summaryrefslogtreecommitdiffstats
path: root/python/tests/grobid_metadata_importer.py
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2018-09-28 15:04:39 -0700
committerBryan Newbold <bnewbold@robocracy.org>2018-10-12 15:32:55 -0400
commitf18d985df6dbe1b96d0fa0df06a57a165982531a (patch)
tree696955137be602a8af279b6af5a1963e35fd24e0 /python/tests/grobid_metadata_importer.py
parent2882100d43b690177ff8faa7897cd423397fa2e7 (diff)
downloadfatcat-f18d985df6dbe1b96d0fa0df06a57a165982531a.tar.gz
fatcat-f18d985df6dbe1b96d0fa0df06a57a165982531a.zip
fix grobid_metadata_importer test for now
Diffstat (limited to 'python/tests/grobid_metadata_importer.py')
-rw-r--r--python/tests/grobid_metadata_importer.py19
1 files changed, 12 insertions, 7 deletions
diff --git a/python/tests/grobid_metadata_importer.py b/python/tests/grobid_metadata_importer.py
index 516d8c44..7bbabb39 100644
--- a/python/tests/grobid_metadata_importer.py
+++ b/python/tests/grobid_metadata_importer.py
@@ -3,6 +3,10 @@ import json
import pytest
from fatcat.grobid_metadata_importer import FatcatGrobidMetadataImporter
+"""
+WARNING: these tests are currently very fragile because they have database
+side-effects. Should probably be disabled or re-written.
+"""
@pytest.fixture(scope="function")
def grobid_metadata_importer():
@@ -13,10 +17,6 @@ def grobid_metadata_importer():
# with open('tests/files/example_grobid_metadata_lines.tsv', 'r') as f:
# grobid_metadata_importer.process_batch(f)
-def test_grobid_metadata_importer(grobid_metadata_importer):
- with open('tests/files/example_grobid_metadata_lines.tsv', 'r') as f:
- grobid_metadata_importer.process_source(f)
-
def test_grobid_metadata_parse(grobid_metadata_importer):
with open('tests/files/example_grobid_metadata_lines.tsv', 'r') as f:
raw = json.loads(f.readline().split('\t')[4])
@@ -32,16 +32,21 @@ def test_grobid_metadata_parse(grobid_metadata_importer):
def test_file_metadata_parse(grobid_metadata_importer):
with open('tests/files/example_grobid_metadata_lines.tsv', 'r') as f:
+ f.readline()
raw = f.readline().split('\t')
fe = grobid_metadata_importer.parse_file_metadata(
raw[0], json.loads(raw[1]), raw[2], int(raw[3]))
assert fe
- assert fe.sha1 == "38d725127246895368e4d9f950e377b4f21b6d75" # "sha1:HDLSKETSI2EVG2HE3H4VBY3XWTZBW3LV"
+ assert fe.sha1 == "d4a841744719518bf8bdd5d91576ccedc55efbb5" # "sha1:2SUEC5CHDFIYX6F52XMRK5WM5XCV565V"
assert fe.md5 == None
assert fe.mimetype == "application/pdf"
- assert fe.size == 260608
- assert fe.urls[1].url.startswith("http://e-journal.hamzanwadi.ac.id")
+ assert fe.size == 142710
+ assert fe.urls[1].url.startswith("http://via.library.depaul.edu")
assert fe.urls[1].rel == "web"
assert fe.urls[0].url.startswith("https://web.archive.org/")
assert fe.urls[0].rel == "webarchive"
assert len(fe.releases) == 0
+
+def test_grobid_metadata_importer(grobid_metadata_importer):
+ with open('tests/files/example_grobid_metadata_lines.tsv', 'r') as f:
+ grobid_metadata_importer.process_source(f)