feat: add Crossref PID provider

* register DOIs with Crossref, functionality analogous to DataCitePIDProvider
* CrossrefXMLSerializer for metadata export
* includes test coverage similar to DataCite PID provider

Author: mfenner

invenio_rdm_records/config.py

@@ -513,11 +513,24 @@ def always_valid(identifier):
         client=providers.DataCiteClient("datacite", config_prefix="DATACITE"),
         label=_("DOI"),
     ),
+    # Crossref DOI provider
+    providers.CrossrefPIDProvider(
+        "crossref",
+        client=providers.CrossrefClient("crossref", config_prefix="CROSSREF"),
+        label=_("DOI"),
+    ),
     # DOI provider for externally managed DOIs
     providers.ExternalPIDProvider(
         "external",
         "doi",
-        validators=[providers.BlockedPrefixes(config_names=["DATACITE_PREFIX"])],
+        validators=[
+            providers.BlockedPrefixes(
+                config_names=[
+                    "DATACITE_PREFIX",
+                    "CROSSREF_PREFIX",
+                ]
+            )
+        ],
         label=_("DOI"),
     ),
     # OAI identifier
@@ -537,14 +550,15 @@ def always_valid(identifier):
 """
 
 RDM_PERSISTENT_IDENTIFIERS = {
-    # DOI automatically removed if DATACITE_ENABLED is False.
+    # DOI automatically removed if DATACITE_ENABLED and CROSSREF_ENABLED are False.
     "doi": {
-        "providers": ["datacite", "external"],
+        "providers": ["datacite", "crossref", "external"],
         "required": True,
         "label": _("DOI"),
         "validator": idutils.is_doi,
         "normalizer": idutils.normalize_doi,
-        "is_enabled": providers.DataCitePIDProvider.is_enabled,
+        "is_enabled": providers.DataCitePIDProvider.is_enabled
+        or providers.CrossrefPIDProvider.is_enabled,
         "ui": {"default_selected": "yes"},  # "yes", "no" or "not_needed"
     },
     "oai": {
@@ -602,7 +616,7 @@ def always_valid(identifier):
 # Configuration for the DataCiteClient used by the DataCitePIDProvider
 
 DATACITE_ENABLED = False
-"""Flag to enable/disable DOI registration."""
+"""Flag to enable/disable DataCite DOI registration."""
 
 DATACITE_USERNAME = ""
 """DataCite username."""
@@ -613,6 +627,9 @@ def always_valid(identifier):
 DATACITE_PREFIX = ""
 """DataCite DOI prefix."""
 
+DATACITE_ADDITIONAL_PREFIXES = []
+"""List of additional DataCite DOI prefixes supported for registration."""
+
 DATACITE_TEST_MODE = True
 """DataCite test mode enabled."""
 
@@ -638,6 +655,50 @@ def make_doi(prefix, record):
 in DataCite XML format.
 """
 
+# Configuration for the CrossrefClient used by the CrossrefPIDProvider
+
+CROSSREF_ENABLED = False
+"""Flag to enable/disable Crossref DOI registration."""
+
+CROSSREF_USERNAME = ""
+"""Crossref username."""
+
+CROSSREF_PASSWORD = ""
+"""Crossref password."""
+
+CROSSREF_PREFIX = ""
+"""Crossref DOI prefix."""
+
+CROSSREF_ADDITIONAL_PREFIXES = []
+"""List of additional Crossref DOI prefixes supported for registration."""
+
+CROSSREF_DEPOSITOR = ""
+"""Crossref depositor name."""
+
+CROSSREF_EMAIL = ""
+"""Crossref depositor email."""
+
+CROSSREF_REGISTRANT = ""
+"""Crossref registrant."""
+
+CROSSREF_TEST_MODE = True
+"""Crossref test mode enabled."""
+
+CROSSREF_FORMAT = "{prefix}/{id}"
+"""A string used for formatting the DOI or a callable.
+
+If set to a string, you can used ``{prefix}`` and ``{id}`` inside the string.
+
+You can also provide a callable instead:
+
+.. code-block:: python
+
+    def make_doi(prefix, record):
+        return f"{prefix}/{record.pid.pid_value}"
+
+    CROSSREF_FORMAT = make_doi
+"""
+
 #
 # Custom fields
 #

invenio_rdm_records/resources/serializers/__init__.py

@@ -19,6 +19,7 @@
 from .bibtex import BibtexSerializer
 from .cff import CFFSerializer
 from .codemeta import CodemetaSerializer
+from .crossref import CrossrefXMLSerializer
 from .csl import CSLJSONSerializer, StringCitationSerializer
 from .csv import CSVRecordSerializer
 from .datacite import (
@@ -48,6 +49,7 @@
 __all__ = (
     "BibtexSerializer",
     "CFFSerializer",
+    "CrossrefXMLSerializer",
     "CSLJSONSerializer",
     "CSVRecordSerializer",
     "DataCite43JSONSerializer",

