:04:30:10' })] }
+
+      it 'returns expected value' do
+        expect(helper.added(record)).to eq(DateTime.strptime('2023-06-29 11:04:30:10', '%Y-%m-%d %H:%M:%S'))
+      end
+
+      it 'returns a year value' do
+        expect(helper.added(record).year).to eq(2023)
+      end
+    end
+
+    context 'with multiple date added values' do
+      let(:fields) do
+        [marc_field(tag: 'itm', subfields: { q: '2023-06-28' }),
+         marc_field(tag: 'itm', subfields: { q: '2023-06-29' })]
+      end
+
+      it 'returns most recent date' do
+        expect(helper.added(record)).to eq(DateTime.strptime('2023-06-29', '%Y-%m-%d'))
+      end
+    end
+
+    context 'with invalid date' do
+      let(:fields) { [marc_field(tag: 'itm', subfields: { q: 'invalid date' })] }
+
+      it 'returns nil' do
+        expect(helper.added(record)).to be_nil
+      end
+
+      it 'outputs error message' do
+        expect do
+          helper.added(record)
+        end.to output("Error parsing date in date added subfield: invalid date - invalid date\n").to_stdout
+      end
+    end
+  end
+
+  describe '.last_updated' do
+    let(:fields) { [marc_field(tag: '005', subfields: { q: '20230213163851.1' })] }
+
+    it 'returns date last updated' do
+      expect(helper.last_updated(record)).to eq(DateTime.iso8601('20230213163851.1').to_datetime)
+    end
+
+    it 'returns year value' do
+      expect(helper.last_updated(record).year).to eq(2023)
+    end
+
+    context 'with invalid date' do
+      let(:fields) { [marc_field(tag: '005', subfields: { q: 'invalid date' })] }
+
+      it 'returns nil' do
+        expect(helper.last_updated(record)).to be_nil
+      end
+
+      it 'outputs error message' do
+        expect do
+          helper.last_updated(record)
+        end.to output("Error parsing last updated date: invalid date - invalid date\n").to_stdout
+      end
+    end
+  end
+end

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

@@ -0,0 +1,38 @@
+# frozen_string_literal: true
+
+describe 'PennMARC::Edition' do
+  include MarcSpecHelpers
+
+  let(:helper) { PennMARC::Edition }
+  let(:mapping) { { aut: 'Author' } }
+  let(:record) do
+    marc_record fields: [marc_field(tag: '250', subfields: { a: '5th Edition', b: 'Remastered' }),
+                         marc_field(tag: '880', subfields: { '6': '250', b: 'رمستر' }),
+                         marc_field(tag: '775', subfields: { i: 'Other Edition (Remove)',
+                                                             h: 'Cool Book',
+                                                             t: 'aut'}),
+                         marc_field(tag: '880', subfields: { '6': '775', i: 'Autre Editione' })]
+  end
+
+  describe '.show' do
+    it 'returns the editions' do
+      expect(helper.show(record)).to contain_exactly('5th Edition Remastered', 'رمستر')
+    end
+  end
+
+  describe '.values' do
+    it 'returns the values' do
+      expect(helper.values(record)).to eq('5th Edition Remastered')
+    end
+  end
+
+  describe '.other_show' do
+    it 'returns other edition values' do
+      expect(helper.other_show(record, mapping)).to contain_exactly('Autre Editione',
+                                                                    'Other Edition: Author. (Cool Book)')
+    end
+  end
+end
+
+
+