Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-08 10:44:09 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update tls_mag.recipe

Browse Source 
fix for missing title and lead image.
This commit is contained in:
unkn0w7n 2024-09-06 15:29:53 +05:30
parent 961ef505d1
commit 741aed541a
1 changed files with 10 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
recipes/tls_mag.recipe
View File

@ -93,11 +93,9 @@ class tls(BasicNewsRecipe):
self.log(' ', title, '\n\t', desc, '\n\t', url)
return ({ 'title': title, 'description': desc, 'url': url })
def print_version(self, url):
return 'https://www.the-tls.co.uk/wp-json/tls/v2/single-article/' + get_id(url)
def preprocess_raw_html(self, raw, *a):
data = json.loads(raw)
pg = re.search(r'var tlsPageObject = ({.+)', raw).group(1)
data = json.JSONDecoder().raw_decode(pg)[0]
if 'articleIntroPrimary' in data:
prim = data['articleIntroPrimary']
title = '<h1>' + prim['headline'] + '</h1>\n'
@ -121,7 +119,7 @@ class tls(BasicNewsRecipe):
prim = title = desc = label = auth = lede = ''
bks = ''
if data['bookdetails']:
if 'bookdetails' in data and data['bookdetails']:
bks += '<br>'
for a in data['bookdetails']:
for x, y in a.items():
@ -132,14 +130,15 @@ class tls(BasicNewsRecipe):
bks += '<div class="det">' + y + '</div>\n'
bks += '<br>'
if 'leadimage' in data:
if 'full_image' in data['leadimage'] and data['leadimage']['full_image']:
if 'article_data_leadimage' in data:
i = data['article_data_leadimage']
if 'full_image' in i and i['full_image']:
lede = '<br><img src="{}"><div class="figc">{}</div>'.format(
data['leadimage']['full_image'] + '?w600', data['leadimage']['imagecaption'] + ' <i>' \
+ data['leadimage']['imagecredit'] + '</i>'
i['full_image'] + '?w600', i['imagecaption'] + ' <i>' \
+ i['imagecredit'] + '</i>'
)
body = data['content']
cont = self.index_to_soup('https://www.the-tls.co.uk/wp-json/tls/v2/single-article/' + data['ID'], raw=True)
body = json.loads(cont)['content']
html = '<html><body><div>' \
+ label + title + desc + auth + lede + bks + body + \