mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-14 07:04:27 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix comments in links.py

Browse source
This commit is contained in:
mlazana 2019-03-24 19:04:22 +02:00
parent 4d10568477
commit 81d846427e
1 changed files with 2 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

5
archivebox/links.py
View file

@ -39,8 +39,6 @@ def validate_links(links):
links = sorted_links(links) # deterministically sort the links based on timstamp, url links = sorted_links(links) # deterministically sort the links based on timstamp, url
links = exclude_links(links) # exclude links that are in blacklist links = exclude_links(links) # exclude links that are in blacklist
print(links)
if not links: if not links:
print('[X] No links found :(') print('[X] No links found :(')
raise SystemExit(1) raise SystemExit(1)
@ -49,7 +47,6 @@ def validate_links(links):
link['title'] = unescape(link['title'].strip()) if link['title'] else None link['title'] = unescape(link['title'].strip()) if link['title'] else None
check_link_structure(link) check_link_structure(link)
print("FINAL LIST", list(links))
return list(links) return list(links)
@ -124,7 +121,7 @@ def lowest_uniq_timestamp(used_timestamps, timestamp):
return new_timestamp return new_timestamp
def exclude_links(links): def exclude_links(links):
""" exclude links that are in blacklist""" """exclude links that are in blacklist"""
links = [link for link in links if not URL_BLACKLIST.match(link['url'])] links = [link for link in links if not URL_BLACKLIST.match(link['url'])]