gutenberg_rdf 0.2.3 → 0.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 085a4adadf222928edc61b7660ae01c77f519682
-  data.tar.gz: 374a75dfb36da499bf3657775630ebda8ace5e64
+  metadata.gz: 5eb92b0007804d96f27544f3e4248ed20d9a6a33
+  data.tar.gz: 4d58bb7b868157dfe461a7a0633ed75eb7933af2
 SHA512:
-  metadata.gz: 3b19af4c7c7f1b58b10adff65dd47d43900248966fd155cd1e76bd4e4879660148ea4d1dfba136ab0403ae049311241ade185056a8fa91b7ea5ebb737900c82d
-  data.tar.gz: 2485ba29c8f3e31f17004366e9fe3cc0f3c76453d793ec55db7c5cbbbc7dafdd804e01b2196fc31479004fd9e5726a6c83809ea034d5875cd52bc754d7d63ca7
+  metadata.gz: 74a4672ad2629c43853abcd4d15f6c033572ad0f71b0eb5177fa1c84dac697a35a775e0e1ba660fea2920b1edf69066aa2b11429fb31b36d5539c44f3d99fcfb
+  data.tar.gz: bcf6c1bc87aa2e6692ae7db25ae69d2a137467dba227e0702eacc02d8345e901dda2de9f6f5e7fedf32bba98ee47fc639f8008d9dcd55d2757644dccbc6321e7

data/gutenberg_rdf.gemspec

@@ -9,7 +9,7 @@ Gem::Specification.new do |spec|
   spec.email         = ["m@mikecook.co.uk"]
   spec.summary       = %q{A Ruby wrapper for the Project Gutenberg RDF catalog files.}
   spec.description   = %q{A Ruby wrapper for the Project Gutenberg RDF catalog files (requires Ruby 2).}
-  spec.homepage      = ""
+  spec.homepage      = "https://github.com/mrcook/gutenberg_rdf"
   spec.license       = "MIT"
 
   spec.files         = `git ls-files`.split($/)
@@ -19,5 +19,5 @@ Gem::Specification.new do |spec|
 
   spec.required_ruby_version = ">= 2.0.0" # so we have UTF-8 by default
 
-  spec.add_development_dependency "rspec", "~> 2.14.1"
+  spec.add_development_dependency "rspec", "~> 2.14"
 end

data/lib/gutenberg_rdf/rdf.rb

@@ -80,14 +80,32 @@ module GutenbergRdf
       title_array = t.split(/\n/)
       title_array = title_array.first.split(/:/) if title_array.count == 1
       title_array = title_array.first.split(/;/) if title_array.count == 1
+      title_array = title_array.first.split(/, or,/) if title_array.count == 1
 
-      title_array.each(&:strip!)
+      title_array.map(&:strip)
+    end
+
+    def roles
+      @roles ||= extract_roles
+    end
+
+    def extract_roles
+      entries = Hash.new
+      xml.elements.each('pgterms:ebook/dcterms:creator') do |entry|
+        entries["#{entry.attributes['rdf:resource'].sub('2009/agents/', '')}"] = 'aut'
+      end
+      xml.elements.each('pgterms:ebook/marcrel:*') do |entry|
+        entries["#{entry.attributes['rdf:resource'].sub('2009/agents/', '')}"] = entry.name
+      end
+      entries
     end
 
     def extract_authors
       entries = Array.new
       xml.elements.each('pgterms:agent') do |agent|
-        entries << Agent.new(agent)
+        entry = Agent.new(agent)
+        entry.assign_role(roles)
+        entries << entry
       end
       entries
     end

data/lib/gutenberg_rdf/rdf/agent.rb

@@ -2,6 +2,7 @@ module GutenbergRdf
   class Rdf
     class Agent
       attr_reader :xml
+      attr_accessor :role
 
       def initialize(xml)
         @xml = xml
@@ -15,6 +16,14 @@ module GutenbergRdf
         xml.attributes['about'].match(/\A\d\d\d\d\/agents\/(\d+)\z/)[1]
       end
 
