diff options
author | Bryan Newbold <bnewbold@archive.org> | 2020-11-03 22:40:14 -0800 |
---|---|---|
committer | Bryan Newbold <bnewbold@archive.org> | 2020-11-03 22:40:14 -0800 |
commit | 653fac9632c6ae9dd036ad844454cf419cd5320b (patch) | |
tree | c09d8a3d8a2524a991f082ab500bce53d1986caa /python/tests/test_xml.py | |
parent | 9beafd7c5fc98571ec26b49d223ce660378d7b9e (diff) | |
download | sandcrawler-653fac9632c6ae9dd036ad844454cf419cd5320b.tar.gz sandcrawler-653fac9632c6ae9dd036ad844454cf419cd5320b.zip |
xml: re-encode XML docs into UTF-8 for persisting
Diffstat (limited to 'python/tests/test_xml.py')
-rw-r--r-- | python/tests/test_xml.py | 18 |
1 files changed, 18 insertions, 0 deletions
diff --git a/python/tests/test_xml.py b/python/tests/test_xml.py new file mode 100644 index 0000000..a996c56 --- /dev/null +++ b/python/tests/test_xml.py @@ -0,0 +1,18 @@ + +import pytest + +from sandcrawler.xml import xml_reserialize + + +def test_xml_reserialize() -> None: + + with open('tests/files/scielo_article.jats.xml', 'rb') as f: + raw_xml = f.read() + + assert b'encoding="ISO-8859-1"' in raw_xml + raw_xml.decode("ISO-8859-1") + with pytest.raises(UnicodeDecodeError): + raw_xml.decode("utf-8") + + str_xml = xml_reserialize(raw_xml) + assert 'encoding="UTF-8"' in str_xml |