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

api changed; fix archive url parser

Browse Source
This commit is contained in:
Akash
2020-07-19 20:26:40 +05:30
committed by GitHub
parent 0aa51c961b
commit 4947fc9b90
1 changed files with 6 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
waybackpy/wrapper.py
View File

@@ -73,16 +73,14 @@ class Url():
raise WaybackError(e) raise WaybackError(e)
header = response.headers header = response.headers
try: def archive_url_parser(header):
arch = re.search(r"rel=\"memento.*?web\.archive\.org(/web/[0-9]{14}/.*?)>", str(header)) arch = re.search(r"X-Cache-Key:\shttps(.*)[A-Z]{2}", str(header))
if arch: if arch:
arch = arch.group(1) return arch.group(1)
else: else:
raise WaybackError("Can't create a new archive. No archive url found in reply headers.") raise WaybackError("No archive url found in the reply headers. Visit https://github.com/akamhy/waybackpy for latest version of waybackpy.")
except KeyError as e:
raise WaybackError(e)
return "https://web.archive.org" + arch return "https://" + archive_url_parser(header)
def get(self, url=None, user_agent=None, encoding=None): def get(self, url=None, user_agent=None, encoding=None):
"""Returns the source code of the supplied URL. Auto detects the encoding if not supplied.""" """Returns the source code of the supplied URL. Auto detects the encoding if not supplied."""