diff --git a/readability/readability.py b/readability/readability.py index a4db7a1..d690861 100755 --- a/readability/readability.py +++ b/readability/readability.py @@ -350,8 +350,9 @@ class Document: if len(s) < 2: continue #self.debug(s) - if REGEXES['unlikelyCandidatesRe'].search(s) and (not REGEXES['okMaybeItsACandidateRe'].search(s)) and elem.tag != 'body': + if REGEXES['unlikelyCandidatesRe'].search(s) and (not REGEXES['okMaybeItsACandidateRe'].search(s)) and elem.tag not in ['html', 'body']: self.debug("Removing unlikely candidate - %s" % describe(elem)) + import ipdb; ipdb.set_trace() elem.drop_tree() def transform_misused_divs_into_paragraphs(self):