From 840529072e4fb7c3ba6e81837a4f5366d72b569b Mon Sep 17 00:00:00 2001 From: unkn0w7n <51942695+unkn0w7n@users.noreply.github.com> Date: Tue, 2 May 2023 13:07:09 +0530 Subject: [PATCH 1/3] website got closed --- recipes/chowk.recipe | 45 -------------------------------------------- 1 file changed, 45 deletions(-) delete mode 100644 recipes/chowk.recipe diff --git a/recipes/chowk.recipe b/recipes/chowk.recipe deleted file mode 100644 index eaab142724..0000000000 --- a/recipes/chowk.recipe +++ /dev/null @@ -1,45 +0,0 @@ -from calibre.web.feeds.news import BasicNewsRecipe - - -class ChowkRecipe(BasicNewsRecipe): - __license__ = 'GPL v3' - __author__ = 'kwetal' - language = 'en_IN' - version = 1 - - title = u'Chowk' - publisher = u'chowk.com' - category = u'Opinion, South Asia' - description = u'Ideas & Identities of South Asia' - - use_embedded_content = False - remove_empty_feeds = True - oldest_article = 30 - max_articles_per_feed = 100 - - remove_javascript = True - encoding = 'utf-8' - - feeds = [] - feeds.append(('Chowk Articles', 'http://www.chowk.com/rss')) - - keep_only_tags = [] - keep_only_tags.append(dict(name='div', attrs={'id': 'content'})) - - conversion_options = {'comments': description, 'tags': category, 'language': 'en', - 'publisher': publisher} - - extra_css = ''' - body{font-family:verdana,arial,helvetica,geneva,sans-serif;} - a {text-decoration: none; color: blue;} - div.pgtitle {font-size: x-large; font-weight: bold;} - div.wname, div.date {font-size: x-small; color: #696969;} - div.wname {margin-top: 1em;} - div.date {margin-bottom: 1em;} - div.title {font-weight: bold;} - ''' - - def print_version(self, url): - main, sep, id = url.rpartition('/') - - return main + '/print/' + id From dbc9e64fb38dac5bb78a549cfea5d53ea1fb9cb7 Mon Sep 17 00:00:00 2001 From: unkn0w7n <51942695+unkn0w7n@users.noreply.github.com> Date: Fri, 5 May 2023 23:12:12 +0530 Subject: [PATCH 2/3] Update irish_times.recipe --- recipes/irish_times.recipe | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/recipes/irish_times.recipe b/recipes/irish_times.recipe index 3c3c97d0c1..5574f975b9 100644 --- a/recipes/irish_times.recipe +++ b/recipes/irish_times.recipe @@ -32,12 +32,11 @@ class IrishTimes(BasicNewsRecipe): no_stylesheets = True temp_files = [] keep_only_tags = [ - dict(name=['h1', 'h2']), - classes('lead-art-wrapper article-body-wrapper byline-text'), + classes('custom-headline custom-subheadline lead-art-wrapper article-body-wrapper byline-text'), ] remove_tags = [ dict(name='button'), - classes('sm-promo-headline top-table-list-container'), + classes('sm-promo-headline top-table-list-container single-divider interstitial-link'), ] remove_attributes = ['width', 'height'] From 303a43fc9ba816958f53d23c1073a93d997e9ea8 Mon Sep 17 00:00:00 2001 From: unkn0w7n <51942695+unkn0w7n@users.noreply.github.com> Date: Fri, 5 May 2023 23:24:37 +0530 Subject: [PATCH 3/3] Update irish_independent.recipe removed duplicate feed and articles. --- recipes/irish_independent.recipe | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/recipes/irish_independent.recipe b/recipes/irish_independent.recipe index 105260fd02..902beff249 100644 --- a/recipes/irish_independent.recipe +++ b/recipes/irish_independent.recipe @@ -15,6 +15,7 @@ class IrishIndependent(BasicNewsRecipe): oldest_article = 2 max_articles_per_feed = 100 no_stylesheets = True + ignore_duplicate_articles = {'url'} keep_only_tags = [ dict(name='div', attrs={'class':lambda x: x and '_contentwrapper' in x}) @@ -26,7 +27,6 @@ class IrishIndependent(BasicNewsRecipe): feeds = [ ('Frontpage News', 'http://www.independent.ie/rss'), - ('News', 'http://www.independent.ie/rss'), ('World News', 'http://www.independent.ie/world-news/rss'), ('Opinion', 'http://www.independent.ie/opinion/rss'), ('Business', 'http://www.independent.ie/business/rss'),