summaryrefslogtreecommitdiffstats
path: root/python/tests/grobid_metadata_importer.py
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2018-09-27 17:06:15 -0700
committerBryan Newbold <bnewbold@robocracy.org>2018-09-27 17:06:17 -0700
commitc7573bf142be405f8cb9003400c6860aeb700457 (patch)
tree552ba89d2dc80e6557519e4bf4ddbcceef058c3b /python/tests/grobid_metadata_importer.py
parent72d14a1ea8113d715e3f7933332829876a438618 (diff)
downloadfatcat-c7573bf142be405f8cb9003400c6860aeb700457.tar.gz
fatcat-c7573bf142be405f8cb9003400c6860aeb700457.zip
improvements to grobid_metadata importer
But still fails tests due to database collision/side-effect on sha1 lookup.
Diffstat (limited to 'python/tests/grobid_metadata_importer.py')
-rw-r--r--python/tests/grobid_metadata_importer.py47
1 files changed, 47 insertions, 0 deletions
diff --git a/python/tests/grobid_metadata_importer.py b/python/tests/grobid_metadata_importer.py
new file mode 100644
index 00000000..516d8c44
--- /dev/null
+++ b/python/tests/grobid_metadata_importer.py
@@ -0,0 +1,47 @@
+
+import json
+import pytest
+from fatcat.grobid_metadata_importer import FatcatGrobidMetadataImporter
+
+
+@pytest.fixture(scope="function")
+def grobid_metadata_importer():
+ yield FatcatGrobidMetadataImporter("http://localhost:9411/v0")
+
+# TODO: use API to check that entities actually created...
+#def test_grobid_metadata_importer_batch(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])
+ re = grobid_metadata_importer.parse_grobid_json(raw)
+ assert re
+ assert re.title == "PEMBELAJARAN FISIKA DENGAN PENDEKATAN KETERAMPILAN PROSES MELALUI METODE DEMONSTRASI MENGGUNAKAN MEDIA RIIL DAN MEDIA VIRTUIL DITINJAU DARI MOTIVASI DAN GAYA BERFIKIR SISWA"
+ assert len(re.contribs) == 5
+ print(re.contribs)
+ assert re.contribs[0].raw_name == "Wahyu Ary"
+ assert re.publisher == None
+ assert re.extra.get('container_name') == None
+ assert len(re.refs) == 27
+
+def test_file_metadata_parse(grobid_metadata_importer):
+ with open('tests/files/example_grobid_metadata_lines.tsv', 'r') as f:
+ 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.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.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