epub-parser 0.0.1

data/lib/epub/parser/version.rb

@@ -0,0 +1,5 @@
+module EPUB
+  class Parser
+    VERSION = "0.0.1"
+  end
+end

data/lib/epub/publication.rb

@@ -0,0 +1,6 @@
+require 'epub/publication/package'
+
+module EPUB
+  module Publication
+  end
+end

data/lib/epub/publication/package.rb

@@ -0,0 +1,25 @@
+%w[ metadata manifest spine guide ].each { |f| require "epub/publication/package/#{f}" }
+
+module EPUB
+  module Publication
+    class Package
+      attr_accessor :version, :unique_identifier, :prefix, :xml_lang, :dir, :id
+      attr_reader :metadata, :manifest, :spine
+      alias lang  xml_lang
+      alias lang= xml_lang=
+
+      def metadata=(metadata)
+      end
+
+      def manifest=(manifest)
+        manifest.package = self
+        @manifest = manifest
+      end
+
+      def spine=(spine)
+        spine.package = self
+        @spine = spine
+      end
+    end
+  end
+end

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

@@ -0,0 +1,18 @@
+module EPUB
+  module Publication
+    class Package
+      class Bindings
+        attr_accessor :media_types
+
+        def <<(media_type)
+          @media_types ||= []
+          @media_types << media_type
+        end
+
+        class MediaType
+          attr_accessor :media_type, :handler
+        end
+      end
+    end
+  end
+end

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

@@ -0,0 +1,11 @@
+# This class is deprecated
+# See OPF2 for specification
+
+module EPUB
+  module Publication
+    class Package
+      class Guide
+      end
+    end
+  end
+end

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

@@ -0,0 +1,44 @@
+require 'enumerabler'
+
+module EPUB
+  module Publication
+    class Package
+      class Manifest
+        attr_accessor :package,
+                      :id
+        attr_reader :items
+
+        # syntax sugar for #items.<<
+        def <<(item)
+          @items ||= []
+          item.manifest = self
+          @items << item
+        end
+
+        # syntax sugar
+        def navs
+          items.selector {|i| i.properties.include? 'nav'}
+        end
+
+        def nav
+          navs.first
+        end
+
+        def [](item_id)
+          items.selector {|item| item.id == item_id}.first
+        end
+
+        class Item
+          attr_accessor :manifest,
+                        :id, :href, :media_type, :fallback, :properties, :media_overlay
+
+          alias path href
+
+          def to_s
+            href.to_s
+          end
+        end
+      end
+    end
+  end
+end

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

@@ -0,0 +1,17 @@
+module EPUB
+  module Publication
+    class Package
+      class Metadata
+        elems = [:identifiers, :titles, :languages] +
+                [:contributers, :coverages, :creators, :dates, :descriptions, :formats, :publishers,
+                 :relations, :rigths, :sources, :subjects, :types]
+        attr_accessor *(elems.collect {|elem| "dc_#{elem}"})
+        elems.each do |elem|
+          alias_method elem, "dc_#{elem}"
+          alias_method "#{elem}=", "dc_#{elem}="
+        end
+        attr_accessor :metas, :links
+      end
+    end
+  end
+end

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

@@ -0,0 +1,46 @@
+module EPUB
+  module Publication
+    class Package
+      class Spine
+        attr_accessor :package,
+                      :id, :toc, :page_progression_direction
+        attr_reader :itemrefs
+
+        def <<(itemref)
+          @itemrefs ||= []
+          itemref.spine = self
+          @itemrefs << itemref
+        end
+
+        # @yield [itemref]
+        # @yieldparam [Itemref] itemref
+        # @yieldreturn [Object] returns the last value of block
+        # @return [Object, Enumerator]
+        #   returns the last value of block when block given, Enumerator when not
+        def each_itemref
+          if block_given?
+            itemrefs.each {|itemref| yield itemref}
+          else
+            enum_for :each_itemref
+          end
+        end
+
+        # @return [Enumerator] Enumerator which yeilds {Manifest::Item}
+        #   referred by each of {#itemrefs}
+        def items
+          itemrefs.collector {|itemref| itemref.item}
+        end
+
+        class Itemref
+          attr_accessor :spine,
+                        :idref, :linear, :id, :properties
+
+          # @return [Package::Manifest::Item] item referred by this object
+          def item
+            @item ||= @spine.package.manifest[idref]
+          end
+        end
+      end
+    end
+  end
+end

