Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-31 14:33:54 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update wsj.recipe

Browse Source 
fix cover-changed source
This commit is contained in:
unkn0w7n 2025-07-27 11:07:14 +05:30
parent 612aa3d81a
commit 5aa877f6dd
1 changed files with 23 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
recipes/wsj.recipe
View File

@ -164,26 +164,32 @@ class WSJ(BasicNewsRecipe):
return soup return soup
def _download_cover(self): def _download_cover(self):
d = self.recipe_specific_options.get('date') import os
if not (d and isinstance(d, str)): from contextlib import closing
import os
from contextlib import closing
from calibre.utils.img import save_cover_data_to from calibre.utils.img import save_cover_data_to
br = browser() br = browser()
raw = br.open('https://www.frontpages.com/the-wall-street-journal/') dt = self.recipe_specific_options.get('date')
if (dt and isinstance(dt, str)):
d, m, y = dt.split('-')
cu = f'https://www.wsj.com/public/resources/documents/WSJNewsPaper-{int(m)}-{int(d)}-{y}.jpg'
else:
raw = br.open('https://wsjtodaysedition.cmail19.com/t/d-e-suujjg-thithlkhlr-r/')
soup = BeautifulSoup(raw.read()) soup = BeautifulSoup(raw.read())
cu = ( cu = soup.find(
'https://www.frontpages.com' 'img',
+ soup.find('img', attrs={'id': 'giornale-img'})['src'] attrs={
) 'class': 'responsive-img',
self.report_progress(1, _('Downloading cover from %s') % cu) 'src': lambda x: x and 'WSJNewsPaper' in x
with closing(br.open(cu, timeout=self.timeout)) as r: }
cdata = r.read() )['src']
cpath = os.path.join(self.output_dir, 'cover.jpg') self.report_progress(1, _('Downloading cover from %s') % cu)
save_cover_data_to(cdata, cpath) with closing(br.open(cu, timeout=self.timeout)) as r:
self.cover_path = cpath cdata = r.read()
cpath = os.path.join(self.output_dir, 'cover.jpg')
save_cover_data_to(cdata, cpath)
self.cover_path = cpath
def get_browser(self, *args, **kw): def get_browser(self, *args, **kw):
br = BasicNewsRecipe.get_browser(self, *args, **kw) br = BasicNewsRecipe.get_browser(self, *args, **kw)