Parsers: include all CC BY variations

cern-sis · Jun 27, 2024 · 61707f5 · 61707f5
1 parent 7521c92
commit 61707f5
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 4 deletions.
diff --git a/dags/open_access/parsers.py b/dags/open_access/parsers.py
@@ -25,9 +25,7 @@ def is_correct_value(value):
 
 
 def field_has_cc_by(field_value):
-    # is CC BY-SA 4.0 falls under the condition of "contains ‘CC-BY’ or ‘CC BY’??
-    #
-    pattern = re.compile(r"CC(\s|-)?BY(\s|-)?4.0", flags=re.I)
+    pattern = re.compile(r"CC(\s|-)?BY(\s|-)?.*", flags=re.I)
     return bool(pattern.match(field_value))
 
 

diff --git a/tests/open_access/test_parser.py b/tests/open_access/test_parser.py
@@ -18,6 +18,7 @@
     "2882429",
     "2882335",
     "2882328",
+    "2882327",
     "2882324",
     "2882322",
     "2882311",
@@ -100,7 +101,7 @@ def test_parse_subset_540_publications(shared_datadir):
             )
             if is_it_wanted_record_by_540:
                 filtered_records_count = filtered_records_count + 1
-        assert filtered_records_count == 14
+        assert filtered_records_count == 15
 
 
 def test_get_green_access_records_dois(shared_datadir):