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 Metro News NL

Browse Source 
Fixes #1186861 [Update of Metronews-NL](https://bugs.launchpad.net/calibre/+bug/1186861)
This commit is contained in:
Kovid Goyal 2013-06-03 10:52:40 +05:30
parent f69504a8f8
commit b598ce3c17
1 changed files with 9 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
recipes/metro_news_nl.recipe
View File

@ -39,6 +39,8 @@ from BeautifulSoup import BeautifulSoup
Version 1.9.4 19-04-2013
Added regex filter for mailto
Updated for new layout of metro-site
Version 1.9.5 28-05-2013
Added some extra id's and classes to remove
'''
class AdvancedUserRecipe1306097511(BasicNewsRecipe):
@ -46,7 +48,7 @@ class AdvancedUserRecipe1306097511(BasicNewsRecipe):
oldest_article = 1.2
max_articles_per_feed = 25
__author__ = u'DrMerry'
description = u'Metro Nederland v1.9.4 2013-04-19'
description = u'Metro Nederland v1.9.5 2013-05-28, Download nieuws van de Nederlandse editie van de krant Metro'
language = u'nl'
simultaneous_downloads = 5
masthead_url = 'http://blog.metronieuws.nl/wp-content/themes/metro/images/header.gif'
@ -70,7 +72,7 @@ class AdvancedUserRecipe1306097511(BasicNewsRecipe):
#(re.compile(r'<(a |/a)[^>]*>', re.DOTALL|re.IGNORECASE),lambda match:'')
#(re.compile('(</?)h2', re.DOTALL|re.IGNORECASE),lambda match:'\1em')
]
remove_tags_before= dict(id='subwrapper')
remove_tags_after = dict(name='div', attrs={'class':['body-area','article-main-area']})
#name='div', attrs={'class':['subwrapper']})]
@ -80,13 +82,13 @@ class AdvancedUserRecipe1306097511(BasicNewsRecipe):
remove_tags = [
dict(name=['iframe','script','noscript','style']),
dict(name='div', attrs={'class':['aside clearfix','aside clearfix middle-col-line','comments','share-tools','article-right-column','column-4-5','column-1-5','ad-msg','col-179 ','col-373 ','clear','ad','navigation',re.compile('share-tools(-top)?'),'tools','metroCommentFormWrap','article-tools-below-title','related-links','padding-top-15',re.compile('^promo.*?$'),'teaser-component',re.compile('fb(-comments|_iframe_widget)'),'promos','header-links','promo-2']}),
dict(id=['article-2','googleads','column-1-5-bottom','column-4-5',re.compile('^ad(\d+|adcomp.*?)?$'),'adadcomp-4','margin-5','sidebar',re.compile('^article-\d'),'comments','gallery-1','sharez_container','ts-container','topshares','ts-title']),
dict(name='div', attrs={'class':['fact-related-box','aside clearfix','aside clearfix middle-col-line','comments','share-tools','article-right-column','column-4-5','column-1-5','ad-msg','col-179 ','col-373 ','clear','ad','navigation',re.compile('share-tools(-top)?'),'tools','metroCommentFormWrap','article-tools-below-title','related-links','padding-top-15',re.compile('^promo.*?$'),'teaser-component',re.compile('fb(-comments|_iframe_widget)'),'promos','header-links','promo-2']}),
dict(id=['super-carousel','article-2','googleads','column-1-5-bottom','column-4-5',re.compile('^ad(\d+|adcomp.*?)?$'),'adadcomp-4','margin-5','sidebar',re.compile('^article-\d'),'comments','gallery-1','sharez_container','ts-container','topshares','ts-title']),
dict(name='a', attrs={'name':'comments'}),
#dict(name='div', attrs={'data-href'}),
dict(name='img', attrs={'class':'top-line','title':'volledig scherm'}),
dict(attrs={'style':re.compile('^(.*(display\s?:\s?none|img-mask|white)\s?;?.*)$'),'title':'volledig scherm'})]
'''removed by before/after:
id:
column-1-5-top,'hidden_div','footer',
@ -182,7 +184,7 @@ class MerryProcess(BeautifulSoup):
except:
pass
return soup
def moveTitleAndAuthor(self, soup):
moveitem = soup.h1
pubdate = soup.find(id="date")
@ -218,4 +220,4 @@ class MerryProcess(BeautifulSoup):
self.removeArrayOfTags(emptytags)
#recursive in case removing empty tag creates new empty tag
self.removeEmptyTags(soup, run=run)
return soup
return soup