# -*- coding: utf-8 -*- require 'rubygems' require 'nokogiri' require 'zip' require 'fileutils' # = GEPUB # Author:: KOJIMA Satoshi # namespace for gepub library. # The core class is GEPUB::Book. It holds metadata and contents of EPUB file. metadata and contents can be accessed # through GEPUB::Meta and GEPUB::Item. # For generating EPUB file, use GEPUB::Builder. # GEPUB::Item holds information and data of resources like xhtml text, css, scripts, images, videos, etc. # GEPUB::Meta holds metadata(title, creator, publisher, etc.) with its information (alternate script, display sequence, etc.) module GEPUB # Book is the class to hold data in EPUB files. # # It can generate and parse EPUB2/EPUB3 files. # # If you want to generate a new EPUB file, consider using GEPUB::Builder instead # of using Book directly. # Builder is a wrapper class of Book specialized for generating EPUB. # # Book delegates many methods to objects in other class, so you can't find # them in Book#methods or in ri/rdoc documentation. Their descriptions are below. # # == \Package Attributes # === Book#version (delegated to Package#version) # returns OPF version. # === Book#version=, Book#set_version (delegated to Package#version=) # set OPF version # === Book#unique_identifier (delegated to Package#unique_identifier) # return unique_identifier ID value. identifier itself can be get by Book#identifier # == \Metadata # \Metadata items(e.g. title, creator, publisher, etc) are GEPUB::Meta objects. # === Book#identifier (delegated to Package#identifier) # return GEPUB::Meta object of unique identifier. # === Book#identifier=(identifier) (delegated to Package#identifier=) # set identifier (i.e. url, uuid, ISBN) as unique-identifier of EPUB. # === Book#set_main_id(identifier, id = nil, type = nil) (delegated to Package#set_main_id) # same as identifier=, but can specify id (in the opf xml) and identifier type(i.e. URL, uuid, ISBN, etc) # === Book#add_identifier(string, id, type=nil) (delegated to Metadata#add_identifier) # Set an identifier metadata. It it not unique-identifier in opf. Many EPUB files do not set identifier other than unique-identifier. # === Book#add_title(content, id = nil, title_type = nil) (delegated to Metadata#add_title) # add title metadata. title_type candidates is defined in TITLE_TYPES. # === Book#set_title(content, id = nil, title_type = nil) (delegated to Metadata#set_title) # clear all titles and then add title. # === Book#title (delegated to Metadata) # returns 'main' title Meta object. 'main' title is determined by this order: # 1. title-type is 'main' # 2. display-seq is smallest # 3. appears first in opf file # === Book#title_list (delegated to Metadata) # returns titles list by display-seq or defined order. # the title without display-seq is appear after titles with display-seq. # === Book#add_creator(content, id = nil, role = 'aut') (delegated to Metadata#add_creator) # add creator. # === Book#creator # returns 'main' creator Meta object. 'main' creator is determined as following: # 1. display-seq is smallest # 2. appears first in opf file # === Book#creator_list (delegated to Metadata) # returns creators list by display-seq or defined order. # the creators without display-seq is appear after creators with display-seq. # === Book#add_contributor(content, id = nil, role = 'aut') (delegated to Metadata#add_contributor) # add contributor. # === Book#contributor(content, id = nil, role = 'aut') (delegated to Metadata#contributor) # returns 'main' contributor. 'main' contributor determined as following: # 1. display-seq is smallest # 2. appears first in opf file # === Book#contributors_list (delegated to Metadata) # returns contributors list by display-seq or defined order. # the contributors without display-seq is appear after contributors with display-seq. # === Book#lastmodified(date) (delegated to Metadata#lastmodified) # set last modified date. # === Book#modified_now (delegated to Metadata#modified_now) # set last modified date to current time. # === Book#lastmodified (delegated to Metadata#lastmodified) # returns Meta object contains last modified time. # === setting and reading other metadata: publisher, language, coverage, date, description, format, relation, rights, source, subject, type (delegated to Metadata) # they all have methods like: publisher(which returns 'main' publisher), add_publisher(content, id) (which add publisher), set_publisher or publisher= (clears and set publisher), and publisher_list(returns publisher Meta object in display-seq order). # === Book#page_progression_direction= (delegated to Spine#page_progression_direction=) # set page-proression-direction attribute to spine. class Book MIMETYPE='mimetype' MIMETYPE_CONTENTS='application/epub+zip' CONTAINER='META-INF/container.xml' ROOTFILE_PATTERN=/^.