data/lib/epub/type.rb

@@ -0,0 +1,7 @@
+module EPUB
+  module Type
+    TOC = :toc
+    PAGE_LIST = :page_list
+    LANDMARKS = :landmarks
+  end
+end

data/schemas/epub-nav-30.rnc

@@ -0,0 +1,10 @@
+
+
+   default namespace = "http://www.w3.org/1999/xhtml"
+   namespace epub = "http://www.idpf.org/2007/ops"
+      
+   include "epub-xhtml-30.rnc" {   
+        html5.nav.content = html5.headings.class?, html5.ol     
+        html5.oli.content = html5.a.phrasing | ((html5.a.phrasing | html5.span), html5.ol)
+   }
+   

data/schemas/epub-nav-30.sch

@@ -0,0 +1,72 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<schema xmlns="http://purl.oclc.org/dsdl/schematron">
+
+    <ns uri="http://www.w3.org/1999/xhtml" prefix="html"/>
+    <ns uri="http://www.idpf.org/2007/ops" prefix="epub"/>
+
+    <pattern id="nav-ocurrence">
+        <rule context="html:body">
+            <assert test="count(.//html:nav[@epub:type='toc']) = 1">Exactly one 'toc' nav element
+                must be present</assert>
+            <assert test="count(.//html:nav[@epub:type='page-list']) &lt; 2">Multiple occurrences of
+                the 'page-list' nav element</assert>
+            <assert test="count(.//html:nav[@epub:type='landmarks']) &lt; 2">Multiple occurrences of
+                the 'landmarks' nav element</assert>
+        </rule>
+    </pattern>
+
+    <pattern id="span-no-sublist">
+        <rule context="html:body//html:nav//html:span">
+            <assert test="count(.//ol) = 0"> The span element must only be used as heading for flat
+                sublists (not hierarchical navigation structures) </assert>
+        </rule>
+    </pattern>
+
+    <pattern id="landmarks">
+        <rule context="html:nav[@epub:type='landmarks']//html:ol//html:a">
+            <assert test="@epub:type">Missing epub:type attribute on anchor inside 'landmarks' nav
+                element</assert>
+        </rule>
+    </pattern>
+
+    <pattern id="link-labels">
+        <rule context="html:nav//html:ol//html:a">
+            <assert test="string-length(normalize-space(string(.))) > 0">Anchors within nav elements
+                must contain text</assert>
+        </rule>
+    </pattern>
+
+    <pattern id="span-labels">
+        <rule context="html:nav//html:ol//html:span">
+            <assert test="string-length(normalize-space(string(.))) > 0">Spans within nav elements
+                must contain text</assert>
+        </rule>
+    </pattern>
+
+    <pattern id="req-heading">
+        <rule
+            context="html:nav[not(@epub:type = 'toc') and not (@epub:type = 'page-list') and not (@epub:type = 'landmarks')]">
+            <let name="fc" value="local-name(./*[1])"/>
+            <assert test="(starts-with($fc,'h') and string-length($fc) = 2) or ($fc = 'hgroup')">nav
+                elements other than 'toc', 'page-list' and 'landmarks' must contain a heading as the
+                first child</assert>
+        </rule>
+    </pattern>
+
+    <pattern id="heading-content">
+        <rule context="html:h1|html:h2|html:h3|html:h4|html:h5|html:h6|html:hgroup">
+            <assert test="string-length(normalize-space(string(.))) > 0">Heading elements must
+                contain text</assert>
+        </rule>
+    </pattern>
+
+
+    <!-- warnings mode <pattern id="page-list-flat">
+        <rule context="html:body//html:nav[@epub:type='page-list']">
+        <assert test="count(.//html:ol) = 1">The page-list navigation structure should be a
+        list, not a nested hierarchy</assert>
+        </rule>
+        </pattern> 
+    -->
+
+</schema>

