iuwqdhzwezdzu

remotes/1693045480750635534/spooky-22
fireworks88 2021-10-27 11:25:18 +02:00
parent 3ac12cf03a
commit b13cd051fa
1 changed files with 4 additions and 4 deletions

View File

@ -860,22 +860,22 @@ def submit_post(v):
body += "\n\n---\n\n" body += "\n\n---\n\n"
else: body = "" else: body = ""
if new_post.url: if new_post.url:
body += f"Snapshots:\n\n* [reveddit.com](https://reveddit.com/{new_post.url})\n* [archive.org](https://web.archive.org/{new_post.url})\n* [archive.ph](https://archive.ph/?url={quote(new_post.url)}&run=1) (click to archive)\n" body += f"Snapshots:\n\n* [reveddit.com](https://reveddit.com/{new_post.url})\n* [archive.org](https://web.archive.org/{new_post.url})\n* [archive.ph](https://archive.ph/?url={quote(new_post.url)}&run=1) (click to archive)\n\n"
gevent.spawn(archiveorg, new_post.url) gevent.spawn(archiveorg, new_post.url)
# archive other urls in post # archive other urls in post
url_regex = '<a href=\"(https?://[a-z]{1,20}\.[^\"]+)\" rel=\"nofollow noopener noreferrer\" target=\"_blank\">([^\"]+)</a>' url_regex = '<a (target=\"_blank\" )?(rel=\"nofollow noopener noreferrer\" )?href=\"(https?://[a-z]{1,20}\.[^\"]+)\"( rel=\"nofollow noopener noreferrer\" target=\"_blank\")?>([^\"]+)</a>'
_body = new_post.body_html _body = new_post.body_html
print(_body) print(_body)
for url_match in re.finditer(url_regex, _body, flags=re.M|re.I): for url_match in re.finditer(url_regex, _body, flags=re.M|re.I):
href = url_match.group(1) href = url_match.group(3)
if not href: if not href:
print(f'{url_match.group(0)} skip') print(f'{url_match.group(0)} skip')
continue continue
print(href) print(href)
title = url_match.group(2) title = url_match.group(5)
body += f'**[{title}]({href})**:\n\n' body += f'**[{title}]({href})**:\n\n'
body += f'* [reveddit.com](https://reveddit.com/{href})\n' body += f'* [reveddit.com](https://reveddit.com/{href})\n'
body += f'* [archive.org](https://web.archive.org/{href})\n' body += f'* [archive.org](https://web.archive.org/{href})\n'