mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-18 17:14:39 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'dev' into link-removal2

Browse source
This commit is contained in:
Nick Sweeting 2021-01-30 03:43:46 -05:00
commit 1fe95474c2
52 changed files with 896 additions and 550 deletions
Download patch file Download diff file Expand all files Collapse all files

11
archivebox/extractors/dom.py
View file

@ -20,18 +20,21 @@ from ..config import (
from ..logging_util import TimedProgress
# output = 'output.html'
@enforce_types
def should_save_dom(snapshot: Model, out_dir: Optional[Path]=None) -> bool:
def should_save_dom(snapshot: Model, overwrite: Optional[bool]=False, out_dir: Optional[Path]=None) -> bool:
out_dir = out_dir or Path(snapshot.snapshot_dir)
if is_static_file(snapshot.url):
return False
if (out_dir / 'output.html').exists():
out_dir = out_dir or Path(link.link_dir)
if not overwrite and (out_dir / 'output.html').exists():
return False
return SAVE_DOM
@enforce_types
def save_dom(snapshot: Model, out_dir: Optional[Path]=None, timeout: int=TIMEOUT) -> ArchiveResult:
"""print HTML of site to file using chrome --dump-html"""