berkeley_library-av-core 0.4.0

data/spec/data/search-b23305516.xml

@@ -0,0 +1,81 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!-- Search-Engine-Total-Number-Of-Results: 1 -->
+<collection xmlns="http://www.loc.gov/MARC21/slim">
+    <record>
+        <controlfield tag="001">19816</controlfield>
+        <controlfield tag="005">20190703231952.0</controlfield>
+        <datafield tag="035" ind1=" " ind2=" ">
+            <subfield code="a">(pacradio)00107</subfield>
+        </datafield>
+        <datafield tag="041" ind1=" " ind2=" ">
+            <subfield code="a">eng</subfield>
+        </datafield>
+        <datafield tag="245" ind1=" " ind2=" ">
+            <subfield code="a">Author Alice Walker reads the short story, Roselily</subfield>
+        </datafield>
+        <datafield tag="260" ind1=" " ind2=" ">
+            <subfield code="a">Los Angeles </subfield>
+            <subfield code="b">Pacifica Radio Archives</subfield>
+            <subfield code="c">1994.</subfield>
+        </datafield>
+        <datafield tag="269" ind1=" " ind2=" ">
+            <subfield code="a">1994</subfield>
+        </datafield>
+        <datafield tag="300" ind1=" " ind2=" ">
+            <subfield code="a">1 online resource.</subfield>
+        </datafield>
+        <datafield tag="336" ind1=" " ind2=" ">
+            <subfield code="a">Audio</subfield>
+        </datafield>
+        <datafield tag="520" ind1=" " ind2=" ">
+            <subfield code="a">Pulitzer Prize-winning author Alice Walker (1944 - ) reads her short story "Roselily" (written ca. 1967). Walker then talks with KPFA's Ginny Z. Berson about the story, how she wrote it about her first marriage to a civil rights attorney and how it questions the dominance of Christianity over all religions in the US. The story was dropped from a 10th grade English test by California state educators based on complaints from Christian conservatives in early 1994. The Traditional Values Coalition applied pressure based on what they read as "anti-religious" elements in the story.</subfield>
+        </datafield>
+        <datafield tag="536" ind1=" " ind2=" ">
+            <subfield code="a">Sponsored by the National Historical Publications and Records Commission at the National Archives and Records Administration as part of Pacifica's American Women Making History and Culture: 1963-1982 grant preservation project.</subfield>
+        </datafield>
+        <datafield tag="540" ind1=" " ind2=" ">
+            <subfield code="a">RESTRICTED.  Permissions, licensing requests, and all other inquiries should be directed in writing to: Director of the Archives, Pacifica Radio Archives, 3729 Cahuenga Blvd. West, North Hollywood, CA 91604, 800-735-0230 x 263, fax 818-506-1084, info@pacificaradioarchives.org, http://www.pacificaradioarchives.org</subfield>
+        </datafield>
+        <datafield tag="700" ind1="1" ind2=" ">
+            <subfield code="a">Berson, Ginny Z.</subfield>
+            <subfield code="e">interviewer.</subfield>
+        </datafield>
+        <datafield tag="700" ind1="1" ind2=" ">
+            <subfield code="a">Walker, Alice.</subfield>
+            <subfield code="e">interviewee.</subfield>
+        </datafield>
+        <datafield tag="710" ind1="2" ind2=" ">
+            <subfield code="a">Pacifica Radio Archive.</subfield>
+        </datafield>
+        <datafield tag="710" ind1="2" ind2=" ">
+            <subfield code="a">KPFA (Radio station : Berkeley, Calif.).</subfield>
+        </datafield>
+        <datafield tag="852" ind1=" " ind2=" ">
+            <subfield code="c">The Library</subfield>
+        </datafield>
+        <datafield tag="856" ind1="4" ind2="1">
+            <subfield code="u">http://oskicat.berkeley.edu/record=b23305516</subfield>
+            <subfield code="y">View library catalog record.</subfield>
+        </datafield>
+        <datafield tag="856" ind1="4" ind2="2">
+            <subfield code="u">http://servlet1.lib.berkeley.edu:8080/audio/stream.play.logic?coll=pacifica&amp;group=b23305516</subfield>
+            <subfield code="y">Play Audio for American Women Making History and Culture. Freely available for streaming.</subfield>
+        </datafield>
+        <datafield tag="901" ind1=" " ind2=" ">
+            <subfield code="m">b23305516</subfield>
+            <subfield code="o">947286769</subfield>
+        </datafield>
+        <datafield tag="903" ind1=" " ind2=" ">
+            <subfield code="b">m</subfield>
+        </datafield>
+        <datafield tag="980" ind1=" " ind2=" ">
+            <subfield code="a">pacradio</subfield>
+        </datafield>
+        <datafield tag="982" ind1=" " ind2=" ">
+            <subfield code="a">Pacifica Radio Archives</subfield>
+            <subfield code="b">Pacifica Radio Archives Social Activism Sound Recording Project</subfield>
+        </datafield>
+    </record>
+
+    
+</collection>