invenio_rdm_records/resources/serializers/crossref/__init__.py

@@ -0,0 +1,110 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2021-2024 CERN.
+# Copyright (C) 2026 Front Matter.
+#
+# 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.
+
+"""Crossref Serializers for Invenio RDM Records."""
+
+from commonmeta import (
+    CrossrefError,
+    CrossrefXMLSchema,
+    Metadata,
+    tostring,
+    write_crossref_xml,
+)
+from flask import current_app
+from flask_resources import BaseListSchema, MarshmallowSerializer
+from flask_resources.serializers import SimpleSerializer
+
+
+class CrossrefXMLSerializer(MarshmallowSerializer):
+    """Marshmallow based Crossref XML serializer for records."""
+
+    def __init__(self, **options):
+        """Constructor."""
+        encoder = options.get("encoder", tostring)
+        super().__init__(
+            format_serializer_cls=SimpleSerializer,
+            object_schema_cls=CrossrefXMLSchema,
+            list_schema_cls=BaseListSchema,
+            encoder=encoder,
+            **options,
+        )
+
+    def serialize_object(self, obj):
+        """Serialize a single record to Crossref XML bytes.
+
+        Overrides the default to avoid double-encoding, since
+        ``dump_obj`` already returns XML bytes.
+        """
+        return self.dump_obj(obj)
+
+    def dump_obj(self, record, url=None):
+        """Dump a single record.
+
+        Config variables for Crossref XML head elements are used in the
+        XML head element.
+
+        :param record: Record instance (dict, Record model, or ChainObject).
+        :param url: the landing page URL for the DOI.
+            Falls back to ``SITE_UI_URL``/records/<id> if not provided.
+        """
+        # Determine the URL that the DOI resolves to, in the following order:
+        #
+        # 1. identifier of type url in ``metadata.identifiers``
+        #    (e.g. archived original content)
+        # 2. The landing page URL passed by the PID service
+        # 3. Default constructed from ``SITE_UI_URL`` and record ID
+        #    (e.g. for Celery tasks or tests without UI endpoints)
+        identifiers = (
+            record.get("metadata", {}).get("identifiers", [])
+            if isinstance(record, dict)
+            else getattr(getattr(record, "metadata", None), "get", lambda *a: [])(
+                "identifiers", []
+            )
+        )
+        registered_url = (
+            next(
+                (
+                    i.get("identifier")
+                    for i in (identifiers or [])
+                    if i.get("scheme") == "url" and i.get("identifier") is not None
+                ),
+                None,
+            )
+            or url
+        )
+
+        if registered_url is None:
+            site_url = current_app.config.get("SITE_UI_URL", "")
+            record_id = (
+                record.get("id")
+                if isinstance(record, dict)
+                else getattr(record, "id", None)
+            )
+            if site_url and record_id:
+                registered_url = f"{site_url}/records/{record_id}"
+
+        # Convert the metadata to crossref_xml format via the commonmeta intermediary format.
+        # XML Schema validation errors raise CrossrefError.
+        try:
+            metadata = Metadata(
+                record,
+                via="inveniordm",
+                url=registered_url,
+            )
+            crossref_xml = write_crossref_xml(metadata)
+            head = {
+                "depositor": current_app.config.get("CROSSREF_DEPOSITOR"),
+                "email": current_app.config.get("CROSSREF_EMAIL"),
+                "registrant": current_app.config.get("CROSSREF_REGISTRANT"),
+            }
+            return tostring(crossref_xml, head=head)
+        except CrossrefError as e:
+            current_app.logger.error(
+                f"CrossrefError while converting {metadata.id} to Crossref XML: {str(e)}"
+            )
+            return ""

invenio_rdm_records/services/pids/providers/__init__.py

@@ -9,12 +9,15 @@
 """PID Providers module."""
 
 from .base import PIDProvider
+from .crossref import CrossrefClient, CrossrefPIDProvider
 from .datacite import DataCiteClient, DataCitePIDProvider
 from .external import BlockedPrefixes, ExternalPIDProvider
 from .oai import OAIPIDProvider
 
 __all__ = (
     "BlockedPrefixes",
+    "CrossrefClient",
+    "CrossrefPIDProvider",
     "DataCiteClient",
     "DataCitePIDProvider",
     "ExternalPIDProvider",