mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 06:34:25 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

fix pinboard RSS parser

Browse source
This commit is contained in:
Nick Sweeting 2019-03-19 18:44:22 -04:00
parent 61ec4971e9
commit f9a7c530b9
2 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

8
archivebox/parse.py
View file

@ -228,9 +228,9 @@ def parse_pinboard_rss_export(rss_file):
items = root.findall("{http://purl.org/rss/1.0/}item") items = root.findall("{http://purl.org/rss/1.0/}item")
for item in items: for item in items:
url = item.find("{http://purl.org/rss/1.0/}link").text url = item.find("{http://purl.org/rss/1.0/}link").text
tags = item.find("{http://purl.org/dc/elements/1.1/}subject").text tags = item.find("{http://purl.org/dc/elements/1.1/}subject").text if item.find("{http://purl.org/dc/elements/1.1/}subject") else None
title = item.find("{http://purl.org/rss/1.0/}title").text.strip() title = item.find("{http://purl.org/rss/1.0/}title").text.strip() if item.find("{http://purl.org/rss/1.0/}title").text.strip() else None
ts_str = item.find("{http://purl.org/dc/elements/1.1/}date").text ts_str = item.find("{http://purl.org/dc/elements/1.1/}date").text if item.find("{http://purl.org/dc/elements/1.1/}date").text else None
# = 🌈🌈🌈🌈 # = 🌈🌈🌈🌈
# = 🌈🌈🌈🌈 # = 🌈🌈🌈🌈
# = 🏆🏆🏆🏆 # = 🏆🏆🏆🏆
@ -243,7 +243,7 @@ def parse_pinboard_rss_export(rss_file):
info = { info = {
'url': url, 'url': url,
'timestamp': str(time.timestamp()), 'timestamp': str(time.timestamp()),
'tags': tags, 'tags': tags or '',
'title': title or None, 'title': title or None,
'sources': [rss_file.name], 'sources': [rss_file.name],
} }

2
archivebox/util.py
View file

@ -413,7 +413,7 @@ def derived_link_info(link):
'is_archived': os.path.exists(os.path.join( 'is_archived': os.path.exists(os.path.join(
ARCHIVE_DIR, ARCHIVE_DIR,
link['timestamp'], link['timestamp'],
wget_output_path(link) or domain(url) domain(url),
)), )),
'num_outputs': len([entry for entry in link['latest'].values() if entry]) if 'latest' in link else 0, 'num_outputs': len([entry for entry in link['latest'].values() if entry]) if 'latest' in link else 0,
} }