+      def role
+        @role ||= 'oth'
+      end
+
+      def assign_role(roles)
+        self.role = roles["#{id}"]
+      end
+
       def fullname
         [firstname, lastname].reject(&:empty?).join(' ')
       end

data/lib/gutenberg_rdf/version.rb

@@ -1,3 +1,3 @@
 module GutenbergRdf
-  VERSION = "0.2.3"
+  VERSION = "0.3.1"
 end

data/spec/gutenberg_rdf/rdf/agent_spec.rb

@@ -5,53 +5,66 @@ module GutenbergRdf
     describe Agent do
       let(:agent) do
         xml = '<rdf:RDF xmlns:dcterms="http://purl.org/dc/terms/" xmlns:pgterms="http://www.gutenberg.org/2009/pgterms/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
-                <pgterms:agent rdf:about="2009/agents/402">
-                  <pgterms:birthdate rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">1830</pgterms:birthdate>
-                  <pgterms:deathdate rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">1905</pgterms:deathdate>
-                  <pgterms:name>Doe, Jon James</pgterms:name>
-                  <pgterms:alias>Doe, Jon</pgterms:alias>
-                  <pgterms:alias>Doe, J. J.</pgterms:alias>
-                  <pgterms:webpage rdf:resource="http://en.wikipedia.org/wiki/Jon_James_Doe"/>
-                </pgterms:agent>
-              </rdf:RDF>'
+                 <pgterms:agent rdf:about="2009/agents/402">
+                   <pgterms:birthdate rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">1830</pgterms:birthdate>
+                   <pgterms:deathdate rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">1905</pgterms:deathdate>
+                   <pgterms:name>Doe, Jon James</pgterms:name>
+                   <pgterms:alias>Doe, Jon</pgterms:alias>
+                   <pgterms:alias>Doe, J. J.</pgterms:alias>
+                   <pgterms:webpage rdf:resource="http://en.wikipedia.org/wiki/Jon_James_Doe"/>
+                 </pgterms:agent>
+               </rdf:RDF>'
         Agent.new(REXML::Document.new(xml).root.elements['pgterms:agent'])
       end
 
       it "expects an agent ID" do
-        expect(agent.id).to eql '402'
+        expect(agent.id).to eq '402'
+      end
+
+      it "sets a default role" do
+        expect(agent.role).to eq 'oth'
+      end
+
+      describe "Assigning Roles" do
+        it "assigns the correct value to .role" do
+          roles = {'402' => 'aut', '116' => 'ctb'}
+          agent.assign_role(roles)
+
+          expect(agent.role).to eq 'aut'
+        end
       end
 
       it "expects the last name" do
-        expect(agent.lastname).to eql 'Doe'
+        expect(agent.lastname).to eq 'Doe'
       end
 
       it "expects the first name(s)" do
-        expect(agent.firstname).to eql 'Jon James'
+        expect(agent.firstname).to eq 'Jon James'
       end
 
       it "expects the full name" do
-        expect(agent.fullname).to eql 'Jon James Doe'
+        expect(agent.fullname).to eq 'Jon James Doe'
       end
 
       it "returns the #fullname when to_s is called" do
-        expect(agent.to_s).to eql 'Jon James Doe'
+        expect(agent.to_s).to eq 'Jon James Doe'
       end
 
       it "expects a birth date" do
-        expect(agent.birthdate).to eql '1830'
+        expect(agent.birthdate).to eq '1830'
       end
 
       it "expects a death date" do
-        expect(agent.deathdate).to eql '1905'
+        expect(agent.deathdate).to eq '1905'
       end
 
       it "expects a webpage" do
-        expect(agent.webpage).to eql 'http://en.wikipedia.org/wiki/Jon_James_Doe'
+        expect(agent.webpage).to eq 'http://en.wikipedia.org/wiki/Jon_James_Doe'
       end
 
       it "expects any alias names" do
-        expect(agent.aliases[0]).to eql 'Doe, Jon'
-        expect(agent.aliases[1]).to eql 'Doe, J. J.'
+        expect(agent.aliases[0]).to eq 'Doe, Jon'
+        expect(agent.aliases[1]).to eq 'Doe, J. J.'
       end
 
       context "when only a single name is given" do
@@ -65,10 +78,10 @@ module GutenbergRdf
         end
 
         it "expects it to be assigned to the last name" do
-          expect(agent.lastname).to eql 'Dato'
+          expect(agent.lastname).to eq 'Dato'
         end
         it "expects firstname to be an empty string" do
-          expect(agent.firstname).to eql ''
+          expect(agent.firstname).to eq ''
         end
       end
 
