Register unversioned DOI alongside versioned on Crossref submissions

Author: Vlad0n20

api/crossref/views.py

@@ -7,6 +7,7 @@
 
 from api.crossref.permissions import RequestComesFromMailgun
 from osf.models import Preprint, NotificationTypeEnum
+from osf.models.base import Guid
 from website import settings
 from website.preprints.tasks import mint_doi_on_crossref_fail
 
@@ -48,6 +49,12 @@ def post(self, request):
                 if record.get('status').lower() == 'success' and doi:
                     msg = record.find('msg').text
                     created = bool(msg == 'Successfully added')
+                    _, version = Guid.split_guid(guid) if guid else (None, None)
+                    if not version:
+                        logger.info(f'Unversioned DOI confirmed by CrossRef (no identifier update needed): {doi}')
+                        dois_processed += 1
+                        continue
+
                     legacy_doi = preprint.get_identifier(category='legacy_doi')
                     if created or legacy_doi:
                         # Sets preprint_doi_created and saves the preprint

api_tests/crossref/views/test_crossref_email_response.py

@@ -220,3 +220,46 @@ def test_confirmation_marks_legacy_doi_as_deleted(self, app, url, preprint):
             app.post(url, context_data)
 
         assert preprint.identifiers.get(category='legacy_doi').deleted
+
+    def test_unversioned_doi_confirmation_skips_identifier_update(self, app, url, preprint):
+        versioned_doi = settings.DOI_FORMAT.format(
+            prefix=preprint.provider.doi_prefix, guid=preprint._id
+        )
+        preprint.set_identifier_value(category='doi', value=versioned_doi)
+
+        base_guid = preprint.get_guid()._id  # no _vN suffix
+        unversioned_doi = settings.DOI_FORMAT.format(
+            prefix=preprint.provider.doi_prefix, guid=base_guid
+        )
+        dual_confirmation_xml = """
+            <?xml version="1.0" encoding="UTF-8"?>
+            <doi_batch_diagnostic status="completed" sp="cs3.crossref.org">
+               <submission_id>1390675999</submission_id>
+               <batch_id>{batch_id}</batch_id>
+               <record_diagnostic status="Success">
+                  <doi>{versioned_doi}</doi>
+                  <msg>Successfully updated</msg>
+               </record_diagnostic>
+               <record_diagnostic status="Success">
+                  <doi>{unversioned_doi}</doi>
+                  <msg>Successfully added</msg>
+               </record_diagnostic>
+               <batch_data>
+                  <record_count>2</record_count>
+                  <success_count>2</success_count>
+                  <warning_count>0</warning_count>
+                  <failure_count>0</failure_count>
+               </batch_data>
+            </doi_batch_diagnostic>
+        """.format(
+            batch_id=preprint._id,
+            versioned_doi=versioned_doi,
+            unversioned_doi=unversioned_doi,
+        )
+
+        context_data = self.make_mailgun_payload(crossref_response=dual_confirmation_xml)
+        with capture_notifications(expect_none=True):
+            app.post(url, context_data)
+
+        preprint.reload()
+        assert preprint.get_identifier_value('doi') == versioned_doi

tests/identifiers/test_crossref.py

@@ -95,6 +95,18 @@ def test_crossref_build_doi(self, crossref_client, preprint):
 
         assert crossref_client.build_doi(preprint) == settings.DOI_FORMAT.format(prefix=doi_prefix, guid=preprint._id)
 
