waybackpy/tests/test_wrapper.py

180 lines
5.2 KiB
Python
Raw Normal View History

# -*- coding: utf-8 -*-
2020-05-05 12:59:55 +02:00
import sys
import pytest
2020-07-18 13:09:35 +02:00
import random
import time
sys.path.append("..")
import waybackpy.wrapper as waybackpy # noqa: E402
2020-07-20 06:37:31 +02:00
if sys.version_info >= (3, 0): # If the python ver >= 3
from urllib.request import Request, urlopen
from urllib.error import URLError
else: # For python2.x
2020-07-20 06:37:31 +02:00
from urllib2 import Request, urlopen, URLError
2020-05-05 12:59:55 +02:00
user_agent = "Mozilla/5.0 (Windows NT 6.2; rv:20.0) Gecko/20121202 Firefox/20.0"
2020-05-07 05:29:09 +02:00
def test_clean_url():
test_url = " https://en.wikipedia.org/wiki/Network security "
answer = "https://en.wikipedia.org/wiki/Network_security"
target = waybackpy.Url(test_url, user_agent)
test_result = target._clean_url()
2020-05-07 05:29:09 +02:00
assert answer == test_result
2020-05-07 05:29:09 +02:00
def test_url_check():
broken_url = "http://wwwgooglecom/"
with pytest.raises(Exception):
waybackpy.Url(broken_url, user_agent)
2020-05-05 12:59:55 +02:00
2020-05-05 12:59:55 +02:00
def test_save():
# Test for urls that exist and can be archived.
time.sleep(10)
2020-07-18 13:09:35 +02:00
url_list = [
"en.wikipedia.org",
"www.wikidata.org",
"commons.wikimedia.org",
"www.wiktionary.org",
"www.w3schools.com",
"twitter.com",
2020-07-18 13:09:35 +02:00
]
x = random.randint(0, len(url_list) - 1)
2020-07-18 13:28:23 +02:00
url1 = url_list[x]
target = waybackpy.Url(
url1,
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_2) AppleWebKit/537.36 "
"(KHTML, like Gecko) Chrome/36.0.1944.0 Safari/537.36",
)
archived_url1 = target.save()
2020-05-05 12:59:55 +02:00
assert url1 in archived_url1
2020-07-18 13:54:54 +02:00
if sys.version_info > (3, 6):
2020-05-05 12:59:55 +02:00
2020-07-18 13:54:54 +02:00
# Test for urls that are incorrect.
with pytest.raises(Exception):
2020-07-18 13:54:54 +02:00
url2 = "ha ha ha ha"
waybackpy.Url(url2, user_agent)
time.sleep(5)
2020-07-18 13:54:54 +02:00
# Test for urls not allowed to archive by robot.txt.
with pytest.raises(Exception):
2020-07-18 13:54:54 +02:00
url3 = "http://www.archive.is/faq.html"
target = waybackpy.Url(
url3,
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:25.0) "
"Gecko/20100101 Firefox/25.0",
)
2020-07-18 13:54:54 +02:00
target.save()
time.sleep(5)
2020-07-18 13:54:54 +02:00
# Non existent urls, test
with pytest.raises(Exception):
url4 = (
"https://githfgdhshajagjstgeths537agajaajgsagudadhuss87623"
"46887adsiugujsdgahub.us"
)
target = waybackpy.Url(
url3,
"Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US) "
"AppleWebKit/533.20.25 (KHTML, like Gecko) Version/5.0.4 "
"Safari/533.20.27",
)
2020-07-18 13:54:54 +02:00
target.save()
else:
pass
2020-05-07 05:09:24 +02:00
2020-05-05 12:59:55 +02:00
def test_near():
time.sleep(10)
2020-05-05 12:59:55 +02:00
url = "google.com"
target = waybackpy.Url(
url,
"Mozilla/5.0 (Windows; U; Windows NT 6.0; de-DE) AppleWebKit/533.20.25 "
"(KHTML, like Gecko) Version/5.0.3 Safari/533.19.4",
)
archive_near_year = target.near(year=2010)
2020-05-05 12:59:55 +02:00
assert "2010" in archive_near_year
2020-07-18 13:54:54 +02:00
if sys.version_info > (3, 6):
time.sleep(5)
archive_near_month_year = target.near(year=2015, month=2)
assert (
("201502" in archive_near_month_year)
or ("201501" in archive_near_month_year)
or ("201503" in archive_near_month_year)
)
target = waybackpy.Url(
"www.python.org",
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
"(KHTML, like Gecko) Chrome/42.0.2311.135 Safari/537.36 Edge/12.246",
)
archive_near_hour_day_month_year = target.near(
year=2008, month=5, day=9, hour=15
)
assert (
("2008050915" in archive_near_hour_day_month_year)
or ("2008050914" in archive_near_hour_day_month_year)
or ("2008050913" in archive_near_hour_day_month_year)
)
with pytest.raises(Exception):
NeverArchivedUrl = (
"https://ee_3n.wrihkeipef4edia.org/rwti5r_ki/Nertr6w_rork_rse7c_urity"
)
2020-07-18 13:54:54 +02:00
target = waybackpy.Url(NeverArchivedUrl, user_agent)
target.near(year=2010)
else:
pass
2020-05-07 05:29:09 +02:00
2020-05-05 12:59:55 +02:00
def test_oldest():
url = "github.com/akamhy/waybackpy"
target = waybackpy.Url(url, user_agent)
assert "20200504141153" in target.oldest()
2020-05-05 12:59:55 +02:00
2020-05-05 12:59:55 +02:00
def test_newest():
url = "github.com/akamhy/waybackpy"
target = waybackpy.Url(url, user_agent)
assert url in target.newest()
2020-05-07 05:09:24 +02:00
2020-05-05 12:59:55 +02:00
def test_get():
target = waybackpy.Url("google.com", user_agent)
assert "Welcome to Google" in target.get(target.oldest())
2020-05-07 05:09:24 +02:00
2020-07-20 06:37:31 +02:00
def test_wayback_timestamp():
ts = waybackpy._wayback_timestamp(
year=2020, month=1, day=2, hour=3, minute=4
)
2020-07-20 06:37:31 +02:00
assert "202001020304" in str(ts)
2020-07-20 06:37:31 +02:00
def test_get_response():
hdr = {
"User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:78.0) "
"Gecko/20100101 Firefox/78.0"
}
req = Request("https://www.google.com", headers=hdr) # nosec
response = waybackpy._get_response(req)
2020-07-20 07:01:47 +02:00
assert response.code == 200
2020-07-20 06:37:31 +02:00
2020-05-07 11:30:28 +02:00
def test_total_archives():
2020-07-18 13:54:54 +02:00
if sys.version_info > (3, 6):
target = waybackpy.Url(" https://google.com ", user_agent)
assert target.total_archives() > 500000
else:
pass
target = waybackpy.Url(
" https://gaha.e4i3n.m5iai3kip6ied.cima/gahh2718gs/ahkst63t7gad8 ", user_agent
)
assert target.total_archives() == 0