From ed497bbd23fe604ee94ac26b7c681072fa12f9e0 Mon Sep 17 00:00:00 2001 From: akamhy <64683866+akamhy@users.noreply.github.com> Date: Wed, 6 May 2020 20:07:25 +0530 Subject: [PATCH] Update wrapper.py --- waybackpy/wrapper.py | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/waybackpy/wrapper.py b/waybackpy/wrapper.py index 6efad7f..2cf2256 100644 --- a/waybackpy/wrapper.py +++ b/waybackpy/wrapper.py @@ -4,9 +4,9 @@ from datetime import datetime from waybackpy.exceptions import TooManyArchivingRequests, ArchivingNotAllowed, PageNotSaved, ArchiveNotFound, UrlNotFound, BadGateWay, InvalidUrl, WaybackUnavailable try: from urllib.request import Request, urlopen - from urllib.error import HTTPError + from urllib.error import HTTPError, URLError except ImportError: - from urllib2 import Request, urlopen, HTTPError + from urllib2 import Request, urlopen, HTTPError, URLError default_UA = "waybackpy python package" @@ -45,12 +45,16 @@ def get(url,encoding=None,UA=default_UA): url_check(url) hdr = { 'User-Agent' : '%s' % UA } req = Request(clean_url(url), headers=hdr) #nosec - resp=urlopen(req) #nosec + try: + resp=urlopen(req) #nosec + except URLError as e: + raise UrlNotFound(e) if encoding is None: try: encoding= resp.headers['content-type'].split('charset=')[-1] except AttributeError: encoding = "UTF-8" + encoding = encoding.replace("text/html","UTF-8",1) return resp.read().decode(encoding) def wayback_timestamp(year,month,day,hour,minute):