diff options
author | Bryan Newbold <bnewbold@archive.org> | 2020-01-14 16:05:41 -0800 |
---|---|---|
committer | Bryan Newbold <bnewbold@archive.org> | 2020-01-14 16:05:41 -0800 |
commit | ba6f16a02cfde0e4acb499c00b456b42472c0b00 (patch) | |
tree | 9b27d9982e61ee97a82181249c0f9723980a4beb /python/tests | |
parent | 21599839802b8ef3a84ffe90855f7bceaaa12a0d (diff) | |
download | sandcrawler-ba6f16a02cfde0e4acb499c00b456b42472c0b00.tar.gz sandcrawler-ba6f16a02cfde0e4acb499c00b456b42472c0b00.zip |
more ftp status 226 support
Diffstat (limited to 'python/tests')
-rw-r--r-- | python/tests/test_grobid.py | 8 | ||||
-rw-r--r-- | python/tests/test_pushers.py | 2 | ||||
-rw-r--r-- | python/tests/test_wayback.py | 2 |
3 files changed, 9 insertions, 3 deletions
diff --git a/python/tests/test_grobid.py b/python/tests/test_grobid.py index 8c5e080..24ea40e 100644 --- a/python/tests/test_grobid.py +++ b/python/tests/test_grobid.py @@ -64,8 +64,12 @@ def test_grobid_worker_cdx(grobid_client, wayback_client): body=REAL_TEI_XML, content_type='text/xml') with open('tests/files/example.cdx', 'r') as cdx_file: - pusher = CdxLinePusher(worker, cdx_file, - filter_http_statuses=[200], filter_mimetypes=['application/pdf']) + pusher = CdxLinePusher( + worker, + cdx_file, + filter_http_statuses=[200, 226], + filter_mimetypes=['application/pdf'], + ) pusher_counts = pusher.run() assert pusher_counts['total'] assert pusher_counts['pushed'] == 7 diff --git a/python/tests/test_pushers.py b/python/tests/test_pushers.py index ed9c0bb..52f26c0 100644 --- a/python/tests/test_pushers.py +++ b/python/tests/test_pushers.py @@ -19,7 +19,7 @@ def test_cdx_line_pusher(): # 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]) + filter_mimetypes=['application/pdf'], filter_http_statuses=[200, 226]) counts = pusher.run() assert counts['total'] == 20 assert counts['skip-parse'] == 1 diff --git a/python/tests/test_wayback.py b/python/tests/test_wayback.py index efffbe2..6bc1ca4 100644 --- a/python/tests/test_wayback.py +++ b/python/tests/test_wayback.py @@ -119,6 +119,7 @@ def wayback_client(cdx_client, mocker): resource = mocker.Mock() client.rstore.load_resource = mocker.MagicMock(return_value=resource) resource.get_status = mocker.MagicMock(return_value=(200, "Ok")) + resource.is_revisit = mocker.MagicMock(return_value=False) resource.get_location = mocker.MagicMock(return_value=WARC_TARGET) body = mocker.Mock() resource.open_raw_content = mocker.MagicMock(return_value=body) @@ -141,6 +142,7 @@ def wayback_client_pdf(cdx_client, mocker): resource = mocker.Mock() client.rstore.load_resource = mocker.MagicMock(return_value=resource) resource.get_status = mocker.MagicMock(return_value=(200, "Ok")) + resource.is_revisit = mocker.MagicMock(return_value=False) resource.get_location = mocker.MagicMock(return_value=WARC_TARGET) body = mocker.Mock() resource.open_raw_content = mocker.MagicMock(return_value=body) |