aboutsummaryrefslogtreecommitdiffstats
path: root/python/tests/test_grobid2json.py
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2018-08-24 12:28:51 -0700
committerBryan Newbold <bnewbold@archive.org>2018-08-24 12:28:51 -0700
commit3782311e29b7e477e1936c89f55ff6483fd02e65 (patch)
treeb4484b9839f24d799d36881dfc85701ad888b94e /python/tests/test_grobid2json.py
parent2a998189ef49976bf01cc95acc1f18a73e1f0ff6 (diff)
downloadsandcrawler-3782311e29b7e477e1936c89f55ff6483fd02e65.tar.gz
sandcrawler-3782311e29b7e477e1936c89f55ff6483fd02e65.zip
rename ./mapreduce to ./python
Diffstat (limited to 'python/tests/test_grobid2json.py')
-rw-r--r--python/tests/test_grobid2json.py22
1 files changed, 22 insertions, 0 deletions
diff --git a/python/tests/test_grobid2json.py b/python/tests/test_grobid2json.py
new file mode 100644
index 0000000..8497b10
--- /dev/null
+++ b/python/tests/test_grobid2json.py
@@ -0,0 +1,22 @@
+
+import xml
+import json
+import pytest
+from grobid2json import *
+
+
+def test_small_xml():
+
+ with open('tests/files/small.xml', 'r') as f:
+ tei_xml = f.read()
+ with open('tests/files/small.json', 'r') as f:
+ json_form = json.loads(f.read())
+
+ assert teixml2json(tei_xml) == json_form
+
+def test_invalid_xml():
+
+ with pytest.raises(xml.etree.ElementTree.ParseError):
+ teixml2json("this is not XML")
+ with pytest.raises(ValueError):
+ teixml2json("<xml></xml>")