Read metadata from ebooks in zip archives. Fixes #1570 (FB2 metadata from ZIP archive)

src/calibre/customize/builtins.py

@@ -25,6 +25,17 @@ every time you add an HTML file to the library.\
         html2oeb(htmlfile, of)
         return of.name
 
+class OPFMetadataReader(MetadataReaderPlugin):
+    
+    name        = 'Read OPF metadata'
+    file_types  = set(['opf'])
+    description = _('Read metadata from %s files')%'OPF'
+    
+    def get_metadata(self, stream, ftype):
+        from calibre.ebooks.metadata.opf2 import OPF
+        from calibre.ebooks.metadata import MetaInformation
+        return MetaInformation(OPF(stream, os.getcwd()))
+
 class RTFMetadataReader(MetadataReaderPlugin):
     
     name        = 'Read RTF metadata' 
@@ -167,6 +178,16 @@ class ComicMetadataReader(MetadataReaderPlugin):
             ext = os.path.splitext(path)[1][1:]
             mi.cover_data = (ext.lower(), data)
         return mi
+    
+class ZipMetadataReader(MetadataReaderPlugin):
+    
+    name = 'Read ZIP metadata'
+    file_types = set(['zip', 'oebzip'])
+    description = _('Read metadata from ebooks in ZIP archives')
+    
+    def get_metadata(self, stream, ftype):
+        from calibre.ebooks.metadata.zip import get_metadata
+        return get_metadata(stream)
 
 class EPUBMetadataWriter(MetadataWriterPlugin):
     

src/calibre/ebooks/metadata/zip.py

@@ -0,0 +1,24 @@
+from __future__ import with_statement
+__license__   = 'GPL v3'
+__copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
+
+import os
+from zipfile import ZipFile 
+from cStringIO import StringIO 
+
+
+def get_metadata(stream):
+    stream_type = None
+    zf = ZipFile(stream, 'r')
+    for f in zf.namelist():
+        stream_type = os.path.splitext(f)[1].lower()
+        if stream_type:
+            stream_type = stream_type[1:]
+            if stream_type in ('lit', 'opf', 'prc', 'mobi', 'fb2', 'epub',
+                               'rb', 'imp', 'pdf', 'lrf'):
+                from calibre.ebooks.metadata.meta import get_metadata
+                stream = StringIO(zf.read(f))
+                return get_metadata(stream, stream_type)
+    raise ValueError('No ebook found in ZIP archive')
+            
+