X-Git-Url: https://git.mdrn.pl/librarian.git/blobdiff_plain/cba76114f031d47c93e1af947a350230cbef0a1f..3b0b98465bc1862306b05bb8305a1abbf40ca310:/librarian/fb2.py

diff --git a/librarian/fb2.py b/librarian/fb2.py
index 78707a9..bc1504d 100644
--- a/librarian/fb2.py
+++ b/librarian/fb2.py
@@ -7,11 +7,34 @@ import os.path
 from copy import deepcopy
 from lxml import etree
 
-from librarian import functions, OutputFile
+from librarian import functions, IOFile
 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 = {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):
@@ -32,9 +55,10 @@ def transform(wldoc, verbose=False,
     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 IOFile.from_string(unicode(result).encode('utf-8'))
 
 # vim:et