mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 22:54:27 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

add custom WGET_USER_AGENT override option

Browse source
This commit is contained in:
Nick Sweeting 2017-07-06 16:51:41 -05:00
parent 83306391ed
commit acf59faa06
2 changed files with 3 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

1
config.py
View file

@ -28,6 +28,7 @@ RESOLUTION = os.getenv('RESOLUTION', '1440,900'
ARCHIVE_PERMISSIONS = os.getenv('ARCHIVE_PERMISSIONS', '755' ) ARCHIVE_PERMISSIONS = os.getenv('ARCHIVE_PERMISSIONS', '755' )
CHROME_BINARY = os.getenv('CHROME_BINARY', 'chromium-browser' ) # change to google-chrome browser if using google-chrome CHROME_BINARY = os.getenv('CHROME_BINARY', 'chromium-browser' ) # change to google-chrome browser if using google-chrome
WGET_BINARY = os.getenv('WGET_BINARY', 'wget' ) WGET_BINARY = os.getenv('WGET_BINARY', 'wget' )
WGET_USER_AGENT = os.getenv('WGET_USER_AGENT', None)
TIMEOUT = int(os.getenv('TIMEOUT', '60')) TIMEOUT = int(os.getenv('TIMEOUT', '60'))
INDEX_TEMPLATE = os.getenv('INDEX_TEMPLATE', 'templates/index.html') INDEX_TEMPLATE = os.getenv('INDEX_TEMPLATE', 'templates/index.html')
INDEX_ROW_TEMPLATE = os.getenv('INDEX_ROW_TEMPLATE', 'templates/index_row.html') INDEX_ROW_TEMPLATE = os.getenv('INDEX_ROW_TEMPLATE', 'templates/index_row.html')

2
fetch.py
View file

@ -17,6 +17,7 @@ from config import (
FETCH_AUDIO, FETCH_AUDIO,
FETCH_VIDEO, FETCH_VIDEO,
FETCH_FAVICON, FETCH_FAVICON,
WGET_USER_AGENT,
TIMEOUT, TIMEOUT,
ANSI, ANSI,
progress, progress,
@ -32,6 +33,7 @@ def fetch_wget(out_dir, link, overwrite=False, requisites=True, timeout=TIMEOUT)
CMD = [ CMD = [
*'wget --timestamping --adjust-extension --no-parent'.split(' '), # Docs: https://www.gnu.org/software/wget/manual/wget.html *'wget --timestamping --adjust-extension --no-parent'.split(' '), # Docs: https://www.gnu.org/software/wget/manual/wget.html
*(('--page-requisites', '--convert-links') if requisites else ()), *(('--page-requisites', '--convert-links') if requisites else ()),
*(('--user-agent="{}"'.format(WGET_USER_AGENT)) if WGET_USER_AGENT else ()),
link['url'], link['url'],
] ]
end = progress(timeout, prefix=' ') end = progress(timeout, prefix=' ')