mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-14 07:04:27 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

add chrome headless option and improve default data dir finding

Browse source
This commit is contained in:
Nick Sweeting 2019-03-12 17:50:10 -04:00
parent 8630c0fdaa
commit 1c1bc76ac1
2 changed files with 39 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

47
archivebox/archive_methods.py
View file

@ -31,6 +31,7 @@ from config import (
COOKIES_FILE, COOKIES_FILE,
WGET_USER_AGENT, WGET_USER_AGENT,
CHROME_USER_DATA_DIR, CHROME_USER_DATA_DIR,
CHROME_HEADLESS,
CHROME_SANDBOX, CHROME_SANDBOX,
TIMEOUT, TIMEOUT,
MEDIA_TIMEOUT, MEDIA_TIMEOUT,
@ -613,14 +614,42 @@ def fetch_git(link_dir, link, timeout=TIMEOUT):
'output': output, 'output': output,
} }
def chrome_headless(binary=CHROME_BINARY, user_data_dir=CHROME_USER_DATA_DIR): def chrome_headless(binary=CHROME_BINARY, user_data_dir=CHROME_USER_DATA_DIR, headless=CHROME_HEADLESS, sandbox=CHROME_SANDBOX):
args = [binary, '--headless'] global USER_DATA_DIR
if not CHROME_SANDBOX: user_data_dir = user_data_dir or USER_DATA_DIR
cmd_args = [binary]
if headless:
cmd_args += ('--headless',)
if not sandbox:
# dont use GPU or sandbox when running inside docker container # dont use GPU or sandbox when running inside docker container
args += ['--no-sandbox', '--disable-gpu'] cmd_args += ('--no-sandbox', '--disable-gpu')
default_profile = os.path.expanduser('~/Library/Application Support/Google/Chrome')
# Find chrome user data directory
default_profile_paths = (
'~/.config/chromium',
'~/.config/google-chrome',
'~/.config/google-chrome-beta',
'~/.config/google-chrome-unstable',
'~/Library/Application Support/Chromium',
'~/Library/Application Support/Google/Chrome',
'~/Library/Application Support/Google/Chrome Canary',
'~/AppData/Local/Chromium/User Data',
'~/AppData/Local/Google/Chrome/User Data',
'~/AppData/Local/Google/Chrome SxS/User Data',
)
if user_data_dir: if user_data_dir:
args.append('--user-data-dir={}'.format(user_data_dir)) cmd_args.append('--user-data-dir={}'.format(user_data_dir))
elif os.path.exists(default_profile): else:
args.append('--user-data-dir={}'.format(default_profile)) for path in default_profile_paths:
return args full_path = os.path.expanduser(path)
if os.path.exists(full_path):
USER_DATA_DIR = full_path
cmd_args.append('--user-data-dir={}'.format(full_path))
break
return cmd_args
USER_DATA_DIR = CHROME_USER_DATA_DIR

1
archivebox/config.py
View file

@ -37,6 +37,7 @@ GIT_DOMAINS = os.getenv('GIT_DOMAINS', 'github.com,bitbuck
WGET_USER_AGENT = os.getenv('WGET_USER_AGENT', 'ArchiveBox/{GIT_SHA} (+https://github.com/pirate/ArchiveBox/) wget/{WGET_VERSION}') WGET_USER_AGENT = os.getenv('WGET_USER_AGENT', 'ArchiveBox/{GIT_SHA} (+https://github.com/pirate/ArchiveBox/) wget/{WGET_VERSION}')
COOKIES_FILE = os.getenv('COOKIES_FILE', None) COOKIES_FILE = os.getenv('COOKIES_FILE', None)
CHROME_USER_DATA_DIR = os.getenv('CHROME_USER_DATA_DIR', None) CHROME_USER_DATA_DIR = os.getenv('CHROME_USER_DATA_DIR', None)
CHROME_HEADLESS = os.getenv('CHROME_HEADLESS', 'True' ).lower() == 'true'
CURL_BINARY = os.getenv('CURL_BINARY', 'curl') CURL_BINARY = os.getenv('CURL_BINARY', 'curl')
GIT_BINARY = os.getenv('GIT_BINARY', 'git') GIT_BINARY = os.getenv('GIT_BINARY', 'git')