epub-parser 0.0.7 → 0.0.8

data/README.markdown

@@ -1,15 +1,26 @@
 EPUB Parser
 ===========
 
-INSTALL
+INSTALLATION
 -------
     gem install epub-parser  
 
 USAGE
 -----
+
+### As a command line tool
+
+	epubinfo path/to/book.epub
+
+For more info:
+
+	epubinfo -h
+
+### As a library
+
     require 'epub/parser'
     
-    book = EPUB::Parser.parse 'book.epub', 'working_directory'
+    book = EPUB::Parser.parse 'book.epub', 'working/directory'
     book.each_page_on_spine do |page|
       # do somethong...
     end
@@ -23,6 +34,15 @@ REQUIREMENTS
 * UNIX like `unzip` command
 * libxml2 and libxslt for Nokogiri gem
 
+TODOS
+-----
+* Adding tests
+* Implementing metadata/link, navigation document and so on
+* Using zip library instead of `unzip` command, which has security issue
+* Using SAX on parsing
+* Extracting and organizing common behavior from some classes to modules
+* Abstraction of XML parser(making it possible to use REXML, standard bundled XML library of Ruby)
+
 LICENSE
 -------
 This library is distribuetd under the term of the MIT License.

data/bin/epubinfo

@@ -6,22 +6,29 @@ Version = EPUB::Parser::VERSION
 
 options = {:format => :line}
 opt = OptionParser.new do |opt|
-  opt.banner = 'Usage: epubinfo [options] EPUBFILE'
-  opt.on '-f', '--format=FORMAT', [:line, :json, :yaml], 'Format of output(line, json or yaml), defaults to line(for console)' do |format|
+  opt.banner = <<EOB
+Show metadata of an EPUB file
+
+Usage: epubinfo [options] EPUBFILE
+
+EOB
+  opt.on '-f', '--format=FORMAT', [:line, :json, :yaml], 'format of output(line, json or yaml), defaults to line(for console)' do |format|
     options[:format] = format
   end
 end
 opt.parse!(ARGV)
 
-file = ARGV[0]
+file = ARGV.shift
 unless file
+  $stdout.puts "error: you must supply an EPUB file name"
   $stdout.puts opt.help
-  exit 1
+  exit!
 end
 
 Dir.mktmpdir('epub-parser') do |dir|
   book = EPUB::Parser.parse file, dir
-  data = book.package.metadata.to_hash
+  data = {'Title' => [book.title]}
+  data.merge!(book.package.metadata.to_hash)
   data['Unique identifier'] = [book.package.metadata.unique_identifier]
   data['EPUB Version'] = [book.package.version]
   if options[:format] == :line

data/epub-parser.gemspec

@@ -28,6 +28,8 @@ Gem::Specification.new do |s|
   s.add_development_dependency 'simplecov'
   s.add_development_dependency 'thin'
   s.add_development_dependency 'yard'
+  s.add_development_dependency 'gem-man'
+  s.add_development_dependency 'ronn'
   s.add_development_dependency 'epzip'
   s.add_development_dependency 'epubcheck'
   s.add_development_dependency 'epub_validator'

data/lib/epub.rb

@@ -19,6 +19,12 @@ module EPUB
   end
   module_function :parse
 
+  %w[ title main_title subtitle short_title collection_title edition_title extended_title ].each do |met|
+    define_method met do
+      @package.metadata.__send__(met)
+    end
+  end
+
   def each_page_on_spine(&blk)
     enum = @package.spine.items
     if block_given?

data/lib/epub/constants.rb

@@ -1,13 +1,13 @@
 module EPUB
   module Constants
     NAMESPACES = {
-      'dc'        => 'http://purl.org/dc/elements/1.1/',
-      'container' => 'urn:oasis:names:tc:opendocument:xmlns:container',
-      'opf'       => 'http://www.idpf.org/2007/opf',
-      'xhtml'     => 'http://www.w3.org/1999/xhtml',
-      'epub'      => 'http://www.idpf.org/2007/ops',
-      'm'         => 'http://www.w3.org/1998/Math/MathML',
-      'svg'       => 'http://www.w3.org/2000/svg'
+      'dc'    => 'http://purl.org/dc/elements/1.1/',
+      'ocf'   => 'urn:oasis:names:tc:opendocument:xmlns:container',
+      'opf'   => 'http://www.idpf.org/2007/opf',
+      'xhtml' => 'http://www.w3.org/1999/xhtml',
+      'epub'  => 'http://www.idpf.org/2007/ops',
+      'm'     => 'http://www.w3.org/1998/Math/MathML',
+      'svg'   => 'http://www.w3.org/2000/svg'
     }
 
