legacy: serialize custom fields

zenodo · Oct 31, 2023 · 75e4c98 · 75e4c98
1 parent dde3572
commit 75e4c98
Show file tree

Hide file tree

Showing 3 changed files with 153 additions and 0 deletions.
diff --git a/site/tests/legacy/deposits/conftest.py b/site/tests/legacy/deposits/conftest.py
@@ -110,6 +110,71 @@ def test_data():
             ],
             title="Test title",
             upload_type="publication",
+            custom={
+                # dwc
+                "dwc:basisOfRecord": ["foo", "bar"],
+                "dwc:catalogNumber": ["foo", "bar"],
+                "dwc:class": ["foo", "bar"],
+                "dwc:collectionCode": ["MUSM", "ZMPC"],
+                "dwc:country": ["foo", "bar"],
+                "dwc:county": ["foo", "bar"],
+                "dwc:dateIdentified": ["2011-05-17", "2022-08-18"],
+                "dwc:decimalLatitude": [45.90621, -33.85249],
+                "dwc:decimalLongitude": [6.1267, 151.21046],
+                "dwc:eventDate": ["2011-05-17", "2022-08-18"],
+                "dwc:family": ["Cerambycidae"],
+                "dwc:genus": ["Asynapteron"],
+                "dwc:identifiedBy": ["foo", "bar"],
+                "dwc:individualCount": ["foo", "bar"],
+                "dwc:institutionCode": ["foo", "bar"],
+                "dwc:kingdom": ["Animalia"],
+                "dwc:lifeStage": ["foo", "bar"],
+                "dwc:locality": ["foo", "bar"],
+                "dwc:materialSampleID": ["foo", "bar"],
+                "dwc:namePublishedInID": ["foo", "bar"],
+                "dwc:namePublishedInYear": ["foo", "bar"],
+                "dwc:order": ["Coleoptera"],
+                "dwc:otherCatalogNumbers": ["foo", "bar"],
+                "dwc:phylum": ["Arthropoda"],
+                "dwc:preparations": ["foo", "bar"],
+                "dwc:recordedBy": ["foo", "bar"],
+                "dwc:scientificName": ["foo", "bar"],
+                "dwc:scientificNameAuthorship": [
+                    "Ju\u00e1rez-No\u00e9 & Gonz\u00e1lez-Coronado"
+                ],
+                "dwc:scientificNameID": ["foo", "bar"],
+                "dwc:sex": ["foo", "bar"],
+                "dwc:specificEpithet": ["andinum"],
+                "dwc:stateProvince": ["foo", "bar"],
+                "dwc:taxonID": ["foo", "bar"],
+                "dwc:taxonRank": ["species"],
+                "dwc:taxonomicStatus": ["sp. nov."],
+                "dwc:typeStatus": ["holotype"],
+                "dwc:verbatimElevation": ["foo", "bar"],
+                "dwc:verbatimEventDate": ["2011-05-17"],
+                # openbiodiv
+                "openbiodiv:TaxonomicConceptLabel": [
+                    "Asynapteron andinum Ju\u00e1rez-No\u00e9 & Gonz\u00e1lez-Coronado, 2023"
+                ],
+                # ac
+                "ac:associatedSpecimenReference": ["foo", "bar"],
+                "ac:captureDevice": ["foo", "bar"],
+                "ac:physicalSetting": ["foo", "bar"],
+                "ac:resourceCreationTechnique": ["foo", "bar"],
+                "ac:subjectOrientation": ["foo", "bar"],
+                "ac:subjectPart": ["foo", "bar"],
+                # dc
+                "dc:creator": ["foo", "bar"],
+                "dc:rightsHolder": ["foo", "bar"],
+                # obo
+                "obo:RO_0002453": [
+                    {"subject": ["foo", "bar"], "object": ["foo", "bar"]},
+                    {"subject": ["foo", "bar"], "object": ["foo", "bar"]},
+                ],
+                # gbif-dwc
+                "gbif-dwc:identifiedByID": ["foo", "bar"],
+                "gbif-dwc:recordedByID": ["foo", "bar"],
+            },
         )
     )
 
