X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/c51d51f2b1fd8d2eaa63224bb0aeac6b85ac3737..b63b804c605433dde83600f6e6906a19bc8141b0:/librarian/fb2.py

diff --git a/librarian/fb2.py b/librarian/fb2.py
index 78707a9..6dd1c35 100644
--- a/librarian/fb2.py
+++ b/librarian/fb2.py
@@ -3,15 +3,41 @@
 # This file is part of Librarian, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
+from __future__ import unicode_literals
+
 import os.path
 from copy import deepcopy
 from lxml import etree
+import six
 
 from librarian import functions, OutputFile
 from .epub import replace_by_verse
 
 
 functions.reg_substitute_entities()
+functions.reg_person_name()
+
+
+def sectionify(tree):
+    """Finds section headers and adds a tree of _section tags."""
+    sections = [
+        'naglowek_czesc',
+        'naglowek_akt', 'naglowek_rozdzial', 'naglowek_scena',
+        'naglowek_podrozdzial']
+    section_level = dict((v, k) for (k, v) in enumerate(sections))
+
+    # We can assume there are just subelements an no text at section level.
+    for level, section_name in reversed(list(enumerate(sections))):
+        for header in tree.findall('//' + section_name):
+            section = header.makeelement("_section")
+            header.addprevious(section)
+            section.append(header)
+            sibling = section.getnext()
+            while (sibling is not None and
+                    section_level.get(sibling.tag, 1000) > level):
+                section.append(sibling)
+                sibling = section.getnext()
+
 
 def transform(wldoc, verbose=False,
               cover=None, flags=None):
@@ -28,13 +54,17 @@ def transform(wldoc, verbose=False,
         for flag in flags:
             document.edoc.getroot().set(flag, 'yes')
 
+    document.clean_ed_note()
+    document.clean_ed_note('abstrakt')
+
     style_filename = os.path.join(os.path.dirname(__file__), 'fb2/fb2.xslt')
     style = etree.parse(style_filename)
 
     replace_by_verse(document.edoc)
+    sectionify(document.edoc)
 
     result = document.transform(style)
 
-    return OutputFile.from_string(unicode(result).encode('utf-8'))
+    return OutputFile.from_bytes(six.text_type(result).encode('utf-8'))
 
 # vim:et