data/spec/lib/berkeley_library/av/av_spec.rb

@@ -0,0 +1,12 @@
+require 'spec_helper'
+
+module BerkeleyLibrary
+  describe AV do
+    describe :configure do
+      it 'exposes the configuration' do
+        conf = AV.configure { config }
+        expect(conf).to eq(AV::Config)
+      end
+    end
+  end
+end

data/spec/lib/berkeley_library/av/config_spec.rb

@@ -0,0 +1,250 @@
+require 'spec_helper'
+require 'uri'
+
+module BerkeleyLibrary
+  module AV
+    describe Config do
+
+      after do
+        Config.send(:clear!)
+      end
+
+      describe(:configured?) do
+        it 'defaults to false' do
+          expect(AV.configured?).to eq(false)
+        end
+
+        it 'returns true if and only if all values are configured' do
+          settings = {
+            avplayer_base_uri: 'http://avplayer.example.edu',
+            alma_sru_host: 'berkeley.alma.exlibrisgroup.com',
+            alma_primo_host: 'search.library.berkeley.edu',
+            alma_institution_code: '01UCS_BER',
+            alma_permalink_key: 'iqob43',
+            tind_base_uri: 'http://tind.example.edu',
+            wowza_base_uri: 'http://wowza.example.edu'
+          }
+          settings.each { |setting, value| Config.send("#{setting}=", value) }
+          expect(AV.configured?).to eq(true)
+
+          aggregate_failures do
+            settings.each do |setting, value|
+              Config.instance_variable_set("@#{setting}".to_sym, nil)
+              expect(AV.configured?).to eq(false), "Clearing #{setting} did not set configured? to false"
+              Config.send("#{setting}=", value)
+            end
+          end
+        end
+
+        it 'reads values from a Rails config if present' do
+          settings = {
+            avplayer_base_uri: URI.parse('http://avplayer.example.edu'),
+            alma_sru_host: 'berkeley.alma.exlibrisgroup.com',
+            alma_primo_host: 'search.library.berkeley.edu',
+            alma_institution_code: '01UCS_BER',
+            alma_permalink_key: 'iqob43',
+            tind_base_uri: URI.parse('http://tind.example.edu'),
+            wowza_base_uri: URI.parse('http://wowza.example.edu')
+          }
+          rails_config = Struct.new(*settings.keys, keyword_init: true).new(**settings)
+
+          # Mock Rails config
+          expect(defined?(Rails)).to be_nil # just to be sure
+          Object.send(:const_set, 'Rails', Struct.new(:application).new)
+          Rails.application = Struct.new(:config).new
+          Rails.application.config = rails_config
+
+          aggregate_failures do
+            settings.each do |setting, v|
+              expect(Config.send(setting)).to eq(v)
+            end
+          end
+
+          expect(AV.configured?).to eq(true)
+        ensure
+          Object.send(:remove_const, 'Rails')
+        end
+      end
+
+      describe(:missing) do
+        it 'defaults to all settings' do
+          expect(Config.missing).to eq(Config::REQUIRED_SETTINGS)
+        end
+
+        it 'returns an empty array if nothing is missing' do
+          settings = {
+            avplayer_base_uri: 'http://avplayer.example.edu',
+            alma_sru_host: 'berkeley.alma.exlibrisgroup.com',
+            alma_primo_host: 'search.library.berkeley.edu',
+            alma_institution_code: '01UCS_BER',
+            alma_permalink_key: 'iqob43',
+            tind_base_uri: 'http://tind.example.edu',
+            wowza_base_uri: 'http://wowza.example.edu'
+          }
+          settings.each { |setting, value| Config.send("#{setting}=", value) }
+
+          expect(Config.missing).to eq([])
+        end
+
+        it 'returns the missing settings' do
+          settings = {
+            avplayer_base_uri: 'http://avplayer.example.edu',
+            alma_sru_host: 'berkeley.alma.exlibrisgroup.com',
+            alma_primo_host: 'search.library.berkeley.edu',
+            alma_institution_code: '01UCS_BER',
+            alma_permalink_key: 'iqob43',
+            tind_base_uri: 'http://tind.example.edu',
+            wowza_base_uri: 'http://wowza.example.edu'
+          }
+          settings.each { |setting, value| Config.send("#{setting}=", value) }
+
+          expected = []
+          settings.each_key do |setting|
+            Config.instance_variable_set("@#{setting}".to_sym, nil)
+            expected << setting
+            expect(Config.missing).to eq(expected)
+          end
+        end
+      end
+
+      describe :avplayer_base_uri= do
+        it 'converts strings to URIs' do
+          expected_uri = URI.parse('http://avplayer.example.edu')
+          Config.avplayer_base_uri = expected_uri.to_s