@@ -210,4 +275,69 @@ def expected_record_metadata():
         ],
         title="Test title",
         upload_type="publication",
+        custom={
+            # dwc
+            "dwc:basisOfRecord": ["foo", "bar"],
+            "dwc:catalogNumber": ["foo", "bar"],
+            "dwc:class": ["foo", "bar"],
+            "dwc:collectionCode": ["MUSM", "ZMPC"],
+            "dwc:country": ["foo", "bar"],
+            "dwc:county": ["foo", "bar"],
+            "dwc:dateIdentified": ["2011-05-17", "2022-08-18"],
+            "dwc:decimalLatitude": [45.90621, -33.85249],
+            "dwc:decimalLongitude": [6.1267, 151.21046],
+            "dwc:eventDate": ["2011-05-17", "2022-08-18"],
+            "dwc:family": ["Cerambycidae"],
+            "dwc:genus": ["Asynapteron"],
+            "dwc:identifiedBy": ["foo", "bar"],
+            "dwc:individualCount": ["foo", "bar"],
+            "dwc:institutionCode": ["foo", "bar"],
+            "dwc:kingdom": ["Animalia"],
+            "dwc:lifeStage": ["foo", "bar"],
+            "dwc:locality": ["foo", "bar"],
+            "dwc:materialSampleID": ["foo", "bar"],
+            "dwc:namePublishedInID": ["foo", "bar"],
+            "dwc:namePublishedInYear": ["foo", "bar"],
+            "dwc:order": ["Coleoptera"],
+            "dwc:otherCatalogNumbers": ["foo", "bar"],
+            "dwc:phylum": ["Arthropoda"],
+            "dwc:preparations": ["foo", "bar"],
+            "dwc:recordedBy": ["foo", "bar"],
+            "dwc:scientificName": ["foo", "bar"],
+            "dwc:scientificNameAuthorship": [
+                "Ju\u00e1rez-No\u00e9 & Gonz\u00e1lez-Coronado"
+            ],
+            "dwc:scientificNameID": ["foo", "bar"],
+            "dwc:sex": ["foo", "bar"],
+            "dwc:specificEpithet": ["andinum"],
+            "dwc:stateProvince": ["foo", "bar"],
+            "dwc:taxonID": ["foo", "bar"],
+            "dwc:taxonRank": ["species"],
+            "dwc:taxonomicStatus": ["sp. nov."],
+            "dwc:typeStatus": ["holotype"],
+            "dwc:verbatimElevation": ["foo", "bar"],
+            "dwc:verbatimEventDate": ["2011-05-17"],
+            # openbiodiv
+            "openbiodiv:TaxonomicConceptLabel": [
+                "Asynapteron andinum Ju\u00e1rez-No\u00e9 & Gonz\u00e1lez-Coronado, 2023"
+            ],
+            # ac
+            "ac:associatedSpecimenReference": ["foo", "bar"],
+            "ac:captureDevice": ["foo", "bar"],
+            "ac:physicalSetting": ["foo", "bar"],
+            "ac:resourceCreationTechnique": ["foo", "bar"],
+            "ac:subjectOrientation": ["foo", "bar"],
+            "ac:subjectPart": ["foo", "bar"],
+            # dc
+            "dc:creator": ["foo", "bar"],
+            "dc:rightsHolder": ["foo", "bar"],
+            # obo
+            "obo:RO_0002453": [
+                {"subject": ["foo", "bar"], "object": ["foo", "bar"]},
+                {"subject": ["foo", "bar"], "object": ["foo", "bar"]},
+            ],
+            # gbif-dwc
+            "gbif-dwc:identifiedByID": ["foo", "bar"],
+            "gbif-dwc:recordedByID": ["foo", "bar"],
+        },
     )
diff --git a/site/zenodo_rdm/legacy/deserializers/schemas.py b/site/zenodo_rdm/legacy/deserializers/schemas.py
@@ -129,6 +129,9 @@ def _imprint(data):
             "thesis:university": university,
         }
 
+        # Add the legacy custom fields
+        custom_fields.update(metadata.get("custom", {}))
+
         clear_none(custom_fields)
 
         if custom_fields:

diff --git a/site/zenodo_rdm/legacy/serializers/schemas/common.py b/site/zenodo_rdm/legacy/serializers/schemas/common.py
@@ -180,6 +180,8 @@ class MetadataSchema(Schema):
     references = fields.Method("dump_reference")
     language = fields.Method("dump_languages")
 
+    custom = fields.Method("dump_custom_fields")
+
     @pre_dump
     def resolve_communities(self, data, **kwargs):
         """Resolve communities for the draft/record."""
@@ -379,6 +381,24 @@ def dump_languages(self, obj):
         # Zenodo-RDM implements ISO-639-2 so no mapping is needed.
         return languages[0]["id"]
 
+    def dump_custom_fields(self, obj):
+        """Dump custom fields."""
+        custom_fields = obj.get("custom_fields", {})
+        return {
+            k: v
+            for k, v in custom_fields.items()
+            if not k.startswith(
+                (
+                    "legacy:",
+                    "thesis:",
+                    "imprint:",
+                    "meeting:",
+                    "journal:",
+                    "part_of:",
+                )
+            )
+        } or missing
+
 
 class LegacySchema(Schema):
     """Legacy schema."""