"""HTML Parser using ElementTree+TidyLib.
This file is a part of Bookmarks database and Internet robot.
+
"""
__author__ = "Oleg Broytman <phd@phdru.name>"
-__copyright__ = "Copyright (C) 2010-2013 PhiloSoft Design"
+__copyright__ = "Copyright (C) 2010-2014 PhiloSoft Design"
__license__ = "GNU GPL"
__all__ = ['parse_html']
from .bkmk_ph_util import HTMLParser
-def parse_html(filename, charset=None, log=None):
+def parse_html(html_text, charset=None, log=None):
try:
- html_tree = TidyHTMLTreeBuilder.parse(filename)
+ html_tree = TidyHTMLTreeBuilder.parseString(html_text)
except:
return None