Benjamin_Loison/waybackpy
1
0
Fork 0
Code Issues 1 Pull Requests Packages Projects Releases Wiki Activity

Update wrapper.py

Browse Source
This commit is contained in:
akamhy 2020-05-02 15:31:33 +05:30 committed by GitHub
parent 1abb46828b
commit 705f4164e2
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
pywayback/wrapper.py
View File

@ -24,15 +24,13 @@ def save(url,UA="pywayback python module"):
hdr = { 'User-Agent' : '%s' % UA } hdr = { 'User-Agent' : '%s' % UA }
req = Request(request_url, headers=hdr) req = Request(request_url, headers=hdr)
try: try:
response = urlopen(req) response = urlopen(req) #nosec
except urllib.error.HTTPError as e: except urllib.error.HTTPError as e:
raise TooManyArchivingRequestsError(e) raise TooManyArchivingRequestsError(e)
# print(response.read())
header = response.headers header = response.headers
if "exclusion.robots.policy" in str(header): if "exclusion.robots.policy" in str(header):
raise ArchivingNotAllowed("Can not archive %s. Disabled by site owner." % (url)) raise ArchivingNotAllowed("Can not archive %s. Disabled by site owner." % (url))
archive_id = header['Content-Location'] archive_id = header['Content-Location']
print(header)
archived_url = "https://web.archive.org" + archive_id archived_url = "https://web.archive.org" + archive_id
return archived_url return archived_url