Mirrors/ArchiveBox
2
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox synced 2024-11-10 06:34:16 +00:00
Code Issues Projects Releases Packages Wiki Activity

handle new wallabag export format with newlines mid-tag attributes

Browse source
This commit is contained in:
Nick Sweeting 2022-05-09 19:07:42 -07:00
parent 808ae1a351
commit acd53c854d
1 changed files with 11 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
archivebox/parsers/wallabag_atom.py
View file

@ -34,13 +34,19 @@ def parse_wallabag_atom_export(rss_file: IO[str], **_kwargs) -> Iterable[Link]:
trailing_removed = entry.split('</entry>', 1)[0]
leading_removed = trailing_removed.strip()
rows = leading_removed.split('\n')
splits_fixed = leading_removed.replace('"\n href="', '" href="')
rows = splits_fixed.split('\n')
def get_row(key):
return [r.strip() for r in rows if r.strip().startswith('<{}'.format(key))][0]
def get_row(prefix):
return [
row.strip()
for row in rows
if row.strip().startswith('<{}'.format(prefix))
][0]
title = str_between(get_row('title'), '<title><![CDATA[', ']]></title>').strip()
url = str_between(get_row('link rel="via"'), '<link rel="via">', '</link>')
url_inside_link = str_between(get_row('link rel="via"'), '<link rel="via">', '</link>')
url_inside_attr = str_between(get_row('link rel="via"'), 'href="', '"/>')
ts_str = str_between(get_row('published'), '<published>', '</published>')
time = datetime.strptime(ts_str, "%Y-%m-%dT%H:%M:%S%z")
try:
@ -49,7 +55,7 @@ def parse_wallabag_atom_export(rss_file: IO[str], **_kwargs) -> Iterable[Link]:
tags = None
yield Link(
url=htmldecode(url),
url=htmldecode(url_inside_attr or url_inside_link),
timestamp=str(time.timestamp()),
title=htmldecode(title) or None,
tags=tags or '',