data/schemas/epub-xhtml-30.sch

@@ -0,0 +1,377 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<schema xmlns="http://purl.oclc.org/dsdl/schematron" queryBinding="xslt2">
+                        
+    <ns uri="http://www.w3.org/1999/xhtml" prefix="h"/>
+    <ns uri="http://www.idpf.org/2007/ops" prefix="epub"/>
+    <ns uri="http://www.w3.org/1998/Math/MathML" prefix="math"/>
+    <ns uri="http://www.w3.org/2001/10/synthesis" prefix="ssml"/>
+    <ns uri="http://www.w3.org/2001/xml-events" prefix="ev"/>
+    <ns uri="http://www.w3.org/2000/svg" prefix="svg" />
+    
+    <let name="id-set" value="//*[@id]"/>
+    
+    <pattern id="ancestor-area-map" is-a="required-ancestor">        
+        <param name="descendant" value="h:area"/>
+        <param name="ancestor" value="h:map"/>        
+    </pattern>
+
+    <pattern id="ancestor-imgismap-ahref" is-a="required-ancestor">        
+        <param name="descendant" value="h:img[@ismap]"/>
+        <param name="ancestor" value="h:a[@href]"/>        
+    </pattern>
+    
+    <pattern id="descendant-a-interactive" is-a="no-interactive-content-descendants">        
+        <param name="ancestor" value="h:a"/>       
+    </pattern>
+    
+    <pattern id="descendant-button-interactive" is-a="no-interactive-content-descendants">        
+        <param name="ancestor" value="h:button"/>       
+    </pattern>
+    
+    <pattern id="descendant-audio-audio" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:audio"/>
+        <param name="descendant" value="h:audio"/>        
+    </pattern>
+    
+    <pattern id="descendant-audio-video" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:audio"/>
+        <param name="descendant" value="h:video"/>        
+    </pattern>
+    
+    <pattern id="descendant-video-video" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:video"/>
+        <param name="descendant" value="h:video"/>        
+    </pattern>
+    
+    <pattern id="descendant-video-audio" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:video"/>
+        <param name="descendant" value="h:audio"/>        
+    </pattern>
+    
+    <pattern id="descendant-address-address" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:address"/>
+        <param name="descendant" value="h:address"/>        
+    </pattern>
+    
+    <pattern id="descendant-address-header" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:address"/>
+        <param name="descendant" value="h:header"/>        
+    </pattern>
+    
+    <pattern id="descendant-address-footer" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:address"/>
+        <param name="descendant" value="h:footer"/>        
+    </pattern>
+
+    <pattern id="descendant-form-form" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:form"/>
+        <param name="descendant" value="h:form"/>        
+    </pattern>
+
+    <pattern id="descendant-progress-progress" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:progress"/>
+        <param name="descendant" value="h:progress"/>        
+    </pattern>
+
+    <pattern id="descendant-meter-meter" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:meter"/>
+        <param name="descendant" value="h:meter"/>        
+    </pattern>
+
+    <pattern id="descendant-dfn-dfn" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:dfn"/>
+        <param name="descendant" value="h:dfn"/>        
+    </pattern>
+
+    <pattern id="descendant-time-time" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:time"/>
+        <param name="descendant" value="h:time"/>        
+    </pattern>
+
+    <pattern id="descendant-caption-table" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:caption"/>
+        <param name="descendant" value="h:table"/>        
+    </pattern>
+    
+    <pattern id="descendant-header-header" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:header"/>
+        <param name="descendant" value="h:header"/>        
+    </pattern>
+    
+    <pattern id="descendant-header-footer" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:header"/>
+        <param name="descendant" value="h:footer"/>        
