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 various Uruguayan recipes

Browse Source 
Merge branch 'Uruguay' of https://github.com/carlosalvesuy/calibre
This commit is contained in:
Kovid Goyal 2018-02-14 02:24:37 +05:30
commit f8a298cef9
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
7 changed files with 81 additions and 137 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
recipes/10minutos.recipe
View File

@ -1,17 +1,17 @@
#!/usr/bin/env python2 #!/usr/bin/env python2
## ##
# Title: Diario 10minutos.com.uy News and Sports Calibre Recipe # Title: Diario 10minutos.com.uy News and Sports Calibre Recipe
# Contact: Carlos Alves - <carlos@carlosalves.info> # Contact: Carlos Alves - <carlosalves90@gmail.com>
## ##
# License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html # License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html
# Copyright: Carlos Alves - <carlos@carlosalves.info> # Copyright: Carlos Alves - <carlosalves90@gmail.com>
## ##
# Written: September 2013 # Written: September 2013
# Last Edited: 2016-01-11 # Last Edited: 2018-02-13
## ##
__license__ = 'GPL v3' __license__ = 'GPL v3'
__author__ = '2016, Carlos Alves <carlos@carlosalves.info>' __author__ = '2016, Carlos Alves <carlosalves90@gmail.com>'
''' '''
10minutos.com.uy 10minutos.com.uy
''' '''
@ -44,17 +44,20 @@ class General(BasicNewsRecipe):
] ]
extra_css = ''' extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;} h1{font-family: Georgia,"Times New Roman",Times,serif}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;} h3{font-family: Georgia,"Times New Roman",Times,serif}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;} h2{font-family: Georgia,"Times New Roman",Times,serif}
p {font-family:Arial,Helvetica,sans-serif;} p{font-family: Verdana,Arial,Helvetica,sans-serif}
body{font-family: Verdana,Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block;}
''' '''
feeds = [ feeds = [
(u'Articulos', u'http://10minutos.com.uy/?feed=rss2') (u'Articulos', u'http://10minutos.com.uy/?feed=rss2')
] ]
def get_cover_url(self): def get_cover_url(self):
return 'http://10minutos.com.uy/a/img/logo.png' return None
def preprocess_html(self, soup): def preprocess_html(self, soup):
for item in soup.findAll(style=True): for item in soup.findAll(style=True):

22
recipes/180.recipe
View File

@ -1,6 +1,6 @@
#!/usr/bin/env python2 #!/usr/bin/env python2
## ##
# Last Edited: 2016-01-11 Carlos Alves <carlos@carlosalves.info> # Last Edited: 2018-02-13 Carlos Alves <carlosalves90@gmail.com>
## ##
__license__ = 'GPL v3' __license__ = 'GPL v3'
@ -28,22 +28,26 @@ class Noticias(BasicNewsRecipe):
max_articles_per_feed = 100 max_articles_per_feed = 100
remove_tags_after = dict(name='article') remove_tags_after = dict(name='article')
keep_only_tags = [ keep_only_tags = [
dict(name='h3', attrs={'class': 'title'}), dict(name='div', attrs={'class': 'nota'}),
dict(name='div', attrs={'class': 'copete'}), dict(name='h3',),
dict(name='article', attrs={'class': 'texto'}) dict(name='h4',),
dict(name='article',)
] ]
remove_tags = [ remove_tags = [
dict(name=['object', 'link']) dict(name='div', attrs={'class': 'items'})
] ]
remove_attributes = ['width', 'height', 'style', 'font', 'color'] remove_attributes = ['width', 'height', 'style', 'font', 'color']
extra_css = ''' extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;} h1{font-family: Georgia,"Times New Roman",Times,serif}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;} h3{font-family: Georgia,"Times New Roman",Times,serif}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;} h2{font-family: Georgia,"Times New Roman",Times,serif}
p {font-family:Arial,Helvetica,sans-serif;} p{font-family: Verdana,Arial,Helvetica,sans-serif}
body{font-family: Verdana,Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block;}
''' '''
feeds = [ feeds = [
(u'Titulares', u'http://www.180.com.uy/feed.php') (u'Titulares', u'http://www.180.com.uy/feed.php')
] ]