     module Type

data/lib/epub/parser.rb

@@ -25,7 +25,7 @@ module EPUB
       @book = create_book options
 
       unzip_cmd = options['unzip-command'] || 'unzip'
-      unzip_cmd << " #{@filepath.to_s.shellescape} -d #{@dir.to_s.shellescape}"
+      unzip_cmd << " #{@filepath.to_s.shellescape} -d #{@dir.to_s.shellescape} 1>/dev/null"
       system unzip_cmd
     end
 

data/lib/epub/parser/ocf.rb

@@ -30,7 +30,7 @@ module EPUB
         container = EPUB::OCF::Container.new
         doc = Nokogiri.XML open(File.join @dir, DIRECTORY, CONTAINER_FILE)
 
-        doc.xpath('/container:container/container:rootfiles/container:rootfile', EPUB::NAMESPACES).each do |elem|
+        doc.xpath('/ocf:container/ocf:rootfiles/ocf:rootfile', EPUB::NAMESPACES).each do |elem|
           rootfile = EPUB::OCF::Container::Rootfile.new
           %w[full-path media-type].each do |attr|
             rootfile.__send__(attr.gsub(/-/, '_') + '=', elem[attr])

data/lib/epub/parser/publication.rb

@@ -32,7 +32,7 @@ module EPUB
       def parse_package
         elem = @doc.root
         @package.version = elem['version']
-        @package.unique_identifier = elem['unique-identifier']
+        @package.unique_identifier_id = elem['unique-identifier']
 
         @package
       end
@@ -40,29 +40,84 @@ module EPUB
       def parse_metadata
         metadata = @package.metadata = EPUB::Publication::Package::Metadata.new
         elem = @doc.xpath('/opf:package/opf:metadata', EPUB::NAMESPACES).first
+        id_map = {}
 
         metadata.identifiers = elem.xpath('./dc:identifier', EPUB::NAMESPACES).collect do |e|
           identifier = EPUB::Publication::Package::Metadata::Identifier.new
           identifier.content = e.content
           identifier.id = id = e['id']
-          metadata.unique_identifier = identifier if id == @package.unique_identifier
+          metadata.unique_identifier = identifier if id == @package.unique_identifier_id
+
           identifier
         end
+        metadata.identifiers.each {|i| id_map[i.id] = {metadata: i} if i.respond_to?(:id) && i.id}
 
-        metadata.titles = collect_dcmes(elem, './dc:title') do |title, e|
+        metadata.titles = elem.xpath('./dc:title', EPUB::NAMESPACES).collect do |e|
+          title = EPUB::Publication::Package::Metadata::Title.new
+          %w[ id lang dir ].each do |attr|
+            title.__send__("#{attr}=", e[attr])
+          end
           title.content = e.content
+
+          title
         end
+        # metadata.titles = collect_dcmes(elem, './dc:title') do |title, e|
+        #   title.content = e.content
+        # end
+        metadata.titles.each {|t| id_map[t.id] = {metadata: t} if t.respond_to?(:id) && t.id}
 
         metadata.languages = elem.xpath('./dc:language', EPUB::NAMESPACES).collect do |e|
           e.content
         end
+        metadata.languages.each {|l| id_map[l.id] = {metadata: l} if l.respond_to?(:id) && l.id}
 
