diff --git a/resources/images/news/tulsaworld.png b/resources/images/news/tulsaworld.png new file mode 100644 index 0000000000..dc68d07266 Binary files /dev/null and b/resources/images/news/tulsaworld.png differ diff --git a/resources/recipes/tulsaworld.recipe b/resources/recipes/tulsaworld.recipe new file mode 100644 index 0000000000..bdb6969853 --- /dev/null +++ b/resources/recipes/tulsaworld.recipe @@ -0,0 +1,47 @@ +__license__ = 'GPL v3' +__copyright__ = '2010, Darko Miletic ' +''' +tulsaworld.com +''' + +from calibre.web.feeds.news import BasicNewsRecipe + +class TulsaWorld(BasicNewsRecipe): + title = 'Tulsa World' + __author__ = 'Darko Miletic' + description = 'Find breaking news, local news, Oklahoma weather, sports, business, entertainment, lifestyle, opinion, government, movies, books, jobs, education, blogs, video & multimedia.' + publisher = 'World Publishing Co.' + category = 'Tulsa World, tulsa world, daily newspaper, breaking news, stories, articles, news, local, weather, coverage, editorial, government, education, community, sports, business, entertainment, lifestyle, opinion, multimedia, media, blogs, consumer, OU, OSU, TU, ORU, football, basketball, school, schools, sudoku, movie reviews, stocks, classified ads, classifieds, books, job, jobs, careers, real estate, home, homes, Oklahoma, northeastern, reviews, auto, autos, archives, forecasts, Sooners, Cowboys, Hurricane, Golden Eagles, NFL, NBA, MLB, pro football, scores, college basketball, college football, college baseball, sports columns, fashion and style, associated press, regional news coverage, health, obituaries, politics, political news, Jenks, Union, Owasso, Tulsa, Booker T. Washington, Trojans, Rams, Hornets, video, photography, photos, images, games, search, the picker, predictions, satellite, family, food, teens, polls, births, celebrations, death notices, divorces, marriages, obituaries, audio, podcasts.' + oldest_article = 2 + max_articles_per_feed = 200 + no_stylesheets = True + encoding = 'utf8' + use_embedded_content = False + language = 'en' + country = 'US' + remove_empty_feeds = True + masthead_url = 'http://www.tulsaworld.com/images/TW_logo-blue-footer.jpg' + extra_css = ' body{font-family: Arial,Verdana,sans-serif } img{margin-bottom: 0.4em} .articleHeadline{font-size: xx-large; font-weight: bold} .articleKicker{font-size: x-large; font-weight: bold} .articleByline,.articleDate{font-size: small} .leadp{font-size: 1.1em} ' + + conversion_options = { + 'comment' : description + , 'tags' : category + , 'publisher' : publisher + , 'language' : language + , 'linearize_tables' : True + } + keep_only_tags = [dict(name='div',attrs={'id':['ctl00_body1_ArticleControl_divArticleText','ctl00_BodyContent_ArticleControl_divArticleText']})] + + feeds = [ + (u'News' , u'http://www.tulsaworld.com/site/rss.aspx?group=1') + ,(u'Business', u'http://www.tulsaworld.com/site/rss.aspx?group=5') + ,(u'Opinion' , u'http://www.tulsaworld.com/site/rss.aspx?group=7') + ] + + def get_article_url(self, article): + return article.get('link', None).rpartition('&rss')[0] + + def preprocess_html(self, soup): + for item in soup.findAll(style=True): + del item['style'] + return self.adeify_images(soup) \ No newline at end of file