[feature] Add a Data Package serializer

osf/metadata/serializers/__init__.py

@@ -7,6 +7,7 @@
 from osf import exceptions
 from osf.metadata import gather
 from .datacite import DataciteJsonMetadataSerializer, DataciteXmlMetadataSerializer
+from .datapackage import DatapackageSerializer
 from .google_dataset_json_ld import GoogleDatasetJsonLdSerializer
 from .turtle import TurtleMetadataSerializer
 
@@ -15,6 +16,7 @@
  'turtle': TurtleMetadataSerializer,
  'datacite-json': DataciteJsonMetadataSerializer,
  'datacite-xml': DataciteXmlMetadataSerializer,
+ 'datapackage.json': DatapackageSerializer,
  'google-dataset-json-ld': GoogleDatasetJsonLdSerializer,
 }
 

osf/metadata/serializers/datapackage.py

@@ -0,0 +1,99 @@
+import json
+from osf.metadata.serializers import _base
+from osf.metadata.serializers.datacite import DataciteJsonMetadataSerializer
+
+PROFILE_URL = "https://datapackage.org/profiles/2.0/datapackage.json"
+
+
+class DatapackageSerializer(_base.MetadataSerializer):
+ mediatype = "application/json" # type: ignore
+
+ def filename_for_itemid(self, itemid: str):
+ return f"{itemid}.datapackage.json"
+
+ def serialize(self) -> str:
+ return json.dumps(
+ self.metadata_as_dict(),
+ indent=2,
+ sort_keys=True,
+ )
+
+ def metadata_as_dict(self) -> dict:
+ dataset = DataciteJsonMetadataSerializer(self.basket).metadata_as_dict()
+ package = {"$schema": PROFILE_URL, "resources": []}
+
+ # Id
+ for identifier in dataset.get("identifiers", []):
+ type = identifier.get("identifierType")
+ value = identifier.get("identifier")
+ if value and type == "DOI":
+ package["id"] = value
+ break
+
+ # Title
+ for title in dataset.get("titles", []):
+ type = title.get("titleType")
+ value = title.get("title")
+ if value and not type:
+ package["title"] = value
+ break
+
+ # Description
+ for title in dataset.get("descriptions", []):
+ type = title.get("descriptionType")
+ value = title.get("description")
+ if value and type == "Abstract":
+ package["description"] = value
+ break
+
+ # Homepage
+ for identifier in dataset.get("identifiers", []):
+ type = identifier.get("identifierType")
+ value = identifier.get("identifier")
+ if value and type == "URL":
+ package["homepage"] = value
+ break
+
+ # Version
+ version = dataset.get("version")
+ if version:
+ package["version"] = version
+
+ # Keywords
+ for subject in dataset.get("subjects", []):
+ value = subject.get("subject")
+ if value:
+ package.setdefault("keywords", []).append(value)
+
+ # Licenses
+ for right in dataset.get("rightsList", []):
+ license = {}
+ license["name"] = right.get("rightsIdentifier")
+ license["path"] = right.get("rightsUri")
+ license["title"] = right.get("rights")
+ license = {k: v for k, v in license.items() if v is not None}
+ if license:
+ package.setdefault("licenses", []).append(license)
+
+ # Contributors
+ creators = dataset.get("creators", [])
+ contributors = dataset.get("contributors", [])
+ for type, items in [("creator", creators), (None, contributors)]:
+ for item in items:
+ type = item.get("contributorType", type)
+ contributor = {}
+ contributor["title"] = item.get("name")
+ contributor["givenName"] = item.get("givenName")
+ contributor["familyName"] = item.get("familyName")
+ if type:
+ contributor["roles"] = [type]
+ for affiliation in item.get("affiliations", []):
+ name = affiliation.get("name")
+ if name:
+ contributor["organization"] = name
+ break
+ contributor = {k: v for k, v in contributor.items() if v is not None}
+ if contributor:
+ package.setdefault("contributors", []).append(contributor)
+
+ return package