datacite: fix related identifiers serialization.

inveniosoftware · Oct 27, 2023 · 7bca45a · 7bca45a
1 parent 949b339
commit 7bca45a
Show file tree

Hide file tree

Showing 2 changed files with 36 additions and 11 deletions.
diff --git a/invenio_rdm_records/resources/serializers/datacite/schema.py b/invenio_rdm_records/resources/serializers/datacite/schema.py
@@ -27,6 +27,29 @@
 from ...serializers.ui.schema import current_default_locale
 from ..utils import get_preferred_identifier, get_vocabulary_props
 
+RELATED_IDENTIFIER_SCHEMES = {
+    "ark",
+    "arxiv",
+    "bibcode",
+    "doi",
+    "ean13",
+    "eissn",
+    "handle",
+    "igsn",
+    "isbn",
+    "issn",
+    "istc",
+    "lissn",
+    "lsid1",
+    "pmid",
+    "purl",
+    "upc",
+    "url",
+    "urn",
+    "w3id",
+}
+"""Allowed related identifier schemes for DataCite. Vocabulary taken from DataCite 4.3 schema definition."""
+
 
 def get_scheme_datacite(scheme, config_name, default=None):
     """Returns the datacite equivalent of a scheme."""
@@ -362,7 +385,8 @@ def get_related_identifiers(self, obj):
                 default=scheme,
             )
 
-            if id_scheme:
+            # Only serialize related identifiers with a valid scheme for DataCite.
+            if id_scheme and id_scheme.lower() in RELATED_IDENTIFIER_SCHEMES:
                 serialized_identifier = {
                     "relatedIdentifier": rel_id["identifier"],
                     "relationType": props.get("datacite", ""),
@@ -423,13 +447,14 @@ def get_related_identifiers(self, obj):
                     "RDM_RECORDS_IDENTIFIERS_SCHEMES",
                     default="doi",
                 )
-                serialized_identifiers.append(
-                    {
-                        "relatedIdentifier": parent_doi["identifier"],
-                        "relationType": "IsVersionOf",
-                        "relatedIdentifierType": id_scheme,
-                    }
-                )
+                if id_scheme.lower() in RELATED_IDENTIFIER_SCHEMES:
+                    serialized_identifiers.append(
+                        {
+                            "relatedIdentifier": parent_doi["identifier"],
+                            "relationType": "IsVersionOf",
+                            "relatedIdentifierType": id_scheme,
+                        }
+                    )
 
         # adding communities
         communities = obj.get("parent", {}).get("communities", {}).get("ids", [])

diff --git a/tests/resources/serializers/test_datacite_serializer.py b/tests/resources/serializers/test_datacite_serializer.py
@@ -365,8 +365,8 @@ def test_datacite43_serializer_with_unknown_id_schemes(
     assert expected_pid_id_2 in serialized_record["identifiers"]
     assert len(serialized_record["identifiers"]) == 5
 
-    assert expected_related_id in serialized_record["relatedIdentifiers"]
-    assert len(serialized_record["relatedIdentifiers"]) == 2
+    assert expected_related_id not in serialized_record["relatedIdentifiers"]
+    assert len(serialized_record["relatedIdentifiers"]) == 1
 
     creator_ids = serialized_record["creators"][0]["nameIdentifiers"]
     assert expected_creator_id in creator_ids
@@ -386,5 +386,5 @@ def test_datacite43_xml_serializer_with_unknown_id_schemes(
 
     assert expected_pid_id in serialized_record
     assert expected_pid_id_2 in serialized_record
-    assert expected_related_id in serialized_record
+    assert expected_related_id not in serialized_record
     assert expected_creator_id in serialized_record