refactor: prefix doi workflows with doi_ for consistency

- bump deps for datacite schema 4.5 and django cve

django/library/doi.py

@@ -18,7 +18,7 @@
     DataCiteRegistrationLog,
 )
 
-from datacite import DataCiteRESTClient, schema43
+from datacite import DataCiteRESTClient, schema45
 from datacite.errors import (
     DataCiteError,
     DataCiteNoContentError,
@@ -183,7 +183,7 @@ def _datacite_heartbeat_url(self):
 
     def _validate_metadata(self, datacite_metadata: DataCiteSchema):
         metadata_dict = datacite_metadata.to_dict()
-        if not schema43.validate(metadata_dict):
+        if not schema45.validate(metadata_dict):
             logger.error("Invalid DataCite metadata: %s", metadata_dict)
             raise DataCiteError(f"Invalid DataCite metadata: {metadata_dict}")
         return datacite_metadata, metadata_dict
@@ -202,16 +202,16 @@ def mint_public_doi(self, codebase_or_release: Codebase | CodebaseRelease):
             return "XX.DRYXX/XXXX-XRUN", True
         if hasattr(codebase_or_release, "datacite"):
             del codebase_or_release.datacite
-        datacite_metadata, metadata_dict = self._validate_metadata(
-            codebase_or_release.datacite
-        )
+
         doi = "Unassigned"
         http_status = 200
         message = "Minted new DOI successfully."
 
+        datacite_metadata = codebase_or_release.datacite
+
         try:
             doi = self.datacite_client.public_doi(
-                metadata_dict, url=codebase_or_release.permanent_url
+                datacite_metadata.to_dict(), url=codebase_or_release.permanent_url
             )
         except DataCiteError as e:
             logger.error(e)

django/library/management/commands/clean_peer_reviewed_dois_02.py → django/library/management/commands/doi_remove_unreviewed_dois_02.py

@@ -1,4 +1,5 @@
 import logging
+import sys
 from django.core.management.base import BaseCommand
 from library.doi import VERIFICATION_MESSAGE, get_welcome_message
 from library.models import CodebaseRelease
@@ -9,9 +10,7 @@
 def remove_dois_from_unreviewed_releases(interactive=True, dry_run=True):
     print(get_welcome_message(dry_run))
 
-    unreviewed_releases_with_dois = CodebaseRelease.objects.filter(
-        peer_reviewed=False, doi__isnull=False
-    )
+    unreviewed_releases_with_dois = CodebaseRelease.objects.unreviewed().with_doi()
     total_unreviewed_releases_with_dois = unreviewed_releases_with_dois.count()
 
     logger.info(
@@ -24,23 +23,21 @@ def remove_dois_from_unreviewed_releases(interactive=True, dry_run=True):
         )
         if confirm.lower() == "delete":
             unreviewed_releases_with_dois.update(doi=None)
+        else:
+            logger.debug("Aborting...")
+            sys.exit()
 
     """
     assert correctness
     """
     if not dry_run:
         print(VERIFICATION_MESSAGE)
         logger.info(
-            "Checking that DOIs for all not peer reviewed releases have been deleted..."
-        )
-        assert (
-            CodebaseRelease.objects.filter(
-                peer_reviewed=False, doi__isnull=False
-            ).count()
-            == 0
+            "Checking that DOIs for all unreviewed releases have been deleted..."
         )
+        assert not CodebaseRelease.objects.unreviewed().with_doi().exists()
         logger.info(
-            "All DOIs from not peer_reviewed CodebaseReleases %s with DOIs deleted successfully.",
+            "%s unreviewed CodebaseReleases with DOIs updated successfully.",
             total_unreviewed_releases_with_dois,
         )
 

django/library/management/commands/update_metadata_for_all_existing_dois_04.py → django/library/management/commands/doi_update_doi_metadata_04.py

@@ -12,18 +12,19 @@ def update_doi_metadata(interactive=True, dry_run=True):
 
     datacite_api = DataCiteApi(dry_run=dry_run)
     all_codebases_with_dois = Codebase.objects.with_doi()
+    total_number_of_codebases_with_dois = all_codebases_with_dois.count()
 
     logger.info(
         "Updating metadata for all codebases (%s) with DOIs and their releases with DOIs. ...",
-        all_codebases_with_dois.count(),
+        total_number_of_codebases_with_dois,
     )
 
     for i, codebase in enumerate(all_codebases_with_dois):
         logger.debug(
             "Processing codebase %s - %s/%s",
             codebase.pk,
             i + 1,
-            all_codebases_with_dois.count(),
+            total_number_of_codebases_with_dois,
         )
         if interactive:
             input("Press Enter to continue or CTRL+C to quit...")
@@ -68,39 +69,38 @@ def update_doi_metadata(interactive=True, dry_run=True):
     """
     if not dry_run:
         print(VERIFICATION_MESSAGE)
-        logger.info("Checking that Comses metadata is in sync with DataCite...")
-        invalid_codebases = []
+        logger.info("Checking that local metadata is in sync with DataCite...")
         invalid_releases = []
 
