subject.location in dublin core

[librarian.git] / librarian / fb2.py

diff --git a/librarian/fb2.py b/librarian/fb2.py
index 78707a9..25a4c1f 100644 (file)
--- a/librarian/fb2.py
+++ b/librarian/fb2.py
@@ -12,6 +12,29 @@ 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):
@@ -28,10 +51,14 @@ 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)