#1030: fixes for separators and empty verses in html, epub, pdf

[librarian.git] / scripts / book2txt

diff --git a/scripts/book2txt b/scripts/book2txt
index 41a3978..2317688 100755 (executable)
--- a/scripts/book2txt
+++ b/scripts/book2txt
@@ -1,4 +1,9 @@
 #!/usr/bin/env python
+# -*- coding: utf-8 -*-
+#
+# This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
+# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.  
+#
 import os
 import optparse
 
@@ -17,7 +22,9 @@ if __name__ == '__main__':
         help='print status messages to stdout')
     parser.add_option('-w', '--wrap', action='store', type='int', dest='wrapping', default=0,
         help='set line wrap column')
-    
+    parser.add_option('-i', '--ignore-dublin-core', action='store_false', dest='parse_dublincore', default=True,
+        help='don\'t try to parse dublin core metadata')
+            
     options, input_filenames = parser.parse_args()
     
     if len(input_filenames) < 1:
@@ -31,7 +38,8 @@ if __name__ == '__main__':
         
         output_filename = os.path.splitext(input_filename)[0] + '.txt'
         try:
-            text.transform(input_filename, output_filename, wrapping=str(options.wrapping))
+            text.transform(input_filename, output_filename, parse_dublincore=options.parse_dublincore,
+                wrapping=str(options.wrapping))
         except ParseError, e:
             print '%(file)s:%(name)s:%(message)s' % {
                 'file': input_filename,
@@ -50,4 +58,4 @@ if __name__ == '__main__':
                 'etype': e.__class__.__name__,
                 'message': e.message,
             }
-            raise e
+            raise