Historia and Buctaras by Silviu Coatara

This commit is contained in:
Kovid Goyal 2011-02-27 10:33:24 -07:00
parent 4af08c8fa9
commit b8f08346b5
4 changed files with 107 additions and 0 deletions

Binary file not shown.

After

Width:  |  Height:  |  Size: 765 B

Binary file not shown.

After

Width:  |  Height:  |  Size: 521 B

View File

@ -0,0 +1,56 @@
# -*- coding: utf-8 -*-
#!/usr/bin/env python
__license__ = 'GPL v3'
__copyright__ = u'2011, Silviu Cotoar\u0103'
'''
bucataras.ro
'''
from calibre.web.feeds.news import BasicNewsRecipe
class Bucataras(BasicNewsRecipe):
title = u'Bucataras'
__author__ = u'Silviu Cotoar\u0103'
description = ''
publisher = 'Bucataras'
oldest_article = 5
language = 'ro'
max_articles_per_feed = 100
no_stylesheets = True
use_embedded_content = False
category = 'Ziare,Bucatarie,Retete'
encoding = 'utf-8'
cover_url = 'http://www.bucataras.ro/templates/default/images/pink/logo.jpg'
conversion_options = {
'comments' : description
,'tags' : category
,'language' : language
,'publisher' : publisher
}
keep_only_tags = [
dict(name='h1', attrs={'class':'titlu'})
, dict(name='div', attrs={'class':'contentL'})
, dict(name='div', attrs={'class':'contentBottom'})
]
remove_tags = [
dict(name='div', attrs={'class':['sociale']})
, dict(name='div', attrs={'class':['contentR']})
, dict(name='a', attrs={'target':['_self']})
, dict(name='div', attrs={'class':['comentarii']})
]
remove_tags_after = [
dict(name='div', attrs={'class':['comentarii']})
]
feeds = [
(u'Feeds', u'http://www.bucataras.ro/rss/retete/')
]
def preprocess_html(self, soup):
return self.adeify_images(soup)

View File

@ -0,0 +1,51 @@
# -*- coding: utf-8 -*-
#!/usr/bin/env python
__license__ = 'GPL v3'
__copyright__ = u'2011, Silviu Cotoar\u0103'
'''
historia.ro
'''
from calibre.web.feeds.news import BasicNewsRecipe
class HistoriaRo(BasicNewsRecipe):
title = u'Historia'
__author__ = u'Silviu Cotoar\u0103'
description = ''
publisher = 'Historia'
oldest_article = 5
language = 'ro'
max_articles_per_feed = 100
no_stylesheets = True
use_embedded_content = False
category = 'Ziare,Reviste,Istorie'
encoding = 'utf-8'
cover_url = 'http://www.historia.ro/sites/all/themes/historia/images/historia.png'
conversion_options = {
'comments' : description
,'tags' : category
,'language' : language
,'publisher' : publisher
}
keep_only_tags = [
dict(name='div', attrs={'class':'c_antet_title'})
, dict(name='a', attrs={'class':'overlaybox'})
, dict(name='div', attrs={'class':'art_content'})
]
remove_tags = [
dict(name='div', attrs={'class':['fl_left']})
, dict(name='div', attrs={'id':['article_toolbar']})
, dict(name='div', attrs={'class':['zoom_cont']})
]
feeds = [
(u'Feeds', u'http://www.historia.ro/rss.xml')
]
def preprocess_html(self, soup):
return self.adeify_images(soup)