@@ -83,8 +96,8 @@ module GutenbergRdf
         end
 
         it "expects the correct name order" do
-          expect(agent.firstname).to eql 'Sir Jon'
-          expect(agent.lastname).to eql 'Doe'
+          expect(agent.firstname).to eq 'Sir Jon'
+          expect(agent.lastname).to eq 'Doe'
         end
       end
 
@@ -100,8 +113,8 @@ module GutenbergRdf
 
         it "expects initials to replaced by name in brackets" do
           pending "Not yet implemented"
-          expect(agent.firstname).to eql 'Jon James'
-          expect(agent.lastname).to eql 'Doe'
+          expect(agent.firstname).to eq 'Jon James'
+          expect(agent.lastname).to eq 'Doe'
         end
         it "expects the name (excluding name in brackets) to be added to the aliases"
         it "should not have duplicate aliases"

data/spec/gutenberg_rdf/rdf_spec.rb

@@ -86,6 +86,24 @@ module GutenbergRdf
         end
       end
 
+      context "with; title, or, subtitle (we need to split on the 'or')" do
+        let(:xml) do
+          '<rdf:RDF xmlns:dcterms="http://purl.org/dc/terms/" xmlns:pgterms="http://www.gutenberg.org/2009/pgterms/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
+            <pgterms:ebook rdf:about="ebooks/98765">
+              <dcterms:title>A Great Multi-Title, or, a Subtitle</dcterms:title>
+            </pgterms:ebook>
+          </rdf:RDF>'
+        end
+        let(:rdf) { Rdf.new(REXML::Document.new(xml)) }
+
+        it "expects the title to be the first line" do
+          expect(rdf.title).to eql 'A Great Multi-Title'
+        end
+        it "expects the subtitle to be the second line" do
+          expect(rdf.subtitle).to eql 'a Subtitle'
+        end
+      end
+
       context "when title:subtitle are separated by a colon" do
         let(:xml) do
           '<rdf:RDF xmlns:dcterms="http://purl.org/dc/terms/" xmlns:pgterms="http://www.gutenberg.org/2009/pgterms/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
@@ -142,7 +160,11 @@ module GutenbergRdf
 
     describe "#authors" do
       let(:xml) do
-        '<rdf:RDF xmlns:dcterms="http://purl.org/dc/terms/" xmlns:pgterms="http://www.gutenberg.org/2009/pgterms/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
+        '<rdf:RDF xmlns:dcterms="http://purl.org/dc/terms/" xmlns:pgterms="http://www.gutenberg.org/2009/pgterms/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:marcrel="http://id.loc.gov/vocabulary/relators">
+          <pgterms:ebook rdf:about="ebooks/99999999">
+            <marcrel:ctb rdf:resource="2009/agents/402"/>
+            <dcterms:creator rdf:resource="2009/agents/116"/>
+          </pgterms:ebook>
           <pgterms:agent rdf:about="2009/agents/402">
             <pgterms:birthdate rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">1830</pgterms:birthdate>
             <pgterms:deathdate rdf:datatype="http://www.w3.org/2001/XMLSchema#integer">1905</pgterms:deathdate>
@@ -166,7 +188,12 @@ module GutenbergRdf
       end
       it "has the correct author names" do
         expect(rdf.authors.first.fullname).to eq 'Mary Mapes Dodge'
-        expect(rdf.authors.last.fullname).to eq 'Various'
+      end
+      it "expects the author to have an aut role" do
+        expect(rdf.authors.last.role).to eq 'aut'
+      end
+      it "expects other agents to have the correct role" do
+        expect(rdf.authors.first.role).to eq 'ctb'
       end
     end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: gutenberg_rdf
 version: !ruby/object:Gem::Version
-  version: 0.2.3
+  version: 0.3.1
 platform: ruby
 authors:
 - Mike Cook
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-01-27 00:00:00.000000000 Z
+date: 2014-02-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -16,14 +16,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.14.1
+        version: '2.14'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 2.14.1
+        version: '2.14'
 description: A Ruby wrapper for the Project Gutenberg RDF catalog files (requires
   Ruby 2).
 email:
@@ -49,7 +49,7 @@ files:
 - spec/gutenberg_rdf/rdf_spec.rb
 - spec/gutenberg_rdf_spec.rb
 - spec/spec_helper.rb
-homepage: ''
+homepage: https://github.com/mrcook/gutenberg_rdf
 licenses:
 - MIT
 metadata: {}