+          expect(Config.avplayer_base_uri).to eq(expected_uri)
+        end
+
+        it 'strips trailing slashes' do
+          expected_uri = URI.parse('http://avplayer.example.edu')
+          Config.avplayer_base_uri = "#{expected_uri}/"
+          expect(Config.avplayer_base_uri).to eq(expected_uri)
+        end
+      end
+
+      describe :alma_sru_host= do
+        it 'sets the hostname' do
+          expected_host = 'alma.example.org'
+          Config.alma_sru_host = expected_host
+          expect(Config.alma_sru_host).to eq(expected_host)
+        end
+      end
+
+      describe :alma_institution_code= do
+        it 'sets the Alma institution code' do
+          expected_code = '01UCS_BER'
+          Config.alma_institution_code = expected_code
+          expect(Config.alma_institution_code).to eq(expected_code)
+        end
+      end
+
+      describe :alma_sru_base_uri do
+        it 'generates the base URI from the SRU host and institution code' do
+          expected_host = 'alma.example.org'
+          Config.alma_sru_host = expected_host
+
+          expected_code = '01UCS_BER'
+          Config.alma_institution_code = expected_code
+
+          base_uri = Config.alma_sru_base_uri
+          expect(base_uri.host).to eq(expected_host)
+          expect(base_uri.path).to end_with("/#{expected_code}")
+        end
+
+        it 'requires both the SRU host and institution code to be set' do
+          expect { Config.alma_sru_base_uri }.to raise_error(ArgumentError)
+
+          Config.alma_sru_host = 'alma.example.org'
+          Config.alma_institution_code = '01UCS_BER'
+
+          Config.alma_sru_base_uri
+        end
+      end
+
+      describe :wowza_base_uri= do
+        it 'converts strings to URIs' do
+          expected_uri = URI.parse('http://wowza.example.edu')
+          Config.wowza_base_uri = expected_uri.to_s
+          expect(Config.wowza_base_uri).to eq(expected_uri)
+        end
+
+        it 'strips trailing slashes' do
+          expected_uri = URI.parse('http://wowza.example.edu')
+          Config.wowza_base_uri = "#{expected_uri}/"
+          expect(Config.wowza_base_uri).to eq(expected_uri)
+        end
+      end
+
+      describe :tind_base_uri= do
+        it 'converts strings to URIs' do
+          expected_uri = URI.parse('http://tind.example.edu')
+          Config.tind_base_uri = expected_uri.to_s
+          expect(Config.tind_base_uri).to eq(expected_uri)
+        end
+
+        it 'strips trailing slashes' do
+          expected_uri = URI.parse('http://tind.example.edu')
+          Config.tind_base_uri = "#{expected_uri}/"
+          expect(Config.tind_base_uri).to eq(expected_uri)
+        end
+      end
+
+      describe 'with Rails config' do
+        attr_reader :config
+
+        before do
+          @config = double(Config)
+
+          application = double(Object)
+          allow(application).to receive(:config).and_return(config)
+
+          rails = double(Object)
+          allow(rails).to receive(:application).and_return(application)
+
+          Object.const_set(:Rails, rails)
+        end
+
+        after do
+          Object.send(:remove_const, :Rails)
+        end
+
+        describe :avplayer_base_uri do
+          attr_reader :avplayer_base_uri
+
+          before do
+            @avplayer_base_uri = URI.parse('http://avplayer.example.edu')
+            allow(config).to receive(:avplayer_base_uri).and_return(avplayer_base_uri.to_s)
+          end
+
+          it 'falls back to the Rails config, if available' do
+            expect(Config.avplayer_base_uri).to eq(avplayer_base_uri)
+          end
+
+          it 'prefers the explicitly configured URI' do
+            expected_avplayer_uri = URI.parse('https://avplayer-other.example.edu')
+            Config.avplayer_base_uri = expected_avplayer_uri
+            expect(Config.avplayer_base_uri).to eq(expected_avplayer_uri)
+          end
+        end
+
+        describe :tind_base_uri do
+          attr_reader :tind_base_uri
+
+          before do
+            @tind_base_uri = URI.parse('http://tind.example.edu')
+            allow(config).to receive(:tind_base_uri).and_return(tind_base_uri.to_s)
+          end
+
+          it 'falls back to the Rails config, if available' do
+            expect(Config.tind_base_uri).to eq(tind_base_uri)
+          end
+
+          it 'prefers the explicitly configured URI' do
+            expected_tind_uri = URI.parse('https://tind-other.example.edu')
+            Config.tind_base_uri = expected_tind_uri
+            expect(Config.tind_base_uri).to eq(expected_tind_uri)
+          end
+        end
+      end
+    end
+  end
+end

