From 37eacf65b09b5fadb4573e976bc1a75ce5ec37f4 Mon Sep 17 00:00:00 2001 From: Kovid Goyal Date: Tue, 23 Feb 2016 09:02:55 +0530 Subject: [PATCH] Update Today Online - Singapore --- recipes/today_online.recipe | 37 +++++++++++++++++++------------------ 1 file changed, 19 insertions(+), 18 deletions(-) diff --git a/recipes/today_online.recipe b/recipes/today_online.recipe index e2fc229473..86e8bdb78b 100644 --- a/recipes/today_online.recipe +++ b/recipes/today_online.recipe @@ -14,7 +14,7 @@ class AdvancedUserRecipe1276486274(BasicNewsRecipe): language = 'en_SG' temp_files = [] articles_are_obfuscated = True - masthead_url = 'http://www.todayonline.com/App_Themes/Default/images/icons/TodayOnlineLogo.gif' + masthead_url = 'http://www.todayonline.com/sites/all/themes/today/logo.png' conversion_options = {'linearize_tables':True} extra_css = ''' .author{font-style: italic; font-size: small} @@ -23,30 +23,31 @@ class AdvancedUserRecipe1276486274(BasicNewsRecipe): .headerStrap{font-weight: bold; font-size: x-large; font-syle: italic} .bodyText{font-size: 4px;font-family: Times New Roman;} ''' + feeds = [ + (u'Hot News', u'http://www.todayonline.com/hot-news/feed'), + (u'Singapore', u'http://www.todayonline.com/feed/singapore'), + (u'World', u'http://www.todayonline.com/feed/world'), + (u'Business', u'http://www.todayonline.com/feed/business'), + (u'Tech', u'http://www.todayonline.com/feed/tech'), + (u'Voices', u'http://www.todayonline.com/feed/voices'), + (u'Commentary', u'http://www.todayonline.com/feed/Commentary'), + (u'Daily Focus', u'http://www.todayonline.com/feed/daily-focus'), + (u'Lifestyle', u'http://www.todayonline.com/feed/lifestyle'), + ] keep_only_tags = [ - dict(name='div', attrs={'id':['fullPrintBodyHolder']}) + dict(name='div', attrs='print-content') ] - remove_tags_after = [ dict(name='div', attrs={'class':'button'})] - remove_tags = [ - dict(name='div', attrs={'class':['url','button']}) - ] - feeds = [ - (u'Singapore', u'http://www.todayonline.com/RSS/Singapore'), - (u'Hot News', u'http://www.todayonline.com/RSS/Hotnews'), - (u'Today Online', u'http://www.todayonline.com/RSS/Todayonline'), - (u'Voices', u'http://www.todayonline.com/RSS/Voices'), - (u'Commentary', u'http://www.todayonline.com/RSS/Commentary'), - (u'World', u'http://www.todayonline.com/RSS/World'), - (u'Business', u'http://www.todayonline.com/RSS/Business'), - (u'Column', u'http://www.todayonline.com/RSS/Columns'), - ] + dict(name='div', attrs={'class':['url','button']}), + dict(name='div', attrs={'class':'node-type-print-edition'}), + dict(name='div', attrs={'class':['field field-name-field-article-section field-type-taxonomy-term-reference field-label-hidden','field field-name-field-article-abstract field-type-text-long field-label-hidden','authoring']}) + ] def get_obfuscated_article(self, url): br = self.get_browser() br.open(url) - response = br.follow_link(url_regex = r'/Print/', nr = 0) + response = br.follow_link(url_regex = r'/print/', nr = 0) html = response.read() self.temp_files.append(PersistentTemporaryFile('_fa.html')) self.temp_files[-1].write(html) @@ -55,5 +56,5 @@ class AdvancedUserRecipe1276486274(BasicNewsRecipe): def preprocess_html(self, soup): for item in soup.findAll(style=True): - del item['style'] + del item['style'] return soup