Add schemaorg serializer

invenio_rdm_records/contrib/journal/processors.py

@@ -154,3 +154,35 @@ def post_dump(self, data, original=None, **kwargs):
             data["ISSN"] = issn
 
         return data
+
+
+class JournalSchemaorgDumper(DumperMixin):
+    """Dumper for Schemaorg serialization of 'Journal' custom field."""
+
+    def post_dump(self, data, original=None, **kwargs):
+        """Adds serialized journal data to the input data."""
+        _original = original or {}
+        custom_fields = _original.get("custom_fields", {})
+        journal_data = custom_fields.get("journal:journal", {})
+
+        if not journal_data:
+            return data
+
+        title = journal_data.get("title")
+        volume = journal_data.get("volume")
+        issue = journal_data.get("issue")
+        pages = journal_data.get("pages")
+        issn = journal_data.get("issn")
+
+        if title:
+            data["container_title"] = title
+        if pages:
+            data["page"] = pages
+        if volume:
+            data["volume"] = volume
+        if issue:
+            data["issue"] = issue
+        if issn:
+            data["ISSN"] = issn
+
+        return data

invenio_rdm_records/resources/config.py

@@ -54,6 +54,7 @@
     DublinCoreXMLSerializer,
     GeoJSONSerializer,
     MARCXMLSerializer,
+    SchemaorgJSONLDSerializer,
     StringCitationSerializer,
     UIJSONSerializer,
 )
@@ -78,6 +79,7 @@ def _bibliography_headers(obj_or_list, code, many=False):
 
 record_serializers = {
     "application/json": ResponseHandler(JSONSerializer(), headers=etag_headers),
+    "application/ld+json": ResponseHandler(SchemaorgJSONLDSerializer()),
     "application/marcxml+xml": ResponseHandler(
         MARCXMLSerializer(), headers=etag_headers
     ),

invenio_rdm_records/resources/serializers/__init__.py

@@ -22,6 +22,7 @@
     IIIFSequenceV2JSONSerializer,
 )
 from .marcxml import MARCXMLSerializer
+from .schemaorg import SchemaorgJSONLDSerializer
 from .ui import UIJSONSerializer
 
 __all__ = (
@@ -37,6 +38,7 @@
     "IIIFManifestV2JSONSerializer",
     "IIIFSequenceV2JSONSerializer",
     "MARCXMLSerializer",
+    "SchemaorgJSONLDSerializer",
     "StringCitationSerializer",
     "UIJSONSerializer",
     "DCATSerializer",

invenio_rdm_records/resources/serializers/schemaorg/__init__.py

@@ -0,0 +1,27 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2023 CERN.
+#
+# Invenio-RDM-Records is free software; you can redistribute it and/or modify
+# it under the terms of the MIT License; see LICENSE file for more details.
+
+"""Schemaorg Serializers for Invenio RDM Records."""
+from flask_resources import BaseListSchema, MarshmallowSerializer
+from flask_resources.serializers import JSONSerializer
+
+from ....contrib.journal.processors import JournalSchemaorgDumper
+from .schema import SchemaorgSchema
+
+
+class SchemaorgJSONLDSerializer(MarshmallowSerializer):
+    """Marshmallow based Schemaorg serializer for records."""
+
+    def __init__(self, **options):
+        """Constructor."""
+        super().__init__(
+            format_serializer_cls=JSONSerializer,
+            object_schema_cls=SchemaorgSchema,
+            list_schema_cls=BaseListSchema,
+            schema_kwargs={"dumpers": [JournalSchemaorgDumper()]},  # Order matters
+            **options
+        )