+\.opf$/ CONTAINER_NS='urn:oasis:names:tc:opendocument:xmlns:container' def self.rootfile_from_container(rootfile) doc = Nokogiri::XML::Document.parse(rootfile) ns = doc.root.namespaces defaultns = ns.select{ |name, value| value == CONTAINER_NS }.to_a[0][0] doc.css("#{defaultns}|rootfiles > #{defaultns}|rootfile")[0]['full-path'] end # Parses existing EPUB2/EPUB3 files from an IO object, and creates new Book object. # book = self.parse(File.new('some.epub')) def self.parse(io) files = {} package = nil package_path = nil book = nil Zip::InputStream::open(io) { |zis| package, package_path = parse_container(zis, files) check_consistency_of_package(package, package_path) parse_files_into_package(files, package) book = Book.new(package.path) book.instance_eval { @package = package; @optional_files = files } } book end # creates new empty Book object. # usually you do not need to specify any arguments. def initialize(path='OEBPS/package.opf', attributes = {}, &block) if File.extname(path) != '.opf' warn 'GEPUB::Book#new interface changed. You must supply path to package.opf as first argument. If you want to set title, please use GEPUB::Book#title=' end @package = Package.new(path, attributes) @toc = [] if block block.arity < 1 ? instance_eval(&block) : block[self] end end # Get optional(not required in EPUB specification) files in the container. def optional_files @optional_files || {} end # Add an optional file to the container def add_optional_file(path, io_or_filename) io = io_or_filename if io_or_filename.class == String io = File.new(io_or_filename) end io.binmode (@optional_files ||= {})[path] = io.read end def set_singleton_methods_to_item(item) toc = @toc metaclass = (class << item;self;end) metaclass.send(:define_method, :toc_text, Proc.new { |text| toc.push(:item => item, :text => text, :id => nil) item }) metaclass.send(:define_method, :toc_text_with_id, Proc.new { |text, id| toc.push(:item => item, :text => text, :id => id) item }) bindings = @package.bindings metaclass.send(:define_method, :is_handler_of, Proc.new { |media_type| bindings.add(item.id, media_type) item }) end # add an item(i.e. html, images, audios, etc) to Book. # the added item will be referenced by the first argument in the EPUB container. def add_item(href, io_or_filename = nil, id = nil, attributes = {}) item = @package.add_item(href,nil,id,attributes) set_singleton_methods_to_item(item) item.add_content io_or_filename unless io_or_filename.nil? item end # same as add_item, but the item will be added to spine of the EPUB. def add_ordered_item(href, io_or_filename = nil, id = nil, attributes = {}) item = @package.add_ordered_item(href,io_or_filename,id,attributes) set_singleton_methods_to_item(item) yield item if block_given? item end # get handler item which defined in bindings for media type, def get_handler_of(media_type) items[@package.bindings.handler_by_media_type[media_type]] end def method_missing(name,*args, &block) @package.send(name, *args, &block) end # should call ordered() with block. # within the block, all item added by add_item will be added to spine also. def ordered(&block) @package.ordered(&block) end # clenup and maintain consistency of metadata and items included in the Book # object. def cleanup cleanup_for_epub2 cleanup_for_epub3 end # write EPUB to stream specified by the argument. def write_to_epub_container(epub) epub.put_next_entry('mimetype', '', '', Zip::Entry::STORED) epub << "application/epub+zip" entries = {} optional_files.each { |k, content| entries[k] = content } entries['META-INF/container.xml'] = container_xml entries[@package.path] = opf_xml @package.manifest.item_list.each { |k, item| if item.content != nil entries[@package.contents_prefix + item.href] = item.content end } entries.sort_by { |k,v| k }.each { |k,v| epub.put_next_entry(k) epub << v.force_encoding('us-ascii') } end # generates and returns StringIO contains EPUB. def generate_epub_stream cleanup Zip::OutputStream::write_buffer(StringIO.new) do |epub| write_to_epub_container(epub) end end # writes EPUB to file. if file exists, it will be overwritten. def generate_epub(path_to_epub) cleanup File.delete(path_to_epub) if File.exist?