data/spec/lib/berkeley_library/av/marc/util_spec.rb

@@ -0,0 +1,150 @@
+require 'spec_helper'
+require 'marc'
+
+module BerkeleyLibrary
+  module AV
+    module Marc
+      module Util
+        describe :grouped_subfield_values do
+          let(:df) { MARC::DataField.new('999') }
+          let(:original_groups) do
+            [
+              { b: '2', a: '1', c: '7' },
+              { b: '4', a: '3', c: '9' },
+              { b: '6', a: '5', c: '10' }
+            ]
+          end
+
+          let(:order) { %i[c b a] }
+          let(:ordered_groups) do
+            [
+              { c: '7', b: '2', a: '1' },
+              { c: '9', b: '4', a: '3' },
+              { c: '10', b: '6', a: '5' }
+            ]
+          end
+
+          describe 'pre-grouped subfields' do
+            before do
+              original_groups.each do |g|
+                g.each { |c, v| df.subfields << MARC::Subfield.new(c, v) }
+              end
+            end
+
+            it 'returns the groups' do
+              groups = Util.group_subfield_values(df)
+              expect(groups).to eq(original_groups)
+            end
+
+            it 'reorders the groups' do
+              reordered = Util.group_subfield_values(df, order:)
+              # Hash equality doesn't check order, so we do it by hand
+              reordered.each_with_index do |actual, i|
+                expected = ordered_groups[i]
+                expect(actual.keys).to eq(expected.keys), "#{i}: expected: #{expected}, got: #{actual}"
+                expect(actual.values).to eq(expected.values), "#{i}: expected: #{expected}, got: #{actual}"
+              end
+            end
+          end
+
+          describe 'pre-grouped by code' do
+            before do
+              {
+                b: %w[2 4 6],
+                a: %w[1 3 5],
+                c: %w[7 9 10]
+              }.each do |c, vv|
+                vv.each { |v| df.subfields << MARC::Subfield.new(c, v) }
+              end
+            end
+
+            it 'groups the values' do
+              groups = Util.group_subfield_values(df)
+              expect(groups).to eq(original_groups)
+            end
+
+            it 'reorders the groups' do
+              reordered = Util.group_subfield_values(df, order:)
+              # Hash equality doesn't check order, so we do it by hand
+              reordered.each_with_index do |actual, i|
+                expected = ordered_groups[i]
+                expect(actual.keys).to eq(expected.keys), "#{i}: expected: #{expected}, got: #{actual}"
+                expect(actual.values).to eq(expected.values), "#{i}: expected: #{expected}, got: #{actual}"
+              end
+            end
+          end
+
+          describe 'inconsistent order' do
+            let(:inconsistent_groups) do
+              [
+                { a: '1', b: '2', c: '7' },
+                { b: '4', a: '3', c: '9' },
+                { c: '10', a: '5', b: '6' }
+              ]
+            end
+
+            before do
+              inconsistent_groups.each do |g|
+                g.each { |c, v| df.subfields << MARC::Subfield.new(c, v) }
+              end
+            end
+
+            it 'groups the values' do
+              groups = Util.group_subfield_values(df)
+              expect(groups).to eq(inconsistent_groups)
+            end
+
+            it 'reorders the groups' do
+              reordered = Util.group_subfield_values(df, order:)
+              # Hash equality doesn't check order, so we do it by hand
+              reordered.each_with_index do |actual, i|
+                expected = ordered_groups[i]
+                expect(actual.keys).to eq(expected.keys), "#{i}: expected: #{expected}, got: #{actual}"
