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 polter.pl recipe

Browse Source
This commit is contained in:
Tomasz Długosz 2018-10-14 23:39:44 +02:00
parent e9478054a8
commit e8f715d3bf
1 changed files with 6 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
recipes/polter_pl.recipe
View File

@ -21,7 +21,8 @@ class Polter(BasicNewsRecipe):
ignore_duplicate_articles = {'title', 'url'} ignore_duplicate_articles = {'title', 'url'}
keep_only_tags = [dict(attrs={'class': 'boxcontent'})] keep_only_tags = [dict(attrs={'class': 'boxcontent'})]
remove_tags = [dict(id='komentarze')] remove_tags = [dict(id='komentarze'),
dict(name='div',attrs={'class':'ostatnieArtykuly'})]
remove_tags_after = dict(id='komentarze') remove_tags_after = dict(id='komentarze')
feeds = [ feeds = [
@ -36,8 +37,7 @@ class Polter(BasicNewsRecipe):
(u'Gry planszowe', 'http://planszowki.polter.pl/wiesci,rss.html'), (u'Gry planszowe', 'http://planszowki.polter.pl/wiesci,rss.html'),
(u'Gry PC', 'http://gry.polter.pl/wiesci,rss.html'), (u'Gry PC', 'http://gry.polter.pl/wiesci,rss.html'),
(u'Gry konsolowe', 'http://konsole.polter.pl/wiesci,rss.html'), (u'Gry konsolowe', 'http://konsole.polter.pl/wiesci,rss.html'),
(u'Konwenty', 'http://konwenty.polter.pl/wiesci,rss.html'), (u'Konwenty', 'http://konwenty.polter.pl/wiesci,rss.html')]
(u'Blogi', 'http://polter.pl/blogi,rss.html')]
def preprocess_html(self, soup): def preprocess_html(self, soup):
for s in soup.findAll(attrs={'style': re.compile('float: ?left')}): for s in soup.findAll(attrs={'style': re.compile('float: ?left')}):
@ -65,3 +65,6 @@ class Polter(BasicNewsRecipe):
for r in soup.findAll(name='a', href=re.compile(r'^http://www.ceneo.pl/')): for r in soup.findAll(name='a', href=re.compile(r'^http://www.ceneo.pl/')):
r.extract() r.extract()
return soup return soup
def preprocess_raw_html(self, raw_html, url):
return raw_html.replace('<br /><br /><h3>Czytaj również</h3>', '')