mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-16 16:14:28 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Refactor should_save_extractor methods to accept overwrite parameter

Browse source
This commit is contained in:
Dan Arnfield 2021-01-21 15:45:11 -06:00
parent 553c3ca219
commit 5420903102
14 changed files with 56 additions and 50 deletions
Download patch file Download diff file Expand all files Collapse all files

10
archivebox/extractors/readability.py
View file

@ -46,13 +46,15 @@ def get_html(link: Link, path: Path) -> str:
return document
@enforce_types
def should_save_readability(link: Link, out_dir: Optional[str]=None) -> bool:
out_dir = out_dir or link.link_dir
def should_save_readability(link: Link, out_dir: Optional[str]=None, overwrite: Optional[bool]=False) -> bool:
if is_static_file(link.url):
return False
output = Path(out_dir or link.link_dir) / 'readability'
return SAVE_READABILITY and READABILITY_VERSION and (not output.exists())
out_dir = out_dir or Path(link.link_dir)
if not overwrite and (out_dir / 'readability').exists():
return False
return SAVE_READABILITY
@enforce_types