+    </pattern>
+    
+    <pattern id="descendant-footer-footer" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:footer"/>
+        <param name="descendant" value="h:footer"/>        
+    </pattern>
+    
+    <pattern id="descendant-footer-header" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:footer"/>
+        <param name="descendant" value="h:header"/>        
+    </pattern>
+    
+    <pattern id="descendant-label-label" is-a="disallowed-descendants">        
+        <param name="ancestor" value="h:label"/>
+        <param name="descendant" value="h:label"/>        
+    </pattern>
+    
+    <pattern id="descendant-annotation-xml-math" is-a="disallowed-descendants">        
+        <param name="ancestor" value="math:annotation-xml[@encoding='application/xhtml+xml' and @name='alternate-representation']"/>
+        <param name="descendant" value="math:*"/>        
+    </pattern>
+    
+    <pattern id="descendant-svgtitle-svg" is-a="disallowed-descendants">        
+        <param name="ancestor" value="svg:title"/>
+        <param name="descendant" value="svg:*"/>        
+    </pattern>
+    
+    <pattern id="bdo-dir" is-a="required-attr">
+        <param name="elem" value="h:bdo"/>
+        <param name="attr" value="dir"/>
+    </pattern>
+
+    <pattern id="idrefs-aria-describedby" is-a="idrefs-any">
+        <param name="element" value="*"/>
+        <param name="idrefs-attr-name" value="aria-describedby"/>
+    </pattern>
+
+    <pattern id="idrefs-output-for" is-a="idrefs-any">
+        <param name="element" value="h:output"/>
+        <param name="idrefs-attr-name" value="for"/>
+    </pattern>
+
+    <pattern id="idrefs-aria-flowto" is-a="idrefs-any">
+        <param name="element" value="*"/>
+        <param name="idrefs-attr-name" value="aria-flowto"/>
+    </pattern>
+
+    <pattern id="idrefs-aria-labelledby" is-a="idrefs-any">
+        <param name="element" value="*"/>
+        <param name="idrefs-attr-name" value="aria-labelledby"/>
+    </pattern>
+
+    <pattern id="idrefs-aria-owns" is-a="idrefs-any">
+        <param name="element" value="*"/>
+        <param name="idrefs-attr-name" value="aria-owns"/>
+    </pattern>
+
+    <pattern id="idrefs-aria-controls" is-a="idrefs-any">
+        <param name="element" value="*"/>
+        <param name="idrefs-attr-name" value="aria-controls"/>
+    </pattern>
+    
+    <pattern id="idref-mathml-xref" is-a="idref-any">
+        <param name="element" value="math:*"/>
+        <param name="idref-attr-name" value="xref"/>
+    </pattern>
+
+    <pattern id="idref-mathml-indenttarget" is-a="idref-any">
+        <param name="element" value="math:*"/>
+        <param name="idref-attr-name" value="indenttarget"/>
+    </pattern>
+
+    <pattern id="idref-contextmenu" is-a="idref-named">
+        <param name="element" value="h:*"/>
+        <param name="idref-attr-name" value="contextmenu"/>
+        <param name="target-name" value="h:menu"/>
+    </pattern>
+    
+    <pattern id="idref-input-list" is-a="idref-named">
+        <param name="element" value="h:input"/>
+        <param name="idref-attr-name" value="list"/>
+        <param name="target-name" value="h:datalist"/>
+    </pattern>
+    
+    <pattern id="idref-forms-form" is-a="idref-named">
+        <param name="element" value="h:*"/>
+        <param name="idref-attr-name" value="form"/>
+        <param name="target-name" value="h:form"/>
+    </pattern>
+
+    <pattern id="idref-aria-activedescendant">
+        <rule context="*[@aria-activedescendant]">
+            <assert test="descendant::*[@id = current()/@aria-activedescendant]"
+                >The aria-activedescendant attribute must refer to a descendant element.</assert>
+        </rule>
+    </pattern>
+
+    <pattern id="idref-label-for">        
+        <rule context="h:label[@for]">
+            <assert
+                test="some $elem in $id-set satisfies $elem/@id eq current()/@for and 
+                   (local-name($elem) eq 'button' 
+                 or (local-name($elem) eq 'input' and not($elem/@type='hidden'))
