mirror/archivebox
1
0
Fork 0
mirror of https://github.com/pirate/ArchiveBox.git synced 2025-08-25 23:36:22 +02:00
Code Issues Releases Wiki Activity

config.py: update function exclude_blacklisted(links)

Browse Source
This commit is contained in:
mlazana
2019-03-27 20:10:05 +02:00
parent a3705e31c6
commit 8502fa5cc3
1 changed files with 11 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
archivebox/links.py
View File

@@ -37,7 +37,7 @@ def validate_links(links):
links = archivable_links(links) # remove chrome://, about:, mailto: etc. links = archivable_links(links) # remove chrome://, about:, mailto: etc.
links = uniquefied_links(links) # merge/dedupe duplicate timestamps & urls links = uniquefied_links(links) # merge/dedupe duplicate timestamps & urls
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 = list(exclude_links(links)) # exclude URLs that match the blacklisted url pattern regex
if not links: if not links:
print('[X] No links found :(') print('[X] No links found :(')
@@ -120,9 +120,8 @@ def lowest_uniq_timestamp(used_timestamps, timestamp):
return new_timestamp return new_timestamp
def exclude_links(links): def exclude_blacklisted(links):
"""exclude links that are in blacklist""" """exclude URLs that match the blacklisted url pattern regex"""
return (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'])]
return links