X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/2903250a90904488ff3de04ecbbf2b6b5c421839..40b0da924b62ba147588a945718b71dc36ed3702:/lib/librarian/bin/book2txt.py?ds=sidebyside

diff --git a/lib/librarian/bin/book2txt.py b/lib/librarian/bin/book2txt.py
index 86f6f1274..1ca4623fd 100755
--- a/lib/librarian/bin/book2txt.py
+++ b/lib/librarian/bin/book2txt.py
@@ -1,25 +1,8 @@
 #!/usr/bin/env python
-# -*- coding: utf-8 -*-
-import re
 import os
 import optparse
-import codecs
 
-
-REGEXES = [
-    (r'<rdf:RDF[^>]*>(.|\n)*?</rdf:RDF>', ''),
-    (r'<motyw[^>]*>(.|\n)*?</motyw>', ''),
-    ('<(begin|end)\\sid=[\'|"][b|e]\\d+[\'|"]\\s/>', ''),
-    (r'<extra>((<!--<(elementy_poczatkowe|tekst_glowny)>-->)|(<!--</(elementy_poczatkowe|tekst_glowny)>-->))</extra>', ''),
-    (r'<uwaga>[^<]*</uwaga>', ''),
-    (r'<p[a|e|r|t]>(.|\n)*?</p>', ''),
-    (r'<[^>]+>', ''),
-    (r'/$', ''),
-    (r'---', u'—'),
-    (r'--', u'-'),
-    (r',,', u'„'),
-    (r'"', u'”'),
-]
+from librarian import text
 
 
 if __name__ == '__main__':
@@ -44,13 +27,5 @@ if __name__ == '__main__':
             print input_filename
         
         output_filename = os.path.splitext(input_filename)[0] + '.txt'
-        
-        xml = codecs.open(input_filename, 'r', encoding='utf-8').read()
-        for pattern, repl in REGEXES:
-            # print pattern, repl
-            xml, n = re.subn(pattern, repl, xml)
-            # print n
-            
-        output = codecs.open(output_filename, 'w', encoding='utf-8')
-        output.write(xml)
+        text.transform(input_filename, output_filename)