+                expect(actual.values).to eq(expected.values), "#{i}: expected: #{expected}, got: #{actual}"
+              end
+            end
+          end
+
+          describe 'ragged groups' do
+            let(:ragged_groups) do
+              [{ a: '1' }, { b: '2' }, { c: '7' }, { b: '4' }, { c: '9' }, { a: '5' }, { b: '6' }]
+            end
+            let(:ragged_reordered) do
+              [
+                { c: '7', b: '2', a: '1' },
+                { c: '9', b: '4', a: '5' },
+                { b: '6' }
+              ]
+            end
+
+            before do
+              ragged_groups.each do |g|
+                g.each { |c, v| df.subfields << MARC::Subfield.new(c, v) }
+              end
+            end
+
+            it 'groups the values' do
+              expected = [
+                { a: '1', b: '2', c: '7' },
+                { a: '5', b: '4', c: '9' },
+                { b: '6' }
+              ]
+              groups = Util.group_subfield_values(df)
+              expect(groups).to eq(expected)
+            end
+
+            it 'reorders the values' do
+              reordered = Util.group_subfield_values(df, order:)
+              # Hash equality doesn't check order, so we do it by hand
+              reordered.each_with_index do |actual, i|
+                expected = ragged_reordered[i]
+                expect(actual.keys).to eq(expected.keys), "#{i}: expected: #{expected}, got: #{actual}"
+                expect(actual.values).to eq(expected.values), "#{i}: expected: #{expected}, got: #{actual}"
+              end
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/spec/lib/berkeley_library/av/marc_spec.rb

@@ -0,0 +1,62 @@
+require 'spec_helper'
+
+module BerkeleyLibrary
+  module AV
+    describe Marc do
+      describe :from_xml do
+        it 'returns a single record' do
+          marc_xml = File.read('spec/data/search-1993.xml')
+          record = Marc.from_xml(marc_xml)
+          expect(record).to be_a(MARC::Record)
+        end
+      end
+
+      describe :all_from_xml do
+        it 'returns multiple records' do
+          marc_xml = File.read('spec/data/search-1993.xml')
+          reader = Marc.all_from_xml(marc_xml)
+          records = reader.to_a
+          expect(records.size).to eq(2)
+          records.each do |record|
+            expect(record).to be_a(MARC::Record)
+          end
+        end
+      end
+
+      describe :reader_for do
+        it 'returns a reader for XML' do
+          marc_path = 'spec/data/record-(cityarts)00002.xml'
+          reader = Marc.reader_for(marc_path)
+          record = reader.first
+          expect(record['245'].value).to eq('826 Spelling Bee For Cheaters, February 17, 2011')
+        end
+
+        it 'returns a reader for binary MARC' do
+          marc_path = 'spec/data/10.23.19.JessieLaCavalier.02.mrc'
+          reader = Marc.reader_for(marc_path)
+          record = reader.first
+          expect(record['100'].value).to eq('LeCavalier, Jesse.')
+        end
+
+        it 'raises an error for non-MARC files' do
+          marc_path = 'I am not a MARC file'
+          expect { Marc.reader_for(marc_path) }.to raise_error(ArgumentError)
+        end
+      end
+
+      describe :read do
+        it 'reads XML' do
+          marc_path = 'spec/data/record-(cityarts)00002.xml'
+          record = Marc.read(marc_path)
+          expect(record['245'].value).to eq('826 Spelling Bee For Cheaters, February 17, 2011')
+        end
+
+        it 'reads binary MARC' do
+          marc_path = 'spec/data/10.23.19.JessieLaCavalier.02.mrc'
+          record = Marc.read(marc_path)
+          expect(record['100'].value).to eq('LeCavalier, Jesse.')
+        end
+      end
+    end
+  end
+end

