aboutsummaryrefslogtreecommitdiffstats
path: root/python/tests/transform_elasticsearch.py
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2020-01-31 13:31:59 -0800
committerBryan Newbold <bnewbold@robocracy.org>2020-01-31 13:31:59 -0800
commit741c7b1efae5e39f3ee2c082e3ca28c6c5c85b76 (patch)
tree7c380becdc0b2cd8e1e5b40af76e4f85b3fe0a1c /python/tests/transform_elasticsearch.py
parent0d037d0d2f73b18014d8d98a06fa3f7bc2c9b794 (diff)
downloadfatcat-741c7b1efae5e39f3ee2c082e3ca28c6c5c85b76.tar.gz
fatcat-741c7b1efae5e39f3ee2c082e3ca28c6c5c85b76.zip
ES releases: host/domain fixes
Diffstat (limited to 'python/tests/transform_elasticsearch.py')
-rw-r--r--python/tests/transform_elasticsearch.py3
1 files changed, 3 insertions, 0 deletions
diff --git a/python/tests/transform_elasticsearch.py b/python/tests/transform_elasticsearch.py
index e67681c6..c94ab375 100644
--- a/python/tests/transform_elasticsearch.py
+++ b/python/tests/transform_elasticsearch.py
@@ -110,6 +110,9 @@ def test_elasticsearch_file_transform(matched_importer):
assert 'web' in es['rels']
assert 'www.zhros.ru' in es['hosts']
assert 'zhros.ru' in es['domains']
+ assert not '.archive.org' in (es['hosts'] + es['domains'])
+ assert not 'archive.org' in (es['hosts'] + es['domains'])
+ assert not 'web.archive.org' in (es['hosts'] + es['domains'])
def test_elasticsearch_changelog_transform(matched_importer):
ce = entity_from_json(open('./tests/files/changelog_3469683.json', 'r').read(), ChangelogEntry)