diff options
Diffstat (limited to 'python/tests/test_pushers.py')
| -rw-r--r-- | python/tests/test_pushers.py | 33 |
1 files changed, 17 insertions, 16 deletions
diff --git a/python/tests/test_pushers.py b/python/tests/test_pushers.py index 52f26c0..ed17d24 100644 --- a/python/tests/test_pushers.py +++ b/python/tests/test_pushers.py @@ -1,7 +1,4 @@ - -import pytest - -from sandcrawler.workers import CdxLinePusher, BlackholeSink +from sandcrawler.workers import BlackholeSink, CdxLinePusher def test_cdx_line_pusher(): @@ -9,20 +6,24 @@ def test_cdx_line_pusher(): sink = BlackholeSink() # vanilla (only default filters) - with open('tests/files/example.cdx', 'r') as cdx_file: + with open("tests/files/example.cdx", "r") as cdx_file: pusher = CdxLinePusher(sink, cdx_file) counts = pusher.run() - assert counts['total'] == 20 - assert counts['skip-parse'] == 1 - assert counts['pushed'] == 19 + assert counts["total"] == 20 + assert counts["skip-parse"] == 1 + assert counts["pushed"] == 19 # HTTP 200 and application/pdf - with open('tests/files/example.cdx', 'r') as cdx_file: - pusher = CdxLinePusher(sink, cdx_file, - filter_mimetypes=['application/pdf'], filter_http_statuses=[200, 226]) + with open("tests/files/example.cdx", "r") as cdx_file: + pusher = CdxLinePusher( + sink, + cdx_file, + filter_mimetypes=["application/pdf"], + filter_http_statuses=[200, 226], + ) counts = pusher.run() - assert counts['total'] == 20 - assert counts['skip-parse'] == 1 - assert counts['skip-http_status'] == 10 - assert counts['skip-mimetype'] == 2 - assert counts['pushed'] == 7 + assert counts["total"] == 20 + assert counts["skip-parse"] == 1 + assert counts["skip-http_status"] == 10 + assert counts["skip-mimetype"] == 2 + assert counts["pushed"] == 7 |