mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 14:44:29 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

create Chrome profile on first run automagically

Browse source
This commit is contained in:
Nick Sweeting 2024-10-21 01:33:08 -07:00
parent 267fde0138
commit 5e743b6aee
No known key found for this signature in database
1 changed files with 24 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

36
archivebox/plugins_extractor/chrome/config.py
View file

@ -102,7 +102,7 @@ class ChromeConfig(BaseConfigSet):
# Cookies & Auth # Cookies & Auth
CHROME_USER_AGENT: str = Field(default=lambda: ARCHIVING_CONFIG.USER_AGENT) CHROME_USER_AGENT: str = Field(default=lambda: ARCHIVING_CONFIG.USER_AGENT)
CHROME_USER_DATA_DIR: Path | None = Field(default=None) CHROME_USER_DATA_DIR: Path | None = Field(default=CONSTANTS.PERSONAS_DIR / 'Default' / 'chrome_profile')
CHROME_PROFILE_NAME: str = Field(default='Default') CHROME_PROFILE_NAME: str = Field(default='Default')
# Extractor Toggles # Extractor Toggles
@ -110,9 +110,11 @@ class ChromeConfig(BaseConfigSet):
SAVE_DOM: bool = Field(default=True, alias='FETCH_DOM') SAVE_DOM: bool = Field(default=True, alias='FETCH_DOM')
SAVE_PDF: bool = Field(default=True, alias='FETCH_PDF') SAVE_PDF: bool = Field(default=True, alias='FETCH_PDF')
@model_validator(mode='after') def validate(self):
def validate_use_chrome(self): from archivebox.config.paths import create_and_chown_dir
if self.USE_CHROME and self.CHROME_TIMEOUT < 15: if self.USE_CHROME and self.CHROME_TIMEOUT < 15:
STDERR.print()
STDERR.print(f'[red][!] Warning: TIMEOUT is set too low! (currently set to TIMEOUT={self.CHROME_TIMEOUT} seconds)[/red]') STDERR.print(f'[red][!] Warning: TIMEOUT is set too low! (currently set to TIMEOUT={self.CHROME_TIMEOUT} seconds)[/red]')
STDERR.print(' Chrome will fail to archive all sites if set to less than ~15 seconds.') STDERR.print(' Chrome will fail to archive all sites if set to less than ~15 seconds.')
STDERR.print(' (Setting it to somewhere between 30 and 300 seconds is recommended)') STDERR.print(' (Setting it to somewhere between 30 and 300 seconds is recommended)')
@ -122,27 +124,30 @@ class ChromeConfig(BaseConfigSet):
STDERR.print() STDERR.print()
# if user has specified a user data dir, make sure its valid # if user has specified a user data dir, make sure its valid
if self.CHROME_USER_DATA_DIR and os.access(self.CHROME_USER_DATA_DIR, os.R_OK): if self.USE_CHROME and self.CHROME_USER_DATA_DIR:
try:
create_and_chown_dir(self.CHROME_USER_DATA_DIR / self.CHROME_PROFILE_NAME)
except Exception:
pass
# check to make sure user_data_dir/<profile_name> exists # check to make sure user_data_dir/<profile_name> exists
if not (self.CHROME_USER_DATA_DIR / self.CHROME_PROFILE_NAME).is_dir(): if not os.path.isdir(self.CHROME_USER_DATA_DIR / self.CHROME_PROFILE_NAME):
STDERR.print()
STDERR.print()
STDERR.print(f'[red][X] Could not find profile "{self.CHROME_PROFILE_NAME}" in CHROME_USER_DATA_DIR.[/red]') STDERR.print(f'[red][X] Could not find profile "{self.CHROME_PROFILE_NAME}" in CHROME_USER_DATA_DIR.[/red]')
STDERR.print(f' {self.CHROME_USER_DATA_DIR}') STDERR.print(f' {self.CHROME_USER_DATA_DIR}')
STDERR.print(' Make sure you set it to a Chrome user data directory containing a Default profile folder.') STDERR.print(' Make sure you set it to a Chrome user data directory containing a Default profile folder.')
STDERR.print(' For more info see:') STDERR.print(' For more info see:')
STDERR.print(' https://github.com/ArchiveBox/ArchiveBox/wiki/Configuration#CHROME_USER_DATA_DIR') STDERR.print(' https://github.com/ArchiveBox/ArchiveBox/wiki/Configuration#CHROME_USER_DATA_DIR')
if '/Default' in str(self.CHROME_USER_DATA_DIR):
# show special hint if they made the common mistake of putting /Default at the end of the path
if str(self.CHROME_USER_DATA_DIR).replace(str(CONSTANTS.PERSONAS_DIR / 'Default'), '').endswith('/Default'):
STDERR.print() STDERR.print()
STDERR.print(' Try removing /Default from the end e.g.:') STDERR.print(' Try removing /Default from the end e.g.:')
STDERR.print(' CHROME_USER_DATA_DIR="{}"'.format(str(self.CHROME_USER_DATA_DIR).split('/Default')[0])) STDERR.print(' CHROME_USER_DATA_DIR="{}"'.format(str(self.CHROME_USER_DATA_DIR).rsplit('/Default', 1)[0]))
# hard error is too annoying here, instead just set it to nothing
# raise SystemExit(2)
self.update_in_place(CHROME_USER_DATA_DIR=None)
else:
if self.CHROME_USER_DATA_DIR is not None:
self.update_in_place(CHROME_USER_DATA_DIR=None) self.update_in_place(CHROME_USER_DATA_DIR=None)
return self
def chrome_args(self, **options) -> List[str]: def chrome_args(self, **options) -> List[str]:
"""helper to build up a chrome shell command with arguments""" """helper to build up a chrome shell command with arguments"""
@ -187,6 +192,11 @@ class ChromeConfig(BaseConfigSet):
cmd_args.append('--user-data-dir={}'.format(options.CHROME_USER_DATA_DIR)) cmd_args.append('--user-data-dir={}'.format(options.CHROME_USER_DATA_DIR))
cmd_args.append('--profile-directory={}'.format(options.CHROME_PROFILE_NAME or 'Default')) cmd_args.append('--profile-directory={}'.format(options.CHROME_PROFILE_NAME or 'Default'))
if not os.path.isfile(options.CHROME_USER_DATA_DIR / options.CHROME_PROFILE_NAME / 'Preferences'):
STDERR.print(f'[green] + creating new Chrome profile in: {pretty_path(options.CHROME_USER_DATA_DIR / options.CHROME_PROFILE_NAME)}[/green]')
cmd_args.remove('--no-first-run')
cmd_args.append('--first-run')
return dedupe(cmd_args) return dedupe(cmd_args)
CHROME_CONFIG = ChromeConfig() CHROME_CONFIG = ChromeConfig()