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 horizons.recipe

Browse Source 
remove empty p tags
This commit is contained in:
unkn0w7n 2025-03-24 11:06:11 +05:30
parent d72d746c1a
commit 1526d0d06f
1 changed files with 20 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
recipes/horizons.recipe
View File

@ -1,8 +1,8 @@
#!/usr/bin/env python #!/usr/bin/env python
# vim:fileencoding=utf-8 # vim:fileencoding=utf-8
''' """
https://www.cirsd.org/en/horizons https://www.cirsd.org/en/horizons
''' """
from calibre.web.feeds.news import BasicNewsRecipe, classes from calibre.web.feeds.news import BasicNewsRecipe, classes
@ -10,10 +10,12 @@ from calibre.web.feeds.news import BasicNewsRecipe, classes
class horizons(BasicNewsRecipe): class horizons(BasicNewsRecipe):
title = 'Horizons' title = 'Horizons'
__author__ = 'unkn0wn' __author__ = 'unkn0wn'
description = (' Horizons – Journal of International Relations and Sustainable Development.' description = (
' Horizons – Journal of International Relations and Sustainable Development.'
' Horizons serves as a high-level platform for influential voices from around the world to' ' Horizons serves as a high-level platform for influential voices from around the world to'
' provide informed analysis and conduct reasoned exchanges on the full spectrum of issues' ' provide informed analysis and conduct reasoned exchanges on the full spectrum of issues'
' that shape international developments.') ' that shape international developments.'
)
no_stylesheets = True no_stylesheets = True
use_embedded_content = False use_embedded_content = False
encoding = 'utf-8' encoding = 'utf-8'
@ -22,13 +24,12 @@ class horizons(BasicNewsRecipe):
masthead_url = 'https://www.cirsd.org/bundles/olpublic/images/horizons-logo.jpg' masthead_url = 'https://www.cirsd.org/bundles/olpublic/images/horizons-logo.jpg'
ignore_duplicate_articles = {'url'} ignore_duplicate_articles = {'url'}
extra_css = 'em{color:#404040;}' extra_css = 'em{color:#404040;}'
simultaneous_downloads = 1
keep_only_tags = [ keep_only_tags = [dict(name='div', attrs={'class': 'article'})]
dict(name='div', attrs={'class':'article'})
]
remove_tags = [ remove_tags = [
classes('back-link'), classes('back-link'),
dict(name='div', attrs={'class':'single-post-footer'}) dict(name='div', attrs={'class': 'single-post-footer'}),
] ]
recipe_specific_options = { recipe_specific_options = {
@ -38,6 +39,9 @@ class horizons(BasicNewsRecipe):
} }
} }
def preprocess_raw_html(self, raw, *a):
return raw.replace('<p>&nbsp;</p>', '')
def get_browser(self): def get_browser(self):
return BasicNewsRecipe.get_browser(self, verify_ssl_certificates=False) return BasicNewsRecipe.get_browser(self, verify_ssl_certificates=False)
@ -47,7 +51,7 @@ class horizons(BasicNewsRecipe):
url = d url = d
else: else:
soup = self.index_to_soup('https://www.cirsd.org/en/horizons') soup = self.index_to_soup('https://www.cirsd.org/en/horizons')
a = soup.findAll('a', href=True, attrs={'class':'horizon-gallery-box'})[0] # use 1 for previous edition a = soup.find('a', href=True, attrs={'class':'horizon-gallery-box'})
url = a['href'] url = a['href']
if url.startswith('/'): if url.startswith('/'):
url = 'https://www.cirsd.org' + url url = 'https://www.cirsd.org' + url
@ -74,10 +78,7 @@ class horizons(BasicNewsRecipe):
if span: if span:
desc = self.tag_to_string(span).strip() desc = self.tag_to_string(span).strip()
self.log('\t', title, '\n\t', desc, '\n\t\t', url) self.log('\t', title, '\n\t', desc, '\n\t\t', url)
articles.append({ articles.append({'title': title, 'url': url, 'description': desc})
'title': title,
'url': url,
'description': desc})
if articles: if articles:
feeds.append((secname, articles)) feeds.append((secname, articles))
return feeds return feeds