From fcd843204587ba984d1789cc87bfee1e222226a6 Mon Sep 17 00:00:00 2001 From: Florent Date: Thu, 24 Jan 2013 22:21:12 +0100 Subject: [PATCH] Use clean_relative_urls --- modules/taz/pages/article.py | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) diff --git a/modules/taz/pages/article.py b/modules/taz/pages/article.py index b769288372..aad673f5c3 100644 --- a/modules/taz/pages/article.py +++ b/modules/taz/pages/article.py @@ -19,7 +19,7 @@ # along with weboob. If not, see . from weboob.tools.capabilities.messages.genericArticle import GenericNewsPage,\ - try_drop_tree + try_drop_tree, clean_relativ_urls class ArticlePage(GenericNewsPage): @@ -33,15 +33,7 @@ def on_loaded(self): def get_body(self): div = self.document.getroot().find('.//div[@class="sectbody"]') try_drop_tree(self.parser, div, "div.anchor") - for a in div.findall('.//a'): - try: - if a.attrib["href"][0:7] != "http://": - a.attrib["href"] = "http://taz.de/" + a.attrib["href"] - except: - continue - for img in div.findall('.//img'): - if img.attrib["src"][0:7] != "http://": - img.attrib["src"] = "http://taz.de/" + img.attrib["src"] + clean_relativ_urls(div, "http://taz.de") return self.parser.tostring(div) -- GitLab