         %w[ contributor coverage creator date description format publisher relation source subject type ].each do |dcmes|
           metadata.__send__ "#{dcmes}s=", collect_dcmes(elem, "./dc:#{dcmes}")
+          metadata.__send__("#{dcmes}s").each {|d| id_map[d.id] = {metadata: d} if d.respond_to?(:id) && d.id}
         end
+
         metadata.rights = collect_dcmes(elem, './dc:rights')
+        metadata.rights.each {|r| id_map[r.id] = {metadata: r} if r.respond_to?(:id) && r.id}
+
+        metadata.metas = elem.xpath('./opf:meta', EPUB::NAMESPACES).collect do |e|
+          meta = EPUB::Publication::Package::Metadata::Meta.new
+          %w[ property id scheme ].each { |attr| meta.__send__("#{attr}=", e[attr]) }
+          meta.content = e.content
+          if (refines = e['refines']) && refines[0] == '#'
+            id = refines[1..-1]
+            id_map[id] ||= {}
+            id_map[id][:refiners] ||= []
+            id_map[id][:refiners] << meta
+          end
+
+          meta
+        end
+        metadata.metas.each {|m| id_map[m.id] = {metadata: m} if m.respond_to?(:id) && m.id}
+
+        metadata.links = elem.xpath('./opf:link', EPUB::NAMESPACES).collect do |e|
+          link = EPUB::Publication::Package::Metadata::Link.new
+          %w[ href id media-type ].each do |attr|
+            link.__send__(attr.gsub(/-/, '_') + '=', e[attr])
+          end
+          link.iri = Addressable::URI.parse(e['href'])
+          link.rel = e['rel'].strip.split
+          if (refines = e['refines']) && refines[0] == '#'
+            id = refines[1..-1]
+            id_map[id] ||= {}
+            id_map[id][:refiners] ||= []
+            id_map[id][:refiners] << link
+          end
 
-        # To do: handle <meta> elements
+          link
+        end
+        metadata.links.each {|l| id_map[l.id] = {metadata: l} if l.respond_to?(:id) && l.id}
+
+        id_map.values.each do |hsh|
+          next unless hsh[:refiners]
+          next unless hsh[:metadata]
+          hsh[:metadata].refiners = hsh[:refiners]
+          hsh[:refiners].each {|meta| meta.refines = hsh[:metadata]}
+        end
 
         metadata
       end

data/lib/epub/parser/version.rb

@@ -1,5 +1,5 @@
 module EPUB
   class Parser
-    VERSION = "0.0.7"
+    VERSION = "0.0.8"
   end
 end

data/lib/epub/publication/package.rb

@@ -4,8 +4,8 @@ module EPUB
   module Publication
     class Package
       attr_accessor :book,
-                    :version, :unique_identifier, :prefix, :xml_lang, :dir, :id
-      attr_reader :metadata, :manifest, :spine, :guide
+                    :version, :unique_identifier_id, :prefix, :xml_lang, :dir, :id
+      attr_reader :metadata, :manifest, :spine, :guide, :unique_identifier
       alias lang  xml_lang
       alias lang= xml_lang=
 
@@ -28,6 +28,10 @@ module EPUB
         guide.package = self
         @guide = guide
       end
+
+      def unique_identifier
+        @metadata.unique_identifier
+      end
     end
   end
 end

data/lib/epub/publication/package/metadata.rb

@@ -2,25 +2,65 @@ module EPUB
   module Publication
     class Package
       class Metadata
-        ELEMS = [:identifiers, :titles, :languages] +
-                [:contributors, :coverages, :creators, :dates, :descriptions, :formats, :publishers,
-                 :relations, :rights, :sources, :subjects, :types]
-        attr_accessor :package, :unique_identifier,
-                      *(ELEMS.collect {|elem| "dc_#{elem}"})
-        ELEMS.each do |elem|
+        DC_ELEMS = [:identifiers, :titles, :languages] +
+                   [:contributors, :coverages, :creators, :dates, :descriptions, :formats, :publishers,
+                    :relations, :rights, :sources, :subjects, :types]
+        attr_accessor :package, :unique_identifier, :metas, :links,
+                      *(DC_ELEMS.collect {|elem| "dc_#{elem}"})
+        DC_ELEMS.each do |elem|
           alias_method elem, "dc_#{elem}"
           alias_method "#{elem}=", "dc_#{elem}="
         end
-        attr_accessor :links
+
+        def title
+          return extended_title unless extended_title.empty?
+          compositted = titles.select {|title| title.display_seq}.sort.join("\n")
+          return compositted unless compositted.empty?
+          return main_title unless main_title.empty?
+          titles.sort.join("\n")
+        end
+
+        %w[ main short collection edition extended ].each do |type|
+          define_method "#{type}_title" do
+            titles.select {|title| title.title_type.to_s == type}.sort.join(' ')
+          end
+        end
+        def subtitle
+          titles.select {|title| title.title_type.to_s == 'subtitle'}.sort.join(' ')
+        end
 
         def to_hash
-          ELEMS.inject({}) do |hsh, elem|
+          DC_ELEMS.inject({}) do |hsh, elem|
             hsh[elem] = __send__(elem)
             hsh
           end
         end
 