+    def test_crossref_build_unversioned_doi(self, crossref_client, preprint):
+        doi_prefix = preprint.provider.doi_prefix
+        base_guid = preprint.get_guid()._id
+
+        expected = settings.DOI_FORMAT.format(prefix=doi_prefix, guid=base_guid)
+        assert crossref_client.build_unversioned_doi(preprint) == expected
+        assert '_v' not in expected
+
+    def test_crossref_build_unversioned_doi_matches_base_guid_not_versioned(self, crossref_client, preprint, preprint_version):
+        assert crossref_client.build_doi(preprint_version) != crossref_client.build_unversioned_doi(preprint_version)
+        assert crossref_client.build_unversioned_doi(preprint) == crossref_client.build_unversioned_doi(preprint_version)
+
     def test_crossref_build_doi_versioned(self, crossref_client, preprint_version):
         doi_prefix = preprint_version.provider.doi_prefix
 
@@ -338,6 +350,57 @@ def test_metadata_for_non_included_relation(self, crossref_client, preprint):
         root_without_relation = lxml.etree.fromstring(xml_without_relation)
         assert root_without_relation.find('.//{%s}intra_work_relation' % crossref.CROSSREF_RELATIONS) is None
 
+    def test_metadata_includes_unversioned_doi_entry(self, crossref_client, preprint):
+        crossref_xml = crossref_client.build_metadata(preprint, include_unversioned_doi=True)
+        root = lxml.etree.fromstring(crossref_xml)
+
+        posted_contents = root.findall('.//{%s}posted_content' % crossref.CROSSREF_NAMESPACE)
+        assert len(posted_contents) == 2
+
+        versioned_dois = posted_contents[0].findall('.//{%s}doi' % crossref.CROSSREF_NAMESPACE)
+        assert len(versioned_dois) == 1
+        assert versioned_dois[0].text == crossref_client.build_doi(preprint)
+
+        unversioned_dois = posted_contents[1].findall('.//{%s}doi' % crossref.CROSSREF_NAMESPACE)
+        assert len(unversioned_dois) == 1
+        assert unversioned_dois[0].text == crossref_client.build_unversioned_doi(preprint)
+        assert '_v' not in unversioned_dois[0].text
+
+        unversioned_resource = posted_contents[1].find('.//{%s}resource' % crossref.CROSSREF_NAMESPACE)
+        base_guid = preprint.get_guid()._id
+        assert unversioned_resource.text == settings.DOMAIN + base_guid
+
+    def test_metadata_unversioned_doi_uses_latest_version_metadata(self, crossref_client, preprint, preprint_version):
+        crossref_xml = crossref_client.build_metadata(preprint, include_unversioned_doi=True)
+        root = lxml.etree.fromstring(crossref_xml)
+
+        posted_contents = root.findall('.//{%s}posted_content' % crossref.CROSSREF_NAMESPACE)
+        assert len(posted_contents) == 2
+
+        unversioned_resource = posted_contents[1].find('.//{%s}resource' % crossref.CROSSREF_NAMESPACE)
+        base_guid = preprint.get_guid()._id
+        assert unversioned_resource.text == settings.DOMAIN + base_guid
+
+    def test_metadata_unversioned_doi_has_no_version_relations(self, crossref_client, preprint, preprint_version):
+        crossref_xml = crossref_client.build_metadata(preprint_version, include_unversioned_doi=True)
+        root = lxml.etree.fromstring(crossref_xml)
+
+        posted_contents = root.findall('.//{%s}posted_content' % crossref.CROSSREF_NAMESPACE)
+        unversioned_content = posted_contents[1]
+
+        relations = unversioned_content.findall('.//{%s}intra_work_relation' % crossref.CROSSREF_RELATIONS)
+        is_version_of_relations = [
+            r for r in relations if r.get('relationship-type') == 'isVersionOf'
+        ]
+        assert len(is_version_of_relations) == 0
+
+    def test_metadata_bulk_does_not_include_unversioned_doi(self, crossref_client, preprint, preprint_version):
+        crossref_xml = crossref_client.build_metadata([preprint, preprint_version])
+        root = lxml.etree.fromstring(crossref_xml)
+
+        posted_contents = root.findall('.//{%s}posted_content' % crossref.CROSSREF_NAMESPACE)
+        assert len(posted_contents) == 2  # one per preprint, no extras
+
     def test_metadata_for_affiliated_institutions(self, crossref_client, preprint):
         institution = InstitutionFactory()
         institution.ror_uri = 'http://ror.org/WHATisITgoodFOR/'

website/identifiers/clients/crossref.py

@@ -41,13 +41,18 @@ def build_doi(self, preprint):
         prefix = preprint.provider.doi_prefix
         return settings.DOI_FORMAT.format(prefix=prefix, guid=preprint._id)
 
