Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-08-11 09:13:57 -04:00
Code Issues Packages Projects Releases Wiki Activity

Fix #1056270 (Updated recipe for Monitor online)

Browse Source
This commit is contained in:
Kovid Goyal 2012-09-26 10:24:13 +05:30
parent 0ec5e231dc
commit 075fbec828
2 changed files with 42 additions and 75 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
recipes/icons/monitor.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 1.1 KiB

117
recipes/monitor.recipe
View File

@ -1,99 +1,66 @@
#!/usr/bin/env python
__license__ = 'GPL v3' __license__ = 'GPL v3'
__copyright__ = '2009, Darko Miletic <darko.miletic at gmail.com>' __copyright__ = '2009-2012, Darko Miletic <darko.miletic at gmail.com>'
''' '''
monitorcg.com www.monitor.co.me
''' '''
import re import re
from calibre.web.feeds.news import BasicNewsRecipe from calibre.web.feeds.news import BasicNewsRecipe
from calibre.ebooks.BeautifulSoup import Tag
class MonitorCG(BasicNewsRecipe): class MonitorCG(BasicNewsRecipe):
title = 'Monitor online' title = 'MONITOR online'
__author__ = 'Darko Miletic' __author__ = 'Darko Miletic'
description = 'News from Montenegro' description = 'Nezavisni nedjeljnik Monitor'
publisher = 'MONITOR d.o.o. Podgorica' publisher = '"Monitor" D.O.O. Podgorica'
category = 'news, politics, Montenegro' category = 'news, politics, Montenegro'
oldest_article = 15 oldest_article = 15
max_articles_per_feed = 150 max_articles_per_feed = 150
no_stylesheets = True no_stylesheets = True
encoding = 'utf-8' encoding = 'utf-8'
auto_cleanup = False
use_embedded_content = False use_embedded_content = False
language = 'sr' language = 'sr'
remove_empty_feeds = True
lang ='sr-Latn-Me' extra_css = """
INDEX = 'http://www.monitorcg.com' @font-face {font-family: "serif1";src:url(res:///opt/sony/ebook/FONT/tt0011m_.ttf)}
@font-face {font-family: "sans1";src:url(res:///opt/sony/ebook/FONT/tt0003m_.ttf)}
extra_css = ' @font-face {font-family: "serif1";src:url(res:///opt/sony/ebook/FONT/tt0011m_.ttf)} body{font-family: serif1, serif} ' h2{font-family: Cambria,"Times New Roman",Times,serif1,serif}
body{font-family: Arial,sans1,sans-serif}
img{display: block}
"""
conversion_options = { conversion_options = {
'comment' : description 'comment' : description
, 'tags' : category , 'tags' : category
, 'publisher' : publisher , 'publisher' : publisher
, 'language' : lang , 'language' : language
, 'pretty_print' : True , 'pretty_print': True
} }
preprocess_regexps = [(re.compile(u'\u0110'), lambda match: u'\u00D0')] preprocess_regexps = [(re.compile(u'\u0110'), lambda match: u'\u00D0')]
keep_only_tags = [dict(attrs={'class':['contentheading','article-meta','article-content']})]
remove_attributes = ['width','height','font','border','align']
keep_only_tags = [dict(name='div', attrs={'id':'ja-current-content'})] feeds = [
(u'Danas, Sjutra' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=5&Itemid=27&format=feed&type=rss')
remove_tags = [ dict(name=['object','link','embed']) ,(u'Duhankesa' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=13&Itemid=37&format=feed&type=rss')
, dict(attrs={'class':['buttonheading','article-section']})] ,(u'Znaci prepoznavanja', u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=6&Itemid=358&format=feed&type=rss')
,(u'Paralele' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=8&Itemid=359&format=feed&type=rss')
remove_attributes = ['style','width','height','font','border','align'] ,(u'Razbijeno ogledalo' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=18&Itemid=354&format=feed&type=rss')
,(u'Tržište' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=26&Itemid=371&format=feed&type=rss')
def adeify_images2(cls, soup): ,(u'Feljton' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=29&Itemid=471&format=feed&type=rss')
for item in soup.findAll('img'): ,(u'Monitor' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=1&Itemid=1852&format=feed&type=rss')
for attrib in ['height','width','border','align','style']: ,(u'Altervizija' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=31&Itemid=2623&format=feed&type=rss')
if item.has_key(attrib): ,(u'Fenomeni' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=35&Itemid=3549&format=feed&type=rss')
del item[attrib] ,(u'Fokus' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=19&Itemid=252&format=feed&type=rss')
oldParent = item.parent ,(u'Monitoring' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=19&Itemid=252&format=feed&type=rss')
if oldParent.name == 'a': ,(u'Profil' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=21&Itemid=256&format=feed&type=rss')
oldParent.name == 'p' ,(u'Intervju' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=27&Itemid=404&format=feed&type=rss')
myIndex = oldParent.contents.index(item) ,(u'Društvo' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=14&Itemid=2&format=feed&type=rss')
brtag = Tag(soup,'br') ,(u'Region' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=12&Itemid=53&format=feed&type=rss')
oldParent.insert(myIndex+1,brtag) ,(u'Svijet' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=11&Itemid=360&format=feed&type=rss')
return soup ,(u'Kultura' , u'http://www.monitor.co.me/index.php?option=com_content&view=section&layout=blog&id=9&Itemid=361&format=feed&type=rss')
]
def preprocess_html(self, soup):
soup.html['xml:lang'] = self.lang
soup.html['lang'] = self.lang
mlang = Tag(soup,'meta',[("http-equiv","Content-Language"),("content",self.lang)])
soup.html.insert(0,mlang)
return self.adeify_images2(soup)
def parse_index(self):
totalfeeds = []
soup = self.index_to_soup(self.INDEX)
cover_item = soup.find('div',attrs={'class':'ja-catslwi'})
if cover_item:
dt = cover_item['onclick'].partition("location.href=")[2]
curl = self.INDEX + dt.strip("'")
lfeeds = [(u'Svi clanci', curl)]
for feedobj in lfeeds:
feedtitle, feedurl = feedobj
self.report_progress(0, _('Fetching feed')+' %s...'%(feedtitle if feedtitle else feedurl))
articles = []
soup = self.index_to_soup(feedurl)
contitem = soup.find('div',attrs={'class':'article-content'})
if contitem:
img = contitem.find('img')
if img:
self.cover_url = self.INDEX + img['src']
for item in contitem.findAll('a'):
url = self.INDEX + item['href']
title = self.tag_to_string(item)
articles.append({
'title' :title
,'date' :''
,'url' :url
,'description':''
})
totalfeeds.append((feedtitle, articles))
return totalfeeds