mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-18 00:54:26 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

test: Fix tests post-rebase

Browse source
This commit is contained in:
Cristian 2020-09-15 14:05:48 -05:00 • committed by Cristian Vargas
parent 422664079a
commit b18bbf8874
15 changed files with 114 additions and 121 deletions
Download patch file Download diff file Expand all files Collapse all files

19
archivebox/extractors/screenshot.py
View file

@ -1,7 +1,6 @@
__package__ = 'archivebox.extractors'
import os
from pathlib import Path
from typing import Optional
from ..index.schema import Link, ArchiveResult, ArchiveOutput, ArchiveError
@ -21,21 +20,21 @@ from ..logging_util import TimedProgress
@enforce_types
def should_save_screenshot(link: Link, out_dir: Optional[str]=None) -> bool:
out_dir = out_dir or link.link_dir
def should_save_screenshot(link: Link, out_dir: Optional[Path]=None) -> bool:
out_dir = out_dir or Path(link.link_dir)
if is_static_file(link.url):
return False
if os.path.exists(os.path.join(out_dir, 'screenshot.png')):
if (out_dir / "screenshot.png").exists():
return False
return SAVE_SCREENSHOT
@enforce_types
def save_screenshot(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEOUT) -> ArchiveResult:
def save_screenshot(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT) -> ArchiveResult:
"""take screenshot of site using chrome --headless"""
out_dir = out_dir or link.link_dir
out_dir = out_dir or Path(link.link_dir)
output: ArchiveOutput = 'screenshot.png'
cmd = [
*chrome_args(TIMEOUT=timeout),
@ -45,13 +44,13 @@ def save_screenshot(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEOU
status = 'succeeded'
timer = TimedProgress(timeout, prefix=' ')
try:
result = run(cmd, cwd=out_dir, timeout=timeout)
result = run(cmd, cwd=str(out_dir), timeout=timeout)
if result.returncode:
hints = (result.stderr or result.stdout).decode()
raise ArchiveError('Failed to save screenshot', hints)
chmod_file(output, cwd=out_dir)
chmod_file(output, cwd=str(out_dir))
except Exception as err:
status = 'failed'
output = err
@ -60,7 +59,7 @@ def save_screenshot(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEOU
return ArchiveResult(
cmd=cmd,
pwd=out_dir,
pwd=str(out_dir),
cmd_version=CHROME_VERSION,
output=output,
status=status,