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 2020-05-04 08:21:42 +05:30 committed by GitHub
parent e66c70c8ee
commit 52b4bb9ac7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
waybackpy/wrapper.py
View File

@ -26,7 +26,8 @@ class PageNotSavedError(Exception):
class InvalidUrlError(Exception): class InvalidUrlError(Exception):
""" """
When the inpute url doesn't have '.' in it. Files like robots.txt are set to deny robot archiving.
Wayback machine respects these file, will not archive.
""" """
def save(url,UA=default_UA): def save(url,UA=default_UA):
@ -68,15 +69,14 @@ def near(
encoding = response.info().get_content_charset('utf8') encoding = response.info().get_content_charset('utf8')
import json import json
data = json.loads(response.read().decode(encoding)) data = json.loads(response.read().decode(encoding))
print(data)
if not data["archived_snapshots"]: if not data["archived_snapshots"]:
raise PageNotSavedError("'%s' was not archived." % url) raise PageNotSavedError("'%s' was not archived." % url)
archive_url = (data["archived_snapshots"]["closest"]["url"]) archive_url = (data["archived_snapshots"]["closest"]["url"])
return archive_url return archive_url
def oldest(url,UA=default_UA): def oldest(url,UA=default_UA,year=1994):
return near(url,year=1995,UA=UA) return near(url,year=year,UA=UA)
def newest(url,UA=default_UA): def newest(url,UA=default_UA):
return near(url,UA=UA) return near(url,UA=UA)