Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'smh-au-recipe-fix' of https://github.com/ping/calibre

Browse Source
This commit is contained in:
Kovid Goyal 2022-06-14 07:46:24 +05:30
commit 7e72349b5f
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
1 changed files with 36 additions and 52 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

88
recipes/smh.recipe
View File

@ -1,66 +1,50 @@
__license__ = 'GPL v3'
__copyright__ = '2010-2011, Darko Miletic <darko.miletic at gmail.com>'
'''
__license__ = "GPL v3"
__copyright__ = "2010-2011, Darko Miletic <darko.miletic at gmail.com>"
"""
smh.com.au
'''
"""
from calibre.web.feeds.news import BasicNewsRecipe
class Smh_au(BasicNewsRecipe):
title = 'The Sydney Morning Herald'
__author__ = 'Darko Miletic'
description = 'Breaking news from Sydney, Australia and the world. Features the latest business, sport, entertainment, travel, lifestyle, and technology news.' # noqa
publisher = 'Fairfax Digital'
category = 'news, politics, Australia, Sydney'
title = "The Sydney Morning Herald"
__author__ = "Darko Miletic"
description = "Breaking news from Sydney, Australia and the world. Features the latest business, sport, entertainment, travel, lifestyle, and technology news." # noqa
publisher = "Fairfax Digital"
category = "news, politics, Australia, Sydney"
oldest_article = 2
max_articles_per_feed = 200
no_stylesheets = True
ignore_duplicate_articles = {'title', 'url'}
ignore_duplicate_articles = {"title", "url"}
use_embedded_content = False
encoding = 'utf-8'
use_embedded_content = False
language = 'en_AU'
encoding = "utf-8"
language = "en_AU"
remove_empty_feeds = True
masthead_url = 'http://images.smh.com.au/2010/02/02/1087188/smh-620.jpg'
publication_type = 'newspaper'
extra_css = """
h1{font-family: Georgia,"Times New Roman",Times,serif }
body{font-family: Arial,Helvetica,sans-serif}
.cT-imageLandscape,.cT-imagePortrait{font-size: x-small}
"""
masthead_url = "http://images.smh.com.au/2010/02/02/1087188/smh-620.jpg"
publication_type = "newspaper"
feeds = [
('Politics', 'http://www.smh.com.au/rssheadlines/political-news/article/rss.xml'),
('NSW News', 'http://www.smh.com.au/rssheadlines/nsw/article/rss.xml'),
('World', 'http://www.smh.com.au/rssheadlines/world/article/rss.xml'),
('National', 'http://www.smh.com.au/rssheadlines/national/article/rss.xml'),
('Business', 'http://www.smh.com.au/rssheadlines/business/article/rss.xml'),
('Entertainment', 'http://feeds.smh.com.au/rssheadlines/entertainment.xml'),
('Technology', 'http://www.smh.com.au/rssheadlines/technology-news/article/rss.xml'),
('Health', 'http://www.smh.com.au/rssheadlines/health/article/rss.xml'),
('Sport', 'http://feeds.smh.com.au/rssheadlines/sport.xml'),
('Cricket', 'http://www.smh.com.au/rssheadlines/cricket/article/rss.xml'),
]
remove_tags_after = [
dict(name='div', attrs={'class': 'articleBody'})
]
keep_only_tags = [
dict(name='div', attrs={'id': 'content'})
]
keep_only_tags = [dict(name="article")]
remove_tags = [
dict(name='div', attrs={
'id': ['googleAds', 'moreGoogleAds', 'comments', 'video-player-content']}),
dict(name='div', attrs={'class': 'cT-imageMultimedia'}),
dict(name=['object', 'embed', 'iframe']),
dict(attrs={'class': 'hidden'}),
dict(name=['link', 'meta', 'base', 'embed', 'object', 'iframe'])
dict(name=["button"]),
dict(id=["saveTooltip"]),
dict(attrs={"class": "noPrint"}),
]
remove_attributes = ['width', 'height', 'lang']
def preprocess_html(self, soup):
for item in soup.findAll(style=True):
del item['style']
for item in soup.findAll('bod'):
item.name = 'div'
return soup
# https://www.smh.com.au/rssheadlines
feeds = [
("Latest News", "https://www.smh.com.au/rss/feed.xml"),
("Federal Politics", "https://www.smh.com.au/rss/politics/federal.xml"),
("NSW News", "https://www.smh.com.au/rss/national/nsw.xml"),
("World", "https://www.smh.com.au/rss/world.xml"),
("National", "https://www.smh.com.au/rss/national.xml"),
("Business", "https://www.smh.com.au/rss/business.xml"),
("Culture", "https://www.smh.com.au/rss/culture.xml"),
("Technology", "https://www.smh.com.au/rss/technology.xml"),
("Environment", "https://www.smh.com.au/rss/environment.xml"),
("Lifestyle", "https://www.smh.com.au/rss/lifestyle.xml"),
("Property", "https://www.smh.com.au/rss/property.xml"),
("Sport", "https://www.smh.com.au/rss/sport.xml"),
("Ruby League", "https://www.smh.com.au/rss/sport/nrl.xml"),
("AFL", "https://www.smh.com.au/rss/sport/afl.xml"),
]