+        def primary_metas
+          metas.select {|meta| meta.primary_expression?}
+        end
+
+        module Refinable
+          PROPERTIES = %w[ alternate-script display-seq file-as group-position identifier-type meta-auth role title-type ]
+
+          PROPERTIES.each do |voc|
+            met = voc.gsub(/-/, '_')
+            attr_accessor met
+            define_method met do
+              @refiners.select {|refiner| refiner.property == voc}.first
+            end
+          end
+
+          attr_accessor :refiners
+
+          def initialize
+            @refiners = []
+          end
+        end
+
         class Identifier
+          include Refinable
+
           attr_accessor :content, :id
 
           def to_s
@@ -29,20 +69,58 @@ module EPUB
         end
 
         class Title
+          include Refinable
+          include Comparable
+
           attr_accessor :content, :id, :lang, :dir
 
+          def <=>(other)
+            return 1 if other.display_seq.nil?
+            return -1 if display_seq.nil?
+            display_seq.to_s.to_i <=> other.display_seq.to_s.to_i
+          end
+
           def to_s
             content
           end
         end
 
         class DCMES
+          include Refinable
+
           attr_accessor :content, :id, :lang, :dir
 
           def to_s
             content
           end
         end
+
+        class Meta
+          include Refinable
+
+          attr_accessor :property, :refines, :id, :scheme, :content
+
+          def refines?
+            ! refines.nil?
+          end
+
+          alias subexpression? refines?
+
+          def primary_expression?
+            ! subexpression?
+          end
+
+          def to_s
+            content
+          end
+        end
+
+        class Link
+          include Refinable
+
+          attr_accessor :href, :rel, :id, :refines, :media_type,
+                        :iri
+        end
       end
     end
   end

data/man/epubinfo.1.ronn

@@ -0,0 +1,19 @@
+epubinfo(1) -- show metadata of an EPUB book
+============================================
+
+SYNOPSIS
+--------
+
+`epubinfo` <filename><br>
+`epubinfo` `-f`|`--format=<formnt>` <filename>
+
+DESCRIPTION
+-----------
+
+**epubinfo** shows metadata of an EPUB book.
+
+OPTIONS
+-------
+
+ * `-f`, `--format`:
+   Output format; currently "line", "json" and "yaml" are acceptable

data/test/fixtures/book/OPS//343/203/253/343/203/274/343/203/210/343/203/225/343/202/241/343/202/244/343/203/253.opf

@@ -1,7 +1,7 @@
 <?xml version="1.0"?>
 <package version="3.0"
-	 unique-identifier="pub-id"
-	 xmlns="http://www.idpf.org/2007/opf">
+         unique-identifier="pub-id"
+         xmlns="http://www.idpf.org/2007/opf">
   <metadata xmlns:dc="http://purl.org/dc/elements/1.1/">
     <dc:identifier id="pub-id">da265185-8da8-462d-a146-17dd388f61fc</dc:identifier>
 
