Mirrors/ArchiveBox
2
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox synced 2024-11-27 06:30:22 +00:00
Code Issues Projects Releases Packages Wiki Activity

Merge pull request #337 from comsomisha/master

Browse source 
New feature: change user-agent for curl.
This commit is contained in:
Nick Sweeting 2020-04-15 05:11:16 -04:00 committed by GitHub
commit b5f40d9b4e
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 11 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
archivebox/archive_methods.py
View file

@ -30,6 +30,8 @@ from config import (
OUTPUT_DIR, OUTPUT_DIR,
GIT_DOMAINS, GIT_DOMAINS,
GIT_SHA, GIT_SHA,
RESTRICT_FILE_NAMES,
CURL_USER_AGENT,
WGET_USER_AGENT, WGET_USER_AGENT,
CHECK_SSL_VALIDITY, CHECK_SSL_VALIDITY,
COOKIES_FILE, COOKIES_FILE,
@ -226,7 +228,7 @@ def fetch_wget(link_dir, link, timeout=TIMEOUT):
'--span-hosts', '--span-hosts',
'--no-parent', '--no-parent',
'-e', 'robots=off', '-e', 'robots=off',
'--restrict-file-names=windows', *(('--restrict-file-names={}'.format(RESTRICT_FILE_NAMES),) if RESTRICT_FILE_NAMES else ()),
'--timeout={}'.format(timeout), '--timeout={}'.format(timeout),
*(('--compression=auto',) if WGET_AUTO_COMPRESSION else ()), *(('--compression=auto',) if WGET_AUTO_COMPRESSION else ()),
*(() if FETCH_WARC else ('--timestamping',)), *(() if FETCH_WARC else ('--timestamping',)),
@ -561,7 +563,7 @@ def archive_dot_org(link_dir, link, timeout=TIMEOUT):
CURL_BINARY, CURL_BINARY,
'--location', '--location',
'--head', '--head',
'--user-agent', 'ArchiveBox/{} (+https://github.com/pirate/ArchiveBox/)'.format(GIT_SHA), # be nice to the Archive.org people and show them where all this ArchiveBox traffic is coming from *(('--user-agent', '{}'.format(CURL_USER_AGENT),) if CURL_USER_AGENT else ()), # be nice to the Archive.org people and show them where all this ArchiveBox traffic is coming from
'--max-time', str(timeout), '--max-time', str(timeout),
*(() if CHECK_SSL_VALIDITY else ('--insecure',)), *(() if CHECK_SSL_VALIDITY else ('--insecure',)),
submit_url, submit_url,

6
archivebox/config.py
View file

@ -34,7 +34,9 @@ SUBMIT_ARCHIVE_DOT_ORG = os.getenv('SUBMIT_ARCHIVE_DOT_ORG', 'True'
CHECK_SSL_VALIDITY = os.getenv('CHECK_SSL_VALIDITY', 'True' ).lower() == 'true' CHECK_SSL_VALIDITY = os.getenv('CHECK_SSL_VALIDITY', 'True' ).lower() == 'true'
RESOLUTION = os.getenv('RESOLUTION', '1440,2000' ) RESOLUTION = os.getenv('RESOLUTION', '1440,2000' )
RESTRICT_FILE_NAMES = os.getenv('RESTRICT_FILE_NAMES', 'windows' )
GIT_DOMAINS = os.getenv('GIT_DOMAINS', 'github.com,bitbucket.org,gitlab.com').split(',') GIT_DOMAINS = os.getenv('GIT_DOMAINS', 'github.com,bitbucket.org,gitlab.com').split(',')
CURL_USER_AGENT = os.getenv('CURL_USER_AGENT', 'ArchiveBox/{GIT_SHA} (+https://github.com/pirate/ArchiveBox/)')
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)
@ -192,13 +194,15 @@ try:
raise raise
### Make sure curl is installed ### Make sure curl is installed
if FETCH_FAVICON or SUBMIT_ARCHIVE_DOT_ORG: if FETCH_FAVICON or FETCH_TITLE or SUBMIT_ARCHIVE_DOT_ORG:
if run(['which', CURL_BINARY], stdout=DEVNULL, stderr=DEVNULL).returncode or run([CURL_BINARY, '--version'], stdout=DEVNULL, stderr=DEVNULL).returncode: if run(['which', CURL_BINARY], stdout=DEVNULL, stderr=DEVNULL).returncode or run([CURL_BINARY, '--version'], stdout=DEVNULL, stderr=DEVNULL).returncode:
print('{red}[X] Missing dependency: curl{reset}'.format(**ANSI)) print('{red}[X] Missing dependency: curl{reset}'.format(**ANSI))
print(' Install it, then confirm it works with: {} --version'.format(CURL_BINARY)) print(' Install it, then confirm it works with: {} --version'.format(CURL_BINARY))
print(' See https://github.com/pirate/ArchiveBox/wiki/Install for help.') print(' See https://github.com/pirate/ArchiveBox/wiki/Install for help.')
raise SystemExit(1) raise SystemExit(1)
CURL_USER_AGENT = CURL_USER_AGENT.format(GIT_SHA=GIT_SHA[:9])
### Make sure wget is installed and calculate version ### Make sure wget is installed and calculate version
if FETCH_WGET or FETCH_WARC: if FETCH_WGET or FETCH_WARC:
if run(['which', WGET_BINARY], stdout=DEVNULL, stderr=DEVNULL).returncode or run([WGET_BINARY, '--version'], stdout=DEVNULL, stderr=DEVNULL).returncode: if run(['which', WGET_BINARY], stdout=DEVNULL, stderr=DEVNULL).returncode or run([WGET_BINARY, '--version'], stdout=DEVNULL, stderr=DEVNULL).returncode:

2
etc/ArchiveBox.conf.default
View file

@ -39,7 +39,9 @@
#CHECK_SSL_VALIDITY=True #CHECK_SSL_VALIDITY=True
#FETCH_WGET_REQUISITES=True #FETCH_WGET_REQUISITES=True
#RESTRICT_FILE_NAMES="windows"
#RESOLUTION="1440,900" #RESOLUTION="1440,900"
#CURL_USER_AGENT="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.75 Safari/537.36"
#WGET_USER_AGENT="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.75 Safari/537.36" #WGET_USER_AGENT="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.75 Safari/537.36"
#CHROME_USER_AGENT="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.75 Safari/537.36" #CHROME_USER_AGENT="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.75 Safari/537.36"
#GIT_DOMAINS="github.com,bitbucket.org,gitlab.com" #GIT_DOMAINS="github.com,bitbucket.org,gitlab.com"