         results = datacite_api.threaded_metadata_check(all_codebases_with_dois)
-        for pk, is_meta_valid in results:
-            if not is_meta_valid:
-                invalid_codebases.append(pk)
-
+        invalid_codebases = [
+            pk for pk, is_valid_metadata in results if not is_valid_metadata
+        ]
         if invalid_codebases:
             logger.error(
-                "Failure. Metadata not in sync with DataCite for %s codebases: %s",
+                "FAILURE: Metadata not in sync with DataCite for %s codebases: %s",
                 invalid_codebases.count(),
                 invalid_codebases,
             )
         else:
             logger.info(
-                "Success. Metadata in sync with DataCite for all codebases with DOI."
+                "SUCCESS: Metadata in sync with DataCite for all codebases with DOI."
             )
 
         all_releases_with_dois = CodebaseRelease.objects.with_doi()
         results = datacite_api.threaded_metadata_check(all_releases_with_dois)
-        for pk, is_meta_valid in results:
-            if not is_meta_valid:
-                invalid_releases.append(pk)
-
+        invalid_releases = [
+            pk for pk, is_valid_metadata in results if not is_valid_metadata
+        ]
         if invalid_releases:
             logger.error(
-                f"Failure. Metadata not in sync with DataCite for {len(invalid_releases)} releases: {invalid_releases}"
+                "FAILURE: Metadata not in sync with DataCite for %s releases: %s",
+                invalid_releases.count(),
+                invalid_releases,
             )
         else:
             logger.info(
-                f"Success. Metadata in sync with DataCite for all releases with DOI."
+                "SUCCESS: Metadata in sync with DataCite for all releases with DOI."
             )
 
 

django/library/migrations/0031_dataciteregistrationlog_contributor_affiliation.py

@@ -0,0 +1,92 @@
+# Generated by Django 4.2.15 on 2024-10-25 23:47
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("library", "0030_peerreviewinvitation"),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name="DataCiteRegistrationLog",
+            fields=[
+                (
+                    "id",
+                    models.AutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                (
+                    "action",
+                    models.CharField(
+                        choices=[
+                            ("CREATE_RELEASE_DOI", "create release DOI"),
+                            ("CREATE_CODEBASE_DOI", "create codebase DOI"),
+                            ("UPDATE_RELEASE_METADATA", "update release metadata"),
+                            ("UPDATE_CODEBASE_METADATA", "update codebase metadata"),
+                        ],
+                        max_length=50,
+                    ),
+                ),
+                ("timestamp", models.DateTimeField(auto_now_add=True)),
+                ("http_status", models.IntegerField(default=None, null=True)),
+                ("message", models.TextField(default=None, null=True)),
+                ("metadata_hash", models.CharField(max_length=255)),
+                ("doi", models.CharField(blank=True, max_length=255, null=True)),
+            ],
+        ),
+        migrations.RemoveField(
+            model_name="contributor",
+            name="affiliations",
+        ),
+        migrations.AlterField(
+            model_name="codebase",
+            name="date_created",
+            field=models.DateTimeField(auto_now_add=True),
+        ),
+        migrations.AlterField(
+            model_name="codebaserelease",
+            name="date_created",
+            field=models.DateTimeField(auto_now_add=True),
+        ),
+        migrations.AlterField(
+            model_name="codebasereleasedownload",
+            name="date_created",
+            field=models.DateTimeField(auto_now_add=True),
+        ),
+        migrations.AlterField(
+            model_name="peerreviewinvitation",
+            name="date_sent",
+            field=models.DateTimeField(auto_now=True),
+        ),
+        migrations.DeleteModel(
+            name="ContributorAffiliation",
+        ),
+        migrations.AddField(
+            model_name="dataciteregistrationlog",
+            name="codebase",
+            field=models.ForeignKey(
+                null=True,
+                on_delete=django.db.models.deletion.CASCADE,
+                related_name="datacite_logs",
+                to="library.codebase",
+            ),
+        ),
+        migrations.AddField(
+            model_name="dataciteregistrationlog",
+            name="release",
+            field=models.ForeignKey(
+                null=True,
+                on_delete=django.db.models.deletion.CASCADE,
+                related_name="datacite_logs",
+                to="library.codebaserelease",
+            ),
+        ),
+    ]

django/library/models.py

@@ -1132,6 +1132,9 @@ def accessible(self, user):
     def reviewed(self, **kwargs):
         return self.filter(peer_reviewed=True, **kwargs)
 
+    def unreviewed(self, **kwargs):
+        return self.exclude(peer_reviewed=True).filter(**kwargs)
+
     def with_doi(self, **kwargs):
         return self.exclude(Q(doi__isnull=True) | Q(doi="")).filter(**kwargs)
 

django/requirements.txt

@@ -1,6 +1,6 @@
 bagit==1.8.1
 bleach==6.1.0
-datacite==1.1.4
+datacite==1.2.0
 dedupe==3.0.2
 django-allauth==0.63.6
 django-anymail[mailgun]==10.3
@@ -24,7 +24,7 @@ django-vite==2.1.3  # latest is 3.0.4
 django-waffle==4.1.0
 djangorestframework==3.15.2
 djangorestframework-camel-case==1.4.2
-Django==4.2.15
+Django==4.2.16
 elasticsearch-dsl>=7.0.0,<8.0.0
 elasticsearch>=7.0.0,<8.0.0
 html2text>=2016.9.19