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

sleeps to prevent too many requests in a timeframe

Browse Source
This commit is contained in:
Akash 2020-07-18 19:25:19 +05:30 committed by GitHub
parent 3fa0c32064
commit f7de8f5575
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 13 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
tests/test_1.py
View File

@ -4,10 +4,12 @@ sys.path.append("..")
import waybackpy import waybackpy
import pytest import pytest
import random import random
import time
user_agent = "Mozilla/5.0 (Windows NT 6.2; rv:20.0) Gecko/20121202 Firefox/20.0" user_agent = "Mozilla/5.0 (Windows NT 6.2; rv:20.0) Gecko/20121202 Firefox/20.0"
def test_clean_url(): def test_clean_url():
time.sleep(10)
test_url = " https://en.wikipedia.org/wiki/Network security " test_url = " https://en.wikipedia.org/wiki/Network security "
answer = "https://en.wikipedia.org/wiki/Network_security" answer = "https://en.wikipedia.org/wiki/Network_security"
target = waybackpy.Url(test_url, user_agent) target = waybackpy.Url(test_url, user_agent)
@ -15,12 +17,14 @@ def test_clean_url():
assert answer == test_result assert answer == test_result
def test_url_check(): def test_url_check():
time.sleep(10)
broken_url = "http://wwwgooglecom/" broken_url = "http://wwwgooglecom/"
with pytest.raises(Exception) as e_info: with pytest.raises(Exception) as e_info:
waybackpy.Url(broken_url, user_agent) waybackpy.Url(broken_url, user_agent)
def test_save(): def test_save():
# Test for urls that exist and can be archived. # Test for urls that exist and can be archived.
time.sleep(10)
url_list = [ url_list = [
"en.wikipedia.org", "en.wikipedia.org",
@ -42,14 +46,14 @@ def test_save():
with pytest.raises(Exception) as e_info: with pytest.raises(Exception) as e_info:
url2 = "ha ha ha ha" url2 = "ha ha ha ha"
waybackpy.Url(url2, user_agent) waybackpy.Url(url2, user_agent)
time.sleep(5)
# Test for urls not allowed to archive by robot.txt. # Test for urls not allowed to archive by robot.txt.
with pytest.raises(Exception) as e_info: with pytest.raises(Exception) as e_info:
url3 = "http://www.archive.is/faq.html" 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") target = waybackpy.Url(url3, "Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:25.0) Gecko/20100101 Firefox/25.0")
target.save() target.save()
time.sleep(5)
# Non existent urls, test # Non existent urls, test
with pytest.raises(Exception) as e_info: with pytest.raises(Exception) as e_info:
url4 = "https://githfgdhshajagjstgeths537agajaajgsagudadhuss8762346887adsiugujsdgahub.us" url4 = "https://githfgdhshajagjstgeths537agajaajgsagudadhuss8762346887adsiugujsdgahub.us"
@ -60,18 +64,17 @@ def test_save():
pass pass
def test_near(): def test_near():
time.sleep(10)
url = "google.com" 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") 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) archive_near_year = target.near(year=2010)
assert "2010" in archive_near_year assert "2010" in archive_near_year
if sys.version_info > (3, 6): if sys.version_info > (3, 6):
time.sleep(5)
archive_near_month_year = target.near( year=2015, month=2) 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) assert ("201502" in archive_near_month_year) or ("201501" in archive_near_month_year) or ("201503" in archive_near_month_year)
archive_near_day_month_year = target.near(year=2006, month=11, day=15)
assert ("20061114" in archive_near_day_month_year) or ("20061115" in archive_near_day_month_year) or ("2006116" in archive_near_day_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") 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) 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) 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)
@ -84,26 +87,30 @@ def test_near():
pass pass
def test_oldest(): def test_oldest():
time.sleep(10)
url = "github.com/akamhy/waybackpy" url = "github.com/akamhy/waybackpy"
target = waybackpy.Url(url, user_agent) target = waybackpy.Url(url, user_agent)
assert "20200504141153" in target.oldest() assert "20200504141153" in target.oldest()
def test_newest(): def test_newest():
time.sleep(10)
url = "github.com/akamhy/waybackpy" url = "github.com/akamhy/waybackpy"
target = waybackpy.Url(url, user_agent) target = waybackpy.Url(url, user_agent)
assert url in target.newest() assert url in target.newest()
def test_get(): def test_get():
time.sleep(10)
target = waybackpy.Url("google.com", user_agent) target = waybackpy.Url("google.com", user_agent)
assert "Welcome to Google" in target.get(target.oldest()) assert "Welcome to Google" in target.get(target.oldest())
def test_total_archives(): def test_total_archives():
time.sleep(10)
if sys.version_info > (3, 6): if sys.version_info > (3, 6):
target = waybackpy.Url(" https://google.com ", user_agent) target = waybackpy.Url(" https://google.com ", user_agent)
assert target.total_archives() > 500000 assert target.total_archives() > 500000
else: else:
pass pass
time.sleep(5)
target = waybackpy.Url(" https://gaha.e4i3n.m5iai3kip6ied.cima/gahh2718gs/ahkst63t7gad8 ", user_agent) target = waybackpy.Url(" https://gaha.e4i3n.m5iai3kip6ied.cima/gahh2718gs/ahkst63t7gad8 ", user_agent)
assert target.total_archives() == 0 assert target.total_archives() == 0