Super simple plugin for azw4 support.

src/calibre/customize/builtins.py

@@ -255,7 +255,7 @@ class LRXMetadataReader(MetadataReaderPlugin):
 class MOBIMetadataReader(MetadataReaderPlugin):
 
     name        = 'Read MOBI metadata'
-    file_types  = set(['mobi', 'prc', 'azw'])
+    file_types  = set(['mobi', 'prc', 'azw', 'azw4'])
     description = _('Read metadata from %s files')%'MOBI'
 
     def get_metadata(self, stream, ftype):
@@ -510,6 +510,7 @@ from calibre.ebooks.lit.input import LITInput
 from calibre.ebooks.mobi.input import MOBIInput
 from calibre.ebooks.odt.input import ODTInput
 from calibre.ebooks.pdb.input import PDBInput
+from calibre.ebooks.azw4.input import AZW4Input
 from calibre.ebooks.pdf.input import PDFInput
 from calibre.ebooks.pml.input import PMLInput
 from calibre.ebooks.rb.input import RBInput
@@ -606,6 +607,7 @@ plugins += [
     MOBIInput,
     ODTInput,
     PDBInput,
+    AZW4Input,
     PDFInput,
     PMLInput,
     RBInput,

src/calibre/ebooks/azw4/input.py

@@ -0,0 +1,26 @@
+# -*- coding: utf-8 -*-
+
+__license__   = 'GPL v3'
+__copyright__ = '2011, John Schember <john@nachtimwald.com>'
+__docformat__ = 'restructuredtext en'
+
+import os
+
+from calibre.customize.conversion import InputFormatPlugin
+from calibre.ebooks.pdb.header import PdbHeaderReader
+from calibre.ebooks.azw4.reader import Reader
+
+class AZW4Input(InputFormatPlugin):
+
+    name        = 'AZW4 Input'
+    author      = 'John Schember'
+    description = 'Convert AZW4 to HTML'
+    file_types  = set(['azw4'])
+
+    def convert(self, stream, options, file_ext, log,
+                accelerators):
+        header = PdbHeaderReader(stream)
+        reader = Reader(header, stream, log, options)
+        opf = reader.extract_content(os.getcwd())
+
+        return opf

src/calibre/ebooks/azw4/reader.py

@@ -0,0 +1,49 @@
+# -*- coding: utf-8 -*-
+
+'''
+Read content from azw4 file.
+
+azw4 is essentially a PDF stuffed into a MOBI container.
+'''
+
+__license__   = 'GPL v3'
+__copyright__ = '2011, John Schember <john@nachtimwald.com>'
+__docformat__ = 'restructuredtext en'
+
+
+import os
+import re
+
+from calibre.ebooks.pdb.formatreader import FormatReader
+
+class Reader(FormatReader):
+
+    def __init__(self, header, stream, log, options):
+        self.header = header
+        self.stream = stream
+        self.log = log
+        self.options = options
+
+    def extract_content(self, output_dir):
+        self.log.info('Extracting PDF from AZW4 Container...')
+
+        self.stream.seek(0)
+        raw_data = self.stream.read()
+        data = ''
+        mo = re.search(r'(?ums)%PDF.*%%EOF.', raw_data)
+        if mo:
+            data = mo.group()
+        
+        pdf_n = os.path.join(os.getcwdu(), 'tmp.pdf')
+        pdf = open(pdf_n, 'wb')
+        pdf.write(data)
+        pdf.close()
+    
+        from calibre.customize.ui import plugin_for_input_format
+        
+        pdf_plugin = plugin_for_input_format('pdf')
+        for opt in pdf_plugin.options:
+            if not hasattr(self.options, opt.option.name):
+                setattr(self.options, opt.option.name, opt.recommended_value)
+
+        return pdf_plugin.convert(open(pdf_n, 'rb'), self.options, 'pdf', self.log, {})