21
recipes/diario_el_pueblo.recipe
View File

@ -1,16 +1,16 @@
#!/usr/bin/env python2 #!/usr/bin/env python2
## ##
# Title: Diario El Pueblo News and Sports Calibre Recipe # Title: Diario El Pueblo News and Sports Calibre Recipe
# Contact: Carlos Alves - <carlos@carlosalves.info> # Contact: Carlos Alves - <carlosalves90@gmail.com>
## ##
# License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html # License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html
# Copyright: Carlos Alves - <carlos@carlosalves.info> # Copyright: Carlos Alves - <carlosalves90@gmail.com>
## ##
# Written: September 2013 # Written: September 2013
# Last Edited: 2016-01-11 # Last Edited: 2018-02-13
## ##
__license__ = 'GPL v3' __license__ = 'GPL v3'
__author__ = '2013, Carlos Alves <carlos@carlosalves.info>' __author__ = '2013, Carlos Alves <carlosalves90@gmail.com>'
''' '''
diarioelpueblo.com.uy diarioelpueblo.com.uy
@ -46,17 +46,20 @@ class General(BasicNewsRecipe):
] ]
extra_css = ''' extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;} h1{font-family: Georgia,"Times New Roman",Times,serif}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;} h3{font-family: Georgia,"Times New Roman",Times,serif}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;} h2{font-family: Georgia,"Times New Roman",Times,serif}
p {font-family:Arial,Helvetica,sans-serif;} p{font-family: Verdana,Arial,Helvetica,sans-serif}
body{font-family: Verdana,Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block;}
''' '''
feeds = [ feeds = [
(u'Articulos', u'http://www.diarioelpueblo.com.uy/feed') (u'Articulos', u'http://www.diarioelpueblo.com.uy/feed')
] ]
def get_cover_url(self): def get_cover_url(self):
return 'http://www.diarioelpueblo.com.uy/wp-content/uploads/2013/06/Cabezal_Web1.jpg' return None
def preprocess_html(self, soup): def preprocess_html(self, soup):
for item in soup.findAll(style=True): for item in soup.findAll(style=True):

15
recipes/el_observador.recipe
View File

@ -1,6 +1,6 @@
#!/usr/bin/env python2 #!/usr/bin/env python2
## ##
# Last Edited: 2016-01-11 Carlos Alves <carlos@carlosalves.info> # Last Edited: 2018-02-13 Carlos Alves <carlosalves90@gmail.com>
## ##
__license__ = 'GPL v3' __license__ = 'GPL v3'
@ -44,17 +44,20 @@ class Noticias(BasicNewsRecipe):
remove_attributes = ['width', 'height', 'style', 'font', 'color'] remove_attributes = ['width', 'height', 'style', 'font', 'color']
extra_css = ''' extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;} h1{font-family: Georgia,"Times New Roman",Times,serif}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;} h3{font-family: Georgia,"Times New Roman",Times,serif}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;} h2{font-family: Georgia,"Times New Roman",Times,serif}
p {font-family:Arial,Helvetica,sans-serif;} p{font-family: Verdana,Arial,Helvetica,sans-serif}
body{font-family: Verdana,Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block;}
''' '''
feeds = [ feeds = [
(u'Portada', u'http://www.elobservador.com.uy/rss/home.xml'), (u'Portada', u'http://www.elobservador.com.uy/rss/home.xml'),
] ]
def get_cover_url(self): def get_cover_url(self):
return 'http://css.elobservador.com.uy/css/181/images/logo.svg' return None
def preprocess_html(self, soup): def preprocess_html(self, soup):
for item in soup.findAll(style=True): for item in soup.findAll(style=True):

36
recipes/el_pais_uy.recipe
View File

@ -1,6 +1,6 @@
#!/usr/bin/env python2 #!/usr/bin/env python2
## ##
# Last Edited: 2016-01-11 Carlos Alves <carlos@carlosalves.info> # Last Edited: 2018-02-13 Carlos Alves <carlosalves90@gmail.com>
## ##
__license__ = 'GPL v3' __license__ = 'GPL v3'
@ -28,12 +28,11 @@ class General(BasicNewsRecipe):
no_stylesheets = True no_stylesheets = True
oldest_article = 2 oldest_article = 2
max_articles_per_feed = 200 max_articles_per_feed = 20
keep_only_tags = [ keep_only_tags = [
dict(name='div', attrs={'class': 'title'}), dict(name='h1', attrs={'class': 'title'}),
dict(name='div', attrs={'class': 'pc'}), dict(name='div', attrs={'class': 'composite-captioned-image'}),
dict(name='div', attrs={'class': 'image'}), dict(name='div', attrs={'class': 'content-modules'})
dict(name='div', attrs={'class': 'article-content'})
] ]
conversion_options = { conversion_options = {
@ -49,32 +48,27 @@ class General(BasicNewsRecipe):
dict(name=['object', 'table'])] dict(name=['object', 'table'])]
extra_css = ''' extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;} h1{font-family: Georgia,"Times New Roman",Times,serif}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;} h3{font-family: Georgia,"Times New Roman",Times,serif}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;} h2{font-family: Georgia,"Times New Roman",Times,serif}
p {font-family:Arial,Helvetica,sans-serif;} p{font-family: Verdana,Arial,Helvetica,sans-serif}
body{font-family: Verdana,Arial,Helvetica,sans-serif} body{font-family: Verdana,Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block;} img{margin-bottom: 0.4em; display:block;}
''' '''
feeds = [ feeds = [
(u'Ultimo Momento', (u'Ultimo Momento',
u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=umomento'), u'https://www.elpais.com.uy/rss/')
(u'Editorial', u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=editorial'),
(u'Nacional', u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=nacional'),
(u'Internacional', u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=internacional'),
(u'Espectaculos', u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=espectaculos'),
(u'Ciudades', u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=ciudades'),
(u'Economia', u'http://www.elpais.com.uy/formatos/rss/index.asp?seccion=economia')
] ]
def get_cover_url(self): def get_cover_url(self):
cover_url = None cover_url = None
index = 'http://www.elpais.com.uy/impresa/' index = 'https://www.elpais.com.uy/impresa/'
soup = self.index_to_soup(index) soup = self.index_to_soup(index)
link_item = soup.find('div', attrs={'class': 'box-dotted-white'}) link_item = soup.find('a', attrs={'class': 'page-link link-module'})
print link_item #print link_item
if link_item: if link_item:
cover_url = 'http://www.elpais.com.uy' + link_item.img['src'] cover_url = 'https://www.elpais.com.uy' + link_item.get('href')
return cover_url return cover_url
def preprocess_html(self, soup): def preprocess_html(self, soup):