+                 or local-name($elem) eq 'keygen' 
+                 or local-name($elem) eq 'meter'
+                 or local-name($elem) eq 'output' 
+                 or local-name($elem) eq 'progress' 
+                 or local-name($elem) eq 'select' 
+                 or local-name($elem) eq 'textarea')"
+                >The for attribute does not refer to an allowed target element (expecting: button|keygen|meter|output|progress|select|textarea|input[not(@type='hidden')]).</assert>
+        </rule>
+    </pattern>
+            
+    <pattern id="idrefs-headers">        
+        <rule context="h:*[@headers]">    
+            <let name="table" value="ancestor::h:table"/>
+            <assert test="every $idref in tokenize(@headers, '\s+') satisfies (some $elem in $table//h:th satisfies ($elem/@id eq $idref))"
+                >The headers attribute must refer to th elements in the same table.</assert>            
+        </rule>
+    </pattern> 
+    
+    <pattern id="idref-trigger-observer" is-a="idref-any">
+        <param name="element" value="epub:trigger"/>
+        <param name="idref-attr-name" value="ev:observer"/>
+    </pattern>
+    
+    <pattern id="idref-trigger-ref" is-a="idref-any">
+        <param name="element" value="epub:trigger"/>
+        <param name="idref-attr-name" value="ref"/>
+    </pattern>
+    
+    <pattern id="map.name" >
+        <rule context='h:map[@name]'>   
+            <let name="name-set" value="//h:map[@name]"/>
+            <assert test="count($name-set[@name = current()/@name]) = 1">Duplicate map name '<value-of
+                select="current()/@name"/>'</assert>
+        </rule>
+    </pattern>
+    
+    <pattern id="map.id" >
+        <rule context='h:map[@id and @name]'>
+            <assert test='@id = @name'
+                >The id attribute on the map element must have the same value as the name attribute.</assert>
+        </rule>        
+    </pattern>
+    
+    <pattern id='lang-xmllang'>
+        <rule context='h:*[@lang and @xml:lang]'>
+            <assert test="lower-case(@xml:lang) = lower-case(@lang)"
+                >The lang and xml:lang attributes must have the same value.</assert>                
+        </rule>
+    </pattern>
+    
+    <pattern id="id-unique">
+        <rule context="*[@id]">
+            <assert test="count($id-set[@id = current()/@id]) = 1">Duplicate ID '<value-of
+                select="current()/@id"/>'</assert>
+        </rule>
+    </pattern>
+    
+    <pattern id="select-multiple">
+        <rule context="h:select[not(@multiple)]">
+            <report test='count(descendant::h:option[@selected]) > 1'
+                >A select element whose multiple attribute is not specified must not have more than one descendant option element with its selected attribute set.</report>
+        </rule>
+    </pattern> 
+    
+    <pattern id="track">                
+        <rule context="h:track">
+            <report test="@label and normalize-space(@label) = ''"
+                >The track element label attribute value must not be the empty string.</report>
+            <report test="@default and preceding-sibling::h:track[@default]"
+                >There must not be more than one track child of a media element element with the default attribute specified.</report>            
+        </rule>
+    </pattern>
+    
+    <pattern id="ssml-ph">        
+        <rule context="*[@ssml:ph]">
+            <report test="ancestor::*[@ssml:ph]"
+                >The ssml:ph attribute must not be specified on a descendant of an element that also carries this attribute.</report>
+        </rule>            
+    </pattern>
+            
+    <pattern id="style-scoped">
+        <rule context="h:style[ancestor::h:body]">
+            <assert test="every $elem in preceding-sibling::* satisfies local-name($elem) eq 'style'"
+                >The scoped style element must occur before any other flow content other than other style elements and inter-element whitespace.</assert>
+        </rule>
+    </pattern>
+        
+    <pattern id="link-sizes">
+        <rule context="h:link[@sizes]">