@@ -38,6 +38,11 @@
     <dc:creator id="creator02">John Tenniel</dc:creator>
     <meta refines="#creator02" property="role" scheme="marc:relators">ill</meta>
     <meta refines="#creator02" property="display-seq">2</meta>
+
+
+    <link rel="onix-record" href="http://example.org/onix/12389347" refines="#pub-id"/>
+    <link rel="xmp-record" href="http://example.org/xmp/12389347"/>
+    <link rel="foaf:homepage" href="http://example.org/book-info/12389347" />
   </metadata>
   <manifest id="manifest-id">
     <item id="nav"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: epub-parser
 version: !ruby/object:Gem::Version
-  version: 0.0.7
+  version: 0.0.8
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-02-21 00:00:00.000000000Z
+date: 2012-03-11 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rubygems-test
-  requirement: &12346600 !ruby/object:Gem::Requirement
+  requirement: &11201540 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12346600
+  version_requirements: *11201540
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &12342420 !ruby/object:Gem::Requirement
+  requirement: &11189400 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12342420
+  version_requirements: *11189400
 - !ruby/object:Gem::Dependency
   name: pry
-  requirement: &12360020 !ruby/object:Gem::Requirement
+  requirement: &11186320 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12360020
+  version_requirements: *11186320
 - !ruby/object:Gem::Dependency
   name: pry-doc
-  requirement: &12359300 !ruby/object:Gem::Requirement
+  requirement: &11185640 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -54,10 +54,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12359300
+  version_requirements: *11185640
 - !ruby/object:Gem::Dependency
   name: test-unit
-  requirement: &12358560 !ruby/object:Gem::Requirement
+  requirement: &11184560 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -65,10 +65,10 @@ dependencies:
         version: '2'
   type: :development
   prerelease: false
-  version_requirements: *12358560
+  version_requirements: *11184560
 - !ruby/object:Gem::Dependency
   name: simplecov
-  requirement: &12358000 !ruby/object:Gem::Requirement
+  requirement: &11183620 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -76,10 +76,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12358000
+  version_requirements: *11183620
 - !ruby/object:Gem::Dependency
   name: thin
-  requirement: &12357120 !ruby/object:Gem::Requirement
+  requirement: &11182540 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -87,10 +87,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12357120
+  version_requirements: *11182540
 - !ruby/object:Gem::Dependency
   name: yard
-  requirement: &12356360 !ruby/object:Gem::Requirement
+  requirement: &11181840 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -98,10 +98,32 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12356360
+  version_requirements: *11181840
+- !ruby/object:Gem::Dependency
+  name: gem-man
+  requirement: &11122520 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: *11122520
+- !ruby/object:Gem::Dependency
+  name: ronn
+  requirement: &11120840 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: *11120840
 - !ruby/object:Gem::Dependency
   name: epzip
-  requirement: &12355640 !ruby/object:Gem::Requirement
+  requirement: &11119120 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -109,10 +131,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12355640
+  version_requirements: *11119120
 - !ruby/object:Gem::Dependency
   name: epubcheck
-  requirement: &12354540 !ruby/object:Gem::Requirement
+  requirement: &12151080 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -120,10 +142,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12354540
+  version_requirements: *12151080
 - !ruby/object:Gem::Dependency
   name: epub_validator
-  requirement: &12353020 !ruby/object:Gem::Requirement
+  requirement: &12148200 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -131,10 +153,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12353020
+  version_requirements: *12148200
 - !ruby/object:Gem::Dependency
   name: zipruby
-  requirement: &12370280 !ruby/object:Gem::Requirement
+  requirement: &11234320 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -142,10 +164,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12370280
+  version_requirements: *11234320
 - !ruby/object:Gem::Dependency
   name: rubyzip
-  requirement: &12369680 !ruby/object:Gem::Requirement
+  requirement: &11232240 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -153,10 +175,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *12369680
+  version_requirements: *11232240
 - !ruby/object:Gem::Dependency
   name: enumerabler
-  requirement: &12368540 !ruby/object:Gem::Requirement
+  requirement: &11227260 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -164,10 +186,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *12368540
+  version_requirements: *11227260
 - !ruby/object:Gem::Dependency
   name: nokogiri
-  requirement: &12367280 !ruby/object:Gem::Requirement
+  requirement: &11365900 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -175,10 +197,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *12367280
+  version_requirements: *11365900
 - !ruby/object:Gem::Dependency
   name: addressable
-  requirement: &12365820 !ruby/object:Gem::Requirement
+  requirement: &11363620 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -186,7 +208,7 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *12365820
+  version_requirements: *11363620
 description: Parse EPUB 3 book loosely
 email:
 - KitaitiMakoto@gmail.com
@@ -227,6 +249,7 @@ files:
 - lib/epub/publication/package/manifest.rb
 - lib/epub/publication/package/metadata.rb
 - lib/epub/publication/package/spine.rb
+- man/epubinfo.1.ronn
 - schemas/epub-nav-30.rnc
 - schemas/epub-nav-30.sch
 - schemas/epub-xhtml-30.sch
@@ -256,7 +279,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -4537999334465122186
+      hash: -1773548056502971517
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -265,7 +288,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -4537999334465122186
+      hash: -1773548056502971517
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.8