66
recipes/lr21.recipe
View File

@ -1,66 +0,0 @@
#!/usr/bin/env python2
##
# Title: lr21.com.uy News, Sports, and Blog Calibre Recipe
# Contact: Carlos Alves - <carlos@carlosalves.info>
##
# License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html
# Copyright: Carlos Alves - <carlos@carlosalves.info>
##
# Written: January 2016
# Last Edited: 2016-01-11
##
__license__ = 'GPL v3'
__author__ = '2016, Carlos Alves <carlos@carlosalves.info>'
'''
lr21.com.uy
'''
from calibre.web.feeds.news import BasicNewsRecipe
class General(BasicNewsRecipe):
title = 'LaRed21'
__author__ = 'Carlos Alves'
description = 'LaRed21 Diario Digital'
tags = 'soccer, futbol, news'
language = 'es_UY'
timefmt = '[%a, %d %b, %Y]'
use_embedded_content = False
recursion = 5
encoding = None
remove_javascript = True
no_stylesheets = True
oldest_article = 10
max_articles_per_feed = 100
keep_only_tags = [
dict(name='h1', attrs={'id': 'article-title'}),
dict(name='h2', attrs={'class': 'copete'}),
dict(name='div', attrs={'class': 'article-content clear bottom-1'})
]
remove_tags = [
dict(name='div', attrs={
'class': ['br', 'hr', 'titlebar', 'navigation']}),
dict(name='dl', attrs={'class': 'gallery-item'}),
dict(name=['object', 'link'])
]
extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;}
p {font-family:Arial,Helvetica,sans-serif;}
'''
feeds = [
(u'LaRed21 Diario Digital', u'http://www.lr21.com.uy/feed')
]
def get_cover_url(self):
pass
def preprocess_html(self, soup):
for item in soup.findAll(style=True):
del item['style']
return soup

23
recipes/padreydecano.recipe
View File

@ -1,16 +1,16 @@
#!/usr/bin/env python2 #!/usr/bin/env python2
## ##
# Title: Padreydecano.com News, Sports, and Blog Calibre Recipe # Title: Padreydecano.com News, Sports, and Blog
# Contact: Carlos Alves - <carlos@carlosalves.info> # Contact: Carlos Alves - <carlosalves90@gmail.com>
## ##
# License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html # License: GNU General Public License v3 - http://www.gnu.org/copyleft/gpl.html
# Copyright: Carlos Alves - <carlos@carlosalves.info> # Copyright: Carlos Alves - <carlosalves90@gmail.com>
## ##
# Written: September 2013 # Written: September 2013
# Last Edited: 2016-01-11 # Last Edited: 2018-02-13
## ##
__license__ = 'GPL v3' __license__ = 'GPL v3'
__author__ = '2016, Carlos Alves <carlos@carlosalves.info>' __author__ = '2016, Carlos Alves <carlosalves90@gmail.com>'
''' '''
padreydecano.com padreydecano.com
@ -47,18 +47,21 @@ class General(BasicNewsRecipe):
] ]
extra_css = ''' extra_css = '''
h1{font-family:Geneva, Arial, Helvetica, sans-serif;color:#154B7A;} h1{font-family: Georgia,"Times New Roman",Times,serif}
h3{font-size: 14px;color:#999999; font-family:Geneva, Arial, Helvetica, sans-serif;font-weight: bold;} h3{font-family: Georgia,"Times New Roman",Times,serif}
h2{color:#666666; font-family:Geneva, Arial, Helvetica, sans-serif;font-size:small;} h2{font-family: Georgia,"Times New Roman",Times,serif}
p {font-family:Arial,Helvetica,sans-serif;} p{font-family: Verdana,Arial,Helvetica,sans-serif}
body{font-family: Verdana,Arial,Helvetica,sans-serif}
img{margin-bottom: 0.4em; display:block;}
''' '''
feeds = [ feeds = [
(u'Padre y Decano | Club Atlético Peñarol', (u'Padre y Decano | Club Atlético Peñarol',
u'http://www.padreydecano.com/cms/feed/') u'http://www.padreydecano.com/cms/feed/')
] ]
def get_cover_url(self): def get_cover_url(self):
return 'http://www.padreydecano.com/cms/wp-content/uploads/2011/09/editorial.jpg' return None
def preprocess_html(self, soup): def preprocess_html(self, soup):
for item in soup.findAll(style=True): for item in soup.findAll(style=True):