Improve metadata extraction from epubs, and clean up the display of popups in the...

[quanlib.git] / book.rb

diff --git a/book.rb b/book.rb
index 72bb83ad1170f77b29e58279eac85cb32c505146..3cf513957c0a2cc4173266857c95952b8cc0bb14 100644 (file)
--- a/book.rb
+++ b/book.rb
@@ -2,13 +2,16 @@
 require 'nokogiri'
 require 'zip'
 
-require './author'
-require './cover'
+require 'author'
+require 'cover'
 
 class Book
+  @@DC_NS_URL = 'http://purl.org/dc/elements/1.1/'
+
   def initialize(fileName)
     @author = nil
     @cover = nil
+    @description = nil
     @path = fileName
     @series = nil
     @title = nil
@@ -22,6 +25,7 @@ class Book
       return false
     end
 
+    #puts "Filename:  " + fileName.to_s
     lowerName = fileName.downcase()
 
     if lowerName.end_with?(".epub")
@@ -31,6 +35,40 @@ class Book
     return false
   end
 
+  def cover
+    return @cover
+  end
+
+  def description
+    @description
+  end
+
+  def heading
+    result = []
+
+    if nil != @title
+      result.push('<b>' + @title + '</b>')
+    else
+      result.push('<i>(Unknown title)</i>')
+    end
+    if nil != @author
+      result.push('<i>by ' + @author.to_s() + '</i>')
+    end
+    
+    seriesInfo = []
+    if nil != @series
+      seriesInfo.push(@series.to_s)
+    end
+    if nil != @volume
+      seriesInfo.push(@volume.to_s)
+    end
+    if seriesInfo.length > 0
+      result.push(seriesInfo.join(' '))
+    end
+
+    return result.join('<br/>')
+  end
+
   def inspect
     data = []
     if nil != @author
@@ -54,6 +92,10 @@ class Book
     return '(Book:' + data.join(',') + ')'
   end
 
+  def path
+    @path
+  end
+
   def to_s
     return inspect()
   end
@@ -126,12 +168,23 @@ class Book
 
   protected 
   def scanEpub!(fileName)
-    Zip::File.open(fileName) do |zipfile|
-      contXml = zipfile.read('META-INF/container.xml')
-      contDoc = Nokogiri::XML(contXml)
-      opfPath = contDoc.css("container rootfiles rootfile")[0]['full-path']
+    #puts 'Scanning "' + fileName.to_s + '"...'
+    begin
+      Zip::File.open(fileName) do |zipfile|
+        entry = zipfile.find_entry('META-INF/container.xml')
+        if nil == entry
+          return
+        end
+        contXml = zipfile.read('META-INF/container.xml')
+        contDoc = Nokogiri::XML(contXml)
+        opfPath = contDoc.css("container rootfiles rootfile")[0]['full-path']
 
-      scanOpf!(zipfile, opfPath)
+        scanOpf!(zipfile, opfPath)
+      end
+    rescue Zip::Error => exc
+      puts 'ERROR processing file "' + fileName + '":'
+      puts exc.message
+      puts exc.backtrace
     end
   end
 
@@ -145,13 +198,13 @@ class Book
     #-------
     # Author
 
-    creator = opfDoc.css('dc|creator', 'dc' => 'http://purl.org/dc/elements/1.1/')
-    if nil != creator
-      roleNode = creator.attr('role')
-      if nil != roleNode
-        role = roleNode.value
+    creators = opfDoc.css('dc|creator', 'dc' => @@DC_NS_URL)
+    if (creators.length > 0)
+      creator = creators[0]
+      if nil != creator
+        role = creator['opf:role']
         if 'aut' == role
-          name = creator.children[0].content
+          name = creator.content
           parts = name.split(' ')
           if parts.length > 1
             surname = parts[-1]
@@ -164,6 +217,28 @@ class Book
       end
     end
 
+    #---------------------------------------
+    # Title
+
+    titles = opfDoc.css('dc|title', 'dc' => @@DC_NS_URL)
+    if titles.length > 0
+      title = titles[0]
+      if nil != title
+        @title = title.content
+      end
+    end
+
+    #---------------------------------------
+    # Description
+    
+    descrNodes = opfDoc.css('dc|description', 'dc' => @@DC_NS_URL)
+    if (descrNodes.length > 0)
+      descrNode = descrNodes[0]
+      if nil != descrNode
+        @description = descrNode.content
+      end
+    end
+
     #---------------------------------------
     # Other metadata:  series, volume, cover
 
@@ -184,22 +259,47 @@ class Book
     #---------------
     # Load the cover
 
+    @cover = loadCover(zipfile, opfPath, opfDoc, coverId)
+  end
+
+  protected
+  def loadCover(zipfile, opfPath, opfDoc, coverId)
     coverFile = nil
-    if nil != coverId
-      items = opfDoc.css('package manifest item')
-      for i in items
-        href = i['href']
-        id = i['id']
-        mimeType = i['media-type']
-
-        if coverId == id
-          entry = zipfile.find_entry(href)
+    if nil == coverId
+      coverId = "cover-image"
+    end
+
+    items = opfDoc.css('package manifest item')
+    for i in items
+      href = i['href']
+      id = i['id']
+      mimeType = i['media-type']
+
+      if coverId == id
+        entry = zipfile.find_entry(href)
+
+        if nil == entry
+          # Although the epub standard requires the path to be relative 
+          # to the base of the epub (zip), some books encountered in the
+          # wild have been found to use a bath relative to the location 
+          # of the opf file.
+          parts = opfPath.split('/')
+          opfBasePath = opfPath.split('/')[0..-2].join('/')
+          coverPath = opfBasePath + '/' + href
+          entry = zipfile.find_entry(coverPath)
+        end
+
+        if nil == entry
+          puts 'WARNING!  Cover image "' + href + '" not found in file "' + @path + '".'
+          return nil
+        else
           entry.get_input_stream() do |is|
-            @cover = Cover.new(is, href, mimeType)
+            return Cover.new(is, href, mimeType)
           end
         end
       end
     end
+    return nil
   end
 end