def build_link_map(directory):
# find all html files
html_files = []
for root, dirnames, filenames in os.walk(directory):
for filename in fnmatch.filter(filenames, '*.html'):
html_files.append(os.path.join(root, filename))
link_map = LinkMap()
for fn in html_files:
f = open(fn, "r")
text = f.read()
f.close()
m = re.search('<script>[^<]*mw\.config\.set([^<]*wgPageName[^<]*)</script>', text)
if not m:
continue
text = m.group(1)
text = re.sub('\s*', '', text)
m = re.search('"wgPageName":"([^"]*)"', text)
if not m:
continue
title = m.group(1)
target = os.path.relpath(os.path.abspath(fn), os.path.abspath(directory))
link_map.add_link(title, target)
return link_map