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

Update Al Jazeera in English

Browse Source
This commit is contained in:
Kovid Goyal 2015-01-27 06:35:05 +05:30
parent 4fceeb4ddd
commit 50ba062e82
1 changed files with 16 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
recipes/al_jazeera.recipe
View File

@ -6,6 +6,9 @@ english.aljazeera.net
''' '''
from calibre.web.feeds.news import BasicNewsRecipe from calibre.web.feeds.news import BasicNewsRecipe
def has_cls(x):
return dict(attrs={'class':lambda cls: cls and x in cls.split()})
class AlJazeera(BasicNewsRecipe): class AlJazeera(BasicNewsRecipe):
title = 'Al Jazeera in English' title = 'Al Jazeera in English'
__author__ = 'Darko Miletic' __author__ = 'Darko Miletic'
@ -17,7 +20,6 @@ class AlJazeera(BasicNewsRecipe):
oldest_article = 2 oldest_article = 2
max_articles_per_feed = 100 max_articles_per_feed = 100
no_stylesheets = True no_stylesheets = True
encoding = 'iso-8859-1'
use_embedded_content = False use_embedded_content = False
extra_css = """ extra_css = """
body{font-family: Arial,sans-serif} body{font-family: Arial,sans-serif}
@ -25,23 +27,19 @@ class AlJazeera(BasicNewsRecipe):
#dvArticleDate{font-size: small; color: #999999} #dvArticleDate{font-size: small; color: #999999}
""" """
conversion_options = { conversion_options = {
'comment' : description 'comment' : description , 'tags' : category ,
, 'tags' : category 'publisher' : publisher , 'language' : language
, 'publisher' : publisher }
, 'language' : language
}
keep_only_tags = [ keep_only_tags = [
dict(attrs={'id':['DetailedTitle','ctl00_cphBody_dvSummary','dvArticleDate']}) dict(id='main-story'),
,dict(name='td',attrs={'class':'DetailedSummary'}) ]
]
remove_tags = [ remove_tags = [
dict(name=['object','link','table','meta','base','iframe','embed']) has_cls('MoreOnTheStory'), has_cls('ArticleBottomToolbar'), dict(smtitle="ShowMore"),
,dict(name='td', attrs={'class':['MostActiveDescHeader','MostActiveDescBody']}) dict(name=['object','link','table','meta','base','iframe','embed']),
] ]
feeds = [(u'AL JAZEERA ENGLISH (AJE)', u'http://english.aljazeera.net/Services/Rss/?PostingId=2007731105943979989' )] feeds = [(u'Al Jazeera English', u'http://english.aljazeera.net/Services/Rss/?PostingId=2007731105943979989')]
def get_article_url(self, article): def get_article_url(self, article):
artlurl = article.get('link', None) artlurl = article.get('link', None)
@ -54,15 +52,14 @@ class AlJazeera(BasicNewsRecipe):
del item['face'] del item['face']
td = soup.find('td',attrs={'class':'DetailedSummary'}) td = soup.find('td',attrs={'class':'DetailedSummary'})
if td: if td:
td.name = 'div' td.name = 'div'
spn = soup.find('span',attrs={'id':'DetailedTitle'}) spn = soup.find('span',attrs={'id':'DetailedTitle'})
if spn: if spn:
spn.name='h1' spn.name='h1'
for itm in soup.findAll('span', attrs={'id':['dvArticleDate','ctl00_cphBody_lblDate']}): for itm in soup.findAll('span', attrs={'id':['dvArticleDate','ctl00_cphBody_lblDate']}):
itm.name = 'div' itm.name = 'div'
for alink in soup.findAll('a'): for alink in soup.findAll('a'):
if alink.string is not None: if alink.string is not None:
tstr = alink.string tstr = alink.string
alink.replaceWith(tstr) alink.replaceWith(tstr)
return soup return soup