Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-08-01 19:16:15 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix more tests for new approach sans scope['surt']

Browse source
This commit is contained in:
Noah Levitt 2018-03-23 10:43:08 -07:00
parent 85a4757527
commit 05f8ab3495
2 changed files with 20 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

28
tests/test_units.py
View file

@ -94,28 +94,28 @@ blocks:
'url': 'http://example.com/foo/bar?baz=quux#monkey',
'site_id': site.id})
assert site.is_in_scope('http://example.com/foo/bar', page)
assert not site.is_in_scope('http://example.com/foo/baz', page)
assert site.accept_reject_or_neither('http://example.com/foo/bar', page) is True
assert site.accept_reject_or_neither('http://example.com/foo/baz', page) is None
assert not site.is_in_scope('http://foo.com/some.mp3', page)
assert site.is_in_scope('http://foo.com/blah/audio_file/some.mp3', page)
assert site.accept_reject_or_neither('http://foo.com/some.mp3', page) is None
assert site.accept_reject_or_neither('http://foo.com/blah/audio_file/some.mp3', page) is True
assert site.is_in_scope('http://a.b.vimeocdn.com/blahblah', page)
assert not site.is_in_scope('https://a.b.vimeocdn.com/blahblah', page)
assert site.accept_reject_or_neither('http://a.b.vimeocdn.com/blahblah', page) is True
assert site.accept_reject_or_neither('https://a.b.vimeocdn.com/blahblah', page) is None
assert site.is_in_scope('https://twitter.com/twit', page)
assert site.is_in_scope('https://twitter.com/twit?lang=en', page)
assert not site.is_in_scope('https://twitter.com/twit?lang=es', page)
assert site.accept_reject_or_neither('https://twitter.com/twit', page) is True
assert site.accept_reject_or_neither('https://twitter.com/twit?lang=en', page) is True
assert site.accept_reject_or_neither('https://twitter.com/twit?lang=es', page) is False
assert site.is_in_scope('https://www.facebook.com/whatevz', page)
assert site.accept_reject_or_neither('https://www.facebook.com/whatevz', page) is True
assert not site.is_in_scope(
'https://www.youtube.com/watch?v=dUIn5OAPS5s', page)
assert site.accept_reject_or_neither(
'https://www.youtube.com/watch?v=dUIn5OAPS5s', page) is None
yt_user_page = brozzler.Page(None, {
'url': 'https://www.youtube.com/user/SonoraSantaneraVEVO',
'site_id': site.id, 'hops_from_seed': 10})
assert site.is_in_scope(
'https://www.youtube.com/watch?v=dUIn5OAPS5s', yt_user_page)
assert site.accept_reject_or_neither(
'https://www.youtube.com/watch?v=dUIn5OAPS5s', yt_user_page) is True
def test_proxy_down():
'''