waybackpy/tests/test_wrapper.py

133 lines
3.5 KiB
Python
Raw Normal View History

2020-05-05 12:59:55 +02:00
import sys
import pytest
2020-07-18 13:09:35 +02:00
import random
import requests
2021-01-03 20:14:55 +01:00
from datetime import datetime
2020-12-13 20:48:04 +01:00
2021-01-09 21:53:53 +01:00
from waybackpy.wrapper import Url, Cdx
2020-11-21 12:30:11 +01:00
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"
def test_url_check():
"""No API Use"""
broken_url = "http://wwwgooglecom/"
with pytest.raises(Exception):
2021-01-09 21:53:53 +01:00
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.
2020-07-18 13:09:35 +02:00
url_list = [
"en.wikipedia.org",
2021-01-09 21:53:53 +01:00
"akamhy.github.io",
2020-07-18 13:09:35 +02:00
"www.wiktionary.org",
"www.w3schools.com",
2021-01-09 21:53:53 +01:00
"youtube.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]
2021-01-09 21:53:53 +01:00
target = 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 = str(target.save())
2020-05-05 12:59:55 +02:00
assert url1 in archived_url1
2020-11-21 12:30:11 +01:00
# Test for urls that are incorrect.
with pytest.raises(Exception):
url2 = "ha ha ha ha"
2021-01-09 21:53:53 +01:00
Url(url2, user_agent)
2020-05-07 05:09:24 +02:00
2020-05-05 12:59:55 +02:00
def test_near():
url = "google.com"
2021-01-09 21:53:53 +01:00
target = 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)
assert "2010" in str(archive_near_year.timestamp)
2020-05-05 12:59:55 +02:00
archive_near_month_year = str(target.near(year=2015, month=2).timestamp)
2020-11-21 12:30:11 +01:00
assert (
("2015-02" in archive_near_month_year)
or ("2015-01" in archive_near_month_year)
or ("2015-03" in archive_near_month_year)
2020-11-21 12:30:11 +01:00
)
2021-01-09 21:53:53 +01:00
target = Url(
2020-11-21 12:30:11 +01:00
"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",
)
2020-12-13 20:48:04 +01:00
archive_near_hour_day_month_year = str(
target.near(year=2008, month=5, day=9, hour=15)
)
2020-11-21 12:30:11 +01:00
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"
)
2021-01-09 21:53:53 +01:00
target = Url(NeverArchivedUrl, user_agent)
2020-11-21 12:30:11 +01:00
target.near(year=2010)
2020-05-07 05:29:09 +02:00
2020-05-05 12:59:55 +02:00
def test_oldest():
url = "github.com/akamhy/waybackpy"
2021-01-09 21:53:53 +01:00
target = Url(url, user_agent)
2021-01-03 20:14:55 +01:00
o = target.oldest()
assert "20200504141153" in str(o)
assert "2020-05-04" in str(o._timestamp)
2020-05-05 12:59:55 +02:00
2020-12-13 20:48:04 +01:00
def test_json():
url = "github.com/akamhy/waybackpy"
2021-01-09 21:53:53 +01:00
target = Url(url, user_agent)
assert "archived_snapshots" in str(target.JSON)
2020-12-13 20:48:04 +01:00
def test_archive_url():
url = "github.com/akamhy/waybackpy"
2021-01-09 21:53:53 +01:00
target = Url(url, user_agent)
assert "github.com/akamhy" in str(target.archive_url)
2020-12-13 20:48:04 +01:00
2020-05-05 12:59:55 +02:00
def test_newest():
url = "github.com/akamhy/waybackpy"
2021-01-09 21:53:53 +01:00
target = Url(url, user_agent)
assert url in str(target.newest())
2020-05-07 05:09:24 +02:00
2020-05-05 12:59:55 +02:00
def test_get():
2021-01-09 21:53:53 +01:00
target = Url("google.com", user_agent)
assert "Welcome to Google" in target.get(target.oldest())
2020-05-07 11:30:28 +02:00
def test_total_archives():
2021-01-03 19:44:38 +01:00
user_agent = (
"Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:78.0) Gecko/20100101 Firefox/78.0"
)
2021-01-09 21:53:53 +01:00
target = Url(" https://outlook.com ", user_agent)
2021-01-03 19:44:38 +01:00
assert target.total_archives() > 80000
2020-11-21 12:30:11 +01:00
2021-01-09 21:53:53 +01:00
target = Url(
" https://gaha.e4i3n.m5iai3kip6ied.cima/gahh2718gs/ahkst63t7gad8 ", user_agent
)
assert target.total_archives() == 0
2020-12-13 20:48:04 +01:00
def test_known_urls():
2021-01-09 21:53:53 +01:00
target = Url("akamhy.github.io", user_agent)
assert len(target.known_urls()) > 3