(path_to_epub) Zip::OutputStream::open(path_to_epub) { |epub| write_to_epub_container(epub) } end def container_xml < EOF end # add tocdata like this : [ {link: chapter1.xhtml, text: 'Capter 1', level: 1} ] . # if item corresponding to the link does not exists, error will be thrown. def add_tocdata(toc_yaml) newtoc = [] toc_yaml.each do |toc_entry| href, id = toc_entry[:link].split('#') item = @package.manifest.item_by_href(href) throw "#{href} does not exist." if item.nil? newtoc.push({item: item, id: id, text: toc_entry[:text], level: toc_entry[:level] }) end @toc = @toc + newtoc end def generate_nav_doc(title = 'Table of Contents') add_item('nav.xhtml', StringIO.new(nav_doc(title)), 'nav').add_property('nav') end def nav_doc(title = 'Table of Contents') # handle cascaded toc start_level = @toc && @toc[0][:level] || 1 stacked_toc = {level: start_level, tocs: [] } @toc.inject(stacked_toc) do |current_stack, toc_entry| toc_entry_level = toc_entry[:level] || 1 if current_stack[:level] < toc_entry_level new_stack = { level: toc_entry_level, tocs: [], parent: current_stack} current_stack[:tocs].last[:child_stack] = new_stack current_stack = new_stack else while current_stack[:level] > toc_entry_level and !current_stack[:parent].nil? current_stack = current_stack[:parent] end end current_stack[:tocs].push toc_entry current_stack end # write toc def write_toc xml_doc, tocs xml_doc.ol { tocs.each { |x| id = x[:id].nil? ? "" : "##{x[:id]}" toc_text = x[:text] toc_text = x[:item].href if toc_text.nil? or toc_text == '' xml_doc.li { xml_doc.a({'href' => x[:item].href + id} ,toc_text) if x[:child_stack] && x[:child_stack][:tocs].size > 0 write_toc(xml_doc, x[:child_stack][:tocs]) end } } } end # build nav builder = Nokogiri::XML::Builder.new { |doc| doc.html('xmlns' => "http://www.w3.org/1999/xhtml",'xmlns:epub' => "http://www.idpf.org/2007/ops") { doc.head { doc.text ' ' } doc.body { doc.nav('epub:type' => 'toc', 'id' => 'toc') { doc.h1 "#{title}" write_toc(doc, stacked_toc[:tocs]) } } } } builder.to_xml(:encoding => 'utf-8') end def ncx_xml builder = Nokogiri::XML::Builder.new { |xml| xml.ncx('xmlns' => 'http://www.daisy.org/z3986/2005/ncx/', 'version' => '2005-1') { xml.head { xml.meta('name' => 'dtb:uid', 'content' => "#{self.identifier}") xml.meta('name' => 'dtb:depth', 'content' => '1') xml.meta('name' => 'dtb:totalPageCount','content' => '0') xml.meta('name' => 'dtb:maxPageNumber', 'content' => '0') } xml.docTitle { xml.text_ "#{@package.metadata.title}" } count = 1 xml.navMap { @toc.each { |x| xml.navPoint('id' => "#{x[:item].itemid}_#{x[:id]}", 'playOrder' => "#{count}") { xml.navLabel { xml.text_ "#{x[:text]}" } if x[:id].nil? xml.content('src' => "#{x[:item].href}") else xml.content('src' => "#{x[:item].href}##{x[:id]}") end } count += 1 } } } } builder.to_xml(:encoding => 'utf-8') end private def self.parse_container(zis, files) package_path = nil package = nil while entry = zis.get_next_entry if !entry.directory? files[entry.name] = zis.read case entry.name when MIMETYPE then if files[MIMETYPE] != MIMETYPE_CONTENTS warn "#{MIMETYPE} is not valid: should be #{MIMETYPE_CONTENTS} but was #{files[MIMETYPE]}" end files.delete(MIMETYPE) when CONTAINER then package_path = rootfile_from_container(files[CONTAINER]) files.delete(CONTAINER) when ROOTFILE_PATTERN then package = Package.parse_opf(files[entry.name], entry.name) files.delete(entry.name) end end end return package, package_path end def self.check_consistency_of_package(package, package_path) if package.nil? raise 'this container do not cotains publication information file' end if package_path != package.path warn 'inconsistend EPUB file: container says opf is #{package_path}, but actually #{package.path}' end end def self.parse_files_into_package(files, package) files.each { |k, content| item = package.manifest.item_by_href(k.sub(/^#{package.contents_prefix}/,'')) if !item.nil? files.delete(k) item.add_raw_content(content) end } end def cleanup_for_epub2 if version.to_f < 3.0 || @package.epub_backward_compat if @package.manifest.item_list.select { |x,item| item.media_type == 'application/x-dtbncx+xml' }.size == 0 if (@toc.size == 0) @toc << { :item => @package.manifest.item_list[@package.spine.itemref_list[0].idref] } end add_item('toc.ncx', StringIO.new(ncx_xml), 'ncx') end end end def cleanup_for_epub3 if version.to_f >=3.0 @package.metadata.modified_now if @package.manifest.item_list.select { |href, item| (item.properties||[]).member? 'nav' }.size == 0 generate_nav_doc end @package.spine.remove_with_idlist @package.manifest.item_list.map { |href, item| item.fallback }.reject(&:nil?) end end end end