mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-15 07:34:27 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

change mercury atomic_write output order

Browse source
This commit is contained in:
Nick Sweeting 2021-02-16 06:19:16 -05:00
parent e62ecffeb0
commit d0f8a5e710
2 changed files with 5 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

6
archivebox/extractors/mercury.py
View file

@ -59,6 +59,8 @@ def save_mercury(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT)
status = 'succeeded'
timer = TimedProgress(timeout, prefix=' ')
try:
output_folder.mkdir(exist_ok=True)
# Get plain text version of article
cmd = [
DEPENDENCIES['MERCURY_BINARY']['path'],
@ -71,6 +73,8 @@ def save_mercury(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT)
except json.JSONDecodeError:
raise ShellError(cmd, result)
atomic_write(str(output_folder / "content.txt"), article_text["content"])
# Get HTML version of article
cmd = [
DEPENDENCIES['MERCURY_BINARY']['path'],
@ -82,9 +86,7 @@ def save_mercury(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT)
except json.JSONDecodeError:
raise ShellError(cmd, result)
output_folder.mkdir(exist_ok=True)
atomic_write(str(output_folder / "content.html"), article_json.pop("content"))
atomic_write(str(output_folder / "content.txt"), article_text["content"])
atomic_write(str(output_folder / "article.json"), article_json)
# Check for common failure cases

1
archivebox/extractors/readability.py
View file

@ -112,6 +112,7 @@ def save_readability(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEO
except (Exception, OSError) as err:
status = 'failed'
output = err
cmd = [cmd[0], './{singlefile,dom}.html']
finally:
timer.end()