bugfix for <extra> with tail

[librarian.git] / librarian / html.py

diff --git a/librarian/html.py b/librarian/html.py
index 4e9abd0..5f832e3 100644 (file)
--- a/librarian/html.py
+++ b/librarian/html.py
@@ -16,6 +16,7 @@ from librarian import functions
 from lxml.etree import XMLSyntaxError, XSLTApplyError
 
 functions.reg_substitute_entities()
 from lxml.etree import XMLSyntaxError, XSLTApplyError
 
 functions.reg_substitute_entities()

+functions.reg_person_name()

 
 STYLESHEETS = {
     'legacy': 'xslt/book2html.xslt',
 
 STYLESHEETS = {
     'legacy': 'xslt/book2html.xslt',


@@ -126,7 +127,13 @@ def extract_fragments(input_filename):
     open_fragments = {}
     closed_fragments = {}
 
     open_fragments = {}
     closed_fragments = {}
 

-    for event, element in etree.iterparse(input_filename, events=('start', 'end')):
+    # iterparse would die on a HTML document
+    parser = etree.HTMLParser(encoding='utf-8')
+    buf = cStringIO.StringIO()
+    buf.write(etree.tostring(etree.parse(input_filename, parser).getroot()[0][0], encoding='utf-8'))
+    buf.seek(0)
+
+    for event, element in etree.iterparse(buf, events=('start', 'end')):

         # Process begin and end elements
         if element.get('class', '') in ('theme-begin', 'theme-end'):
             if not event == 'end': continue # Process elements only once, on end event
         # Process begin and end elements
         if element.get('class', '') in ('theme-begin', 'theme-end'):
             if not event == 'end': continue # Process elements only once, on end event