forks/archivebox
1
0
Fork 0
Code Releases Activity

merge_links() used wrong index

Browse source 
Because merge_links() use the index, we need to get the new_links() _before_ we manipulate the index with write_links_index(). This has the negative side effect that the "Adding X new links ..." will output twice (because we execute merge_links() twice. For that, we only output stuff when the only_new is not set.
This commit is contained in:
Aaron Fischer 2018-10-19 22:35:08 +02:00 • committed by GitHub
parent 69c007ce85
commit b1b6be4f13
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
archiver/archive.py
View file

@ -64,7 +64,7 @@ def merge_links(archive_path=OUTPUT_DIR, import_path=None, only_new=False):
all_links = validate_links(existing_links + all_links)
num_new_links = len(all_links) - len(existing_links)
if num_new_links:
if num_new_links and not only_new:
print('[{green}+{reset}] [{}] Adding {} new links from {} to {}/index.json'.format(
datetime.now().strftime('%Y-%m-%d %H:%M:%S'),
num_new_links,
@ -166,7 +166,8 @@ if __name__ == '__main__':
# Step 1: Parse the links and dedupe them with existing archive
links = merge_links(archive_path=out_dir, import_path=source, only_new=False)
new_links = merge_links(archive_path=out_dir, import_path=source, only_new=True)
# Step 2: Write new index
write_links_index(out_dir=out_dir, links=links)
@ -175,7 +176,6 @@ if __name__ == '__main__':
# Step 4: Run the archive methods for each link
if ONLY_NEW:
new_links = merge_links(archive_path=out_dir, import_path=source, only_new=True)
update_archive(out_dir, new_links, source=source, resume=resume, append=True)
else:
update_archive(out_dir, links, source=source, resume=resume, append=True)