Improve HTML formatting. Handle more variants of EPUB.

[quanlib.git] / book.rb

diff --git a/book.rb b/book.rb
index 370093da92964413102b286792ab27a4e8ed6415..f59b61a66ea44d17a86e7d999e34496b84fbea99 100644 (file)
--- a/book.rb
+++ b/book.rb
@@ -22,6 +22,7 @@ class Book
       return false
     end
 
+    #puts "Filename:  " + fileName.to_s
     lowerName = fileName.downcase()
 
     if lowerName.end_with?(".epub")
@@ -84,6 +85,10 @@ class Book
     return '(Book:' + data.join(',') + ')'
   end
 
+  def path
+    @path
+  end
+
   def to_s
     return inspect()
   end
@@ -156,13 +161,23 @@ class Book
 
   protected 
   def scanEpub!(fileName)
-    puts 'Scanning "' + fileName.to_s + '"...'
-    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
 
@@ -177,11 +192,11 @@ class Book
     # Author
 
     creator = opfDoc.css('dc|creator', 'dc' => 'http://purl.org/dc/elements/1.1/')
-    if nil != creator
+    if (nil != creator) and (creator.length > 0)
       roleNode = creator.attr('role')
       if nil != roleNode
         role = roleNode.value
-        if 'aut' == role
+        if ('aut' == role) and (creator.children.length > 0) and (nil != creator.children[0])
           name = creator.children[0].content
           parts = name.split(' ')
           if parts.length > 1
@@ -215,22 +230,47 @@ 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