data/spec/lib/berkeley_library/av/metadata/field_spec.rb

@@ -0,0 +1,81 @@
+require 'spec_helper'
+require 'marc'
+
+module BerkeleyLibrary
+  module AV
+    class Metadata
+      describe Field do
+        attr_reader :marc_record
+
+        before do
+          @marc_record = MARC::XMLReader.new('spec/data/record-(pacradio)01469.xml').first
+        end
+
+        it 'extracts the text from whole datafields' do
+          field = Field.new(order: 2, label: 'Creator', spec: '700')
+          value = field.value_from(marc_record)
+          expect(value).to be_a(Value)
+
+          expected_entries = [
+            'Coleman, Wanda. interviewee.',
+            'Adisa, Opal Palmer. interviewer.'
+          ]
+          expect(value.entries).to eq(expected_entries)
+        end
+
+        it 'extracts the text from subfields without groups' do
+          field = Field.new(order: 85, label: 'Usage Statement', spec: '540$a')
+          value = field.value_from(marc_record)
+          expect(value).to be_a(Value)
+
+          expected_entries = ['RESTRICTED.  Permissions, licensing requests, and all other inquiries should be directed in writing to: Director of the Archives, Pacifica Radio Archives, 3729 Cahuenga Blvd. West, North Hollywood, CA 91604, 800-735-0230 x 263, fax 818-506-1084, info@pacificaradioarchives.org, http://www.pacificaradioarchives.org']
+          expect(value.entries).to eq(expected_entries)
+        end
+
+        # TODO: test multiple groups
+        it 'extracts the text from grouped subfields' do
+          field = Field.new(order: 66, label: 'Grant Information', spec: '536', subfield_order: %w[a o m n])
+          value = field.value_from(marc_record)
+          expect(value).to be_a(Value)
+          expected_entries = ["Sponsored by the National Historical Publications and Records Commission at the National Archives and Records Administration as part of Pacifica's American Women Making History and Culture: 1963-1982 grant preservation project."]
+          expect(value.entries).to eq(expected_entries)
+        end
+
+        context 'links' do
+          it 'extracts links from TIND records' do
+            field = Field.new(order: 9999, label: 'Link to Content', spec: '856{^1=\4}{^2=\2}', subfield_order: %w[u y])
+            value = field.value_from(marc_record)
+            expect(value).to be_a(Value)
+            expected_link = AV::Metadata::Link.new(
+              url: 'https://avplayer.lib.berkeley.edu/Pacifica/b23305522',
+              body: 'Play Audio for American Women Making History and Culture. Freely available for streaming.'
+            )
+            expect(value.entries).to contain_exactly(expected_link)
+          end
+
+          it 'extracts links from Alma records' do
+            marc_record = MARC::XMLReader.new('spec/data/alma/991054360089706532-sru.xml').first
+            field = Field.new(order: 9999, label: 'Link to Content', spec: '956{^1=\4}{^2=\0}', subfield_order: %w[u z])
+            value = field.value_from(marc_record)
+            expect(value).to be_a(Value)
+            expected_link = AV::Metadata::Link.new(
+              url: 'https://avplayer.lib.berkeley.edu/Video-UCBOnly-MRC/b25716973',
+              body: 'UCB Access.'
+            )
+            expect(value.entries).to contain_exactly(expected_link)
+          end
+        end
+
+        describe :hash do
+          it 'returns the same hash for identical Fields' do
+            f1 = Field.new(order: 2, label: 'Description', spec: '520$a')
+            f2 = Field.new(order: 2, label: 'Description', spec: '520$a')
+            expect(f1).to eq(f2) # just to be sure
+
+            expect(f1.hash).to eq(f2.hash)
+          end
+        end
+      end
+    end
+  end
+end