+            <assert test="@rel='icon'"
+                >The sizes attribute must not be specified on link elements that do not have a rel attribute that specifies the icon keyword.</assert>
+        </rule>
+    </pattern>   
+    
+    <pattern id="meta-charset">
+        <rule context="h:meta[@charset]">
+            <assert test="count(preceding-sibling::h:meta[@charset]) = 0"
+                >There must not be more than one meta element with a charset attribute per document.</assert>
+        </rule>
+    </pattern>
+    
+    <pattern id="article-pubdate">
+        <rule context="h:article[h:time]">
+            <assert test="count(./h:time[@pubdate]) &lt; 2"
+                >For each article element, there must be no more than one time element child with a pubdate attribute</assert>
+        </rule>
+    </pattern>
+    
+    <pattern id="document-pubdate">
+        <rule context="h:time[not (ancestor::h:article)]">
+            <assert test="count(//h:time[@pubdate and not (ancestor::h:article)]) &lt; 2"
+                >For each Document, there must be no more than one time element with a pubdate 
+                attribute that does not have an ancestor article element.</assert>
+        </rule>
+    </pattern>
+
+    <pattern abstract="true" id="idref-any">
+        <rule context="$element[@$idref-attr-name]">
+            <assert test="some $elem in $id-set satisfies $elem/@id eq current()/@$idref-attr-name"
+                >The <name path="@$idref-attr-name"/> attribute must refer to an element in the same document (the ID '<value-of 
+                    select="current()/@$idref-attr-name"/>' does not exist).</assert>
+        </rule>
+    </pattern>
+
+    <pattern abstract="true" id="idrefs-any">
+        <rule context="$element[@$idrefs-attr-name]">            
+            <assert test="every $idref in tokenize(@$idrefs-attr-name,'\s+') satisfies (some $elem in $id-set satisfies ($elem/@id eq $idref))"
+                >The <name path="@$idrefs-attr-name"/> attribute must refer to elements in the same document (target ID missing)</assert>
+        </rule>
+    </pattern>
+
+    <pattern abstract="true" id="idref-named">
+        <rule context="$element[@$idref-attr-name]">            
+            <assert test="//$target-name[@id = current()/@$idref-attr-name]">The <name
+                    path="@$idref-attr-name"/> attribute does not refer to an allowed target element (expecting: <value-of 
+                    select="replace('$target-name','h:','')"/>).</assert>
+        </rule>
+    </pattern>
+
+    <pattern abstract="true" id="required-attr">
+        <rule context="$elem">
+            <assert test="@$attr"
+                >The <name/> element must have a <value-of select="'$attr'"/> attribute.</assert>            
+        </rule>
+    </pattern>
+
+    <pattern abstract="true" id="disallowed-descendants">
+        <rule context="$descendant">
+            <report test="ancestor::$ancestor"
+                >The <name/> element must not appear inside <value-of select="local-name(ancestor::$ancestor)"/> elements.</report>            
+        </rule>
+    </pattern>
+    
+    <pattern abstract="true" id="required-ancestor">
+        <rule context='$descendant'>            
+            <assert test='ancestor::$ancestor'
+                >The <value-of select="replace('$descendant','h:','')"/> element must have an ancestor <value-of select="replace('$ancestor','h:','')"/> element.</assert>
+        </rule>
+    </pattern>
+    
+    <pattern abstract="true" id="no-interactive-content-descendants">
+        <rule context="h:a|h:audio[@controls]|h:button|h:details|h:embed|h:iframe|h:img[@usemap]|h:input[not(@type='hidden')]
+            |h:keygen|h:label|h:menu[@type='toolbar']|h:object[@usemap]|h:select|h:textarea|h:video[@controls]">
+            <report test="ancestor::$ancestor"
+                >The <name/> element must not appear inside <value-of select="local-name(ancestor::$ancestor)"/> elements.</report>            
+        </rule>
+    </pattern>
+    
+    <include href="./mod/epub-svg11-re.sch"/>
+    
+</schema>