-    def build_metadata(self, preprint, include_relation=True):
+    def build_unversioned_doi(self, preprint):
+        prefix = preprint.provider.doi_prefix
+        return settings.DOI_FORMAT.format(prefix=prefix, guid=preprint.get_guid()._id)
+
+    def build_metadata(self, preprint, include_relation=True, include_unversioned_doi=False):
         """Return the crossref metadata XML document for a given preprint as a string for DOI minting purposes
 
         :param preprint: the preprint, or list of preprints to build metadata for
         """
         if isinstance(preprint, (list, QuerySet)):
             preprints = preprint
+            include_unversioned_doi = False  # not supported for bulk batches
         else:
             preprints = [preprint]
 
@@ -74,6 +79,9 @@ def build_metadata(self, preprint, include_relation=True):
         for preprint in preprints:
             body.append(self.build_posted_content(preprint, element, include_relation))
 
+        if include_unversioned_doi:
+            body.append(self.build_unversioned_posted_content(preprints[0], element))
+
         root = element.doi_batch(
             head,
             body,
@@ -82,10 +90,12 @@ def build_metadata(self, preprint, include_relation=True):
         root.attrib['{%s}schemaLocation' % XSI] = CROSSREF_SCHEMA_LOCATION
         return lxml.etree.tostring(root)
 
-    def build_posted_content(self, preprint, element, include_relation):
+    def build_posted_content(self, preprint, element, include_relation, doi_override=None, resource_override=None):
         """Build the <posted_content> element for a single preprint
         preprint - preprint to build posted_content for
         element - namespace element to use when building parts of the XML structure
+        doi_override - if provided, use this DOI value instead of the preprint's own DOI
+        resource_override - if provided, use this URL as the <resource> instead of the default
         """
         from osf.models import SpamStatus
 
@@ -157,15 +167,27 @@ def build_posted_content(self, preprint, element, include_relation):
             posted_content.append(relations_program)
 
         minted_doi = preprint.get_identifier_value('doi')
-        doi = minted_doi or self.build_doi(preprint)
+        doi = doi_override or minted_doi or self.build_doi(preprint)
+        resource_url = resource_override if resource_override is not None else settings.DOMAIN + preprint._id
         doi_data = [
             element.doi(doi),
-            element.resource(settings.DOMAIN + preprint._id)
+            element.resource(resource_url)
         ]
         posted_content.append(element.doi_data(*doi_data))
 
         return posted_content
 
+    def build_unversioned_posted_content(self, preprint, element):
+        latest = preprint.get_guid().referent
+        base_guid = latest.get_guid()._id
+        return self.build_posted_content(
+            latest,
+            element,
+            include_relation=False,
+            doi_override=self.build_unversioned_doi(latest),
+            resource_override=settings.DOMAIN + base_guid,
+        )
+
     def _process_crossref_name(self, contributor):
         # Adapted from logic used in `api/citations/utils.py`
         # If the user has a family and given name, use those
@@ -249,7 +271,7 @@ def _build_url(self, **query):
 
     def create_identifier(self, preprint, category, include_relation=True):
         if category == 'doi':
-            metadata = self.build_metadata(preprint, include_relation)
+            metadata = self.build_metadata(preprint, include_relation, include_unversioned_doi=True)
             doi = self.build_doi(preprint)
             username, password = self.get_credentials()
             logger.info(f'Sending metadata for DOI {doi}:\n{metadata}')