html: work around firstmonday DOCTYPE issue

1 files changed, 14 insertions, 0 deletions

diff --git a/python/tests/test_html_ingest.py b/python/tests/test_html_ingest.py
new file mode 100644
index 0000000..96b3883
--- /dev/null
+++ b/python/tests/test_html_ingest.py
@@ -0,0 +1,14 @@
+
+import datetime
+import pytest
+
+from sandcrawler.html_ingest import *
+
+
+def test_html_extract_ojs3() -> None:
+
+    with open('tests/files/first_monday_ojs3_fulltext.html', 'rb') as f:
+        ojs3_html = f.read()
+
+    fulltext = html_extract_fulltext_teixml(ojs3_html)
+    assert fulltext['status'] == 'success'