[ENG-7455] easy bulk-delete long-past harvest

share/management/commands/__init__.py

@@ -23,7 +23,7 @@ def rollback_unless_commit(self, commit):
         except Rollback:
             pass
 
-    def input_confirm(self, prompt, default=None):
+    def input_confirm(self, prompt, default=None) -> bool:
         result = input(prompt)
         if not result and default is not None:
             return default

share/management/commands/delete_pretrove_data.py

@@ -0,0 +1,38 @@
+from django.db.models import OuterRef, Exists
+from django.utils.translation import gettext as _
+
+from share.management.commands import BaseShareCommand
+from share import models as _db
+
+
+class Command(BaseShareCommand):
+    def add_arguments(self, parser):
+        parser.add_argument('--really-really', action='store_true', help='skip final confirmation prompt before really deleting')
+
+    def handle(self, *args, really_really: bool, **kwargs):
+        _pretrove_configs = _db.SourceConfig.objects.exclude(transformer_key='rdf')
+        _pretrove_configs_with_rawdata = (
+            _pretrove_configs
+            .annotate(has_rawdata=Exists(
+                _db.RawDatum.objects
+                .filter(suid__source_config_id=OuterRef('pk'))
+            ))
+            .filter(has_rawdata=True)
+        )
+        if not _pretrove_configs_with_rawdata.exists():
+            self.stdout.write(self.style.SUCCESS(_('nothing to delete')))
+            return
+        self.stdout.write(self.style.WARNING(_('pre-trove source-configs with deletable rawdata:')))
+        for _label in _pretrove_configs_with_rawdata.values_list('label', flat=True):
+            self.stdout.write(f'\t{_label}')
+        if really_really or self.input_confirm(self.style.WARNING(_('really DELETE ALL raw metadata records belonging to these source-configs? (y/n)'))):
+            self.stdout.write(_('deleting...'))
+            _rawdata_to_delete = (
+                _db.RawDatum.objects
+                .filter(suid__source_config_id__in=_pretrove_configs)
+            )
+            _deleted_total, _deleted_counts = _rawdata_to_delete.delete()
+            for _name, _count in _deleted_counts.items():
+                self.stdout.write(self.style.SUCCESS(f'{_name}: deleted {_count}'))
+        else:
+            self.stdout.write(self.style.SUCCESS('deleted nothing'))

trove/migrations/0007_rawdata_fks_do_nothing.py

@@ -0,0 +1,30 @@
+# Generated by Django 3.2.25 on 2025-04-21 14:59
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('share', '0077_big_cleanup_2025'),
+        ('trove', '0006_supplementary_indexcard_rdf'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='archivedindexcardrdf',
+            name='from_raw_datum',
+            field=models.ForeignKey(on_delete=django.db.models.deletion.DO_NOTHING, related_name='+', to='share.rawdatum'),
+        ),
+        migrations.AlterField(
+            model_name='latestindexcardrdf',
+            name='from_raw_datum',
+            field=models.ForeignKey(on_delete=django.db.models.deletion.DO_NOTHING, related_name='+', to='share.rawdatum'),
+        ),
+        migrations.AlterField(
+            model_name='supplementaryindexcardrdf',
+            name='from_raw_datum',
+            field=models.ForeignKey(on_delete=django.db.models.deletion.DO_NOTHING, related_name='+', to='share.rawdatum'),
+        ),
+    ]

trove/models/indexcard.py

@@ -295,7 +295,7 @@ class IndexcardRdf(models.Model):
     # required:
     from_raw_datum = models.ForeignKey(
         share_db.RawDatum,
-        on_delete=models.CASCADE,
+        on_delete=models.DO_NOTHING,  # allows faster bulk-deletion of unused RawDatum (but errors deleting used RawDatum)
         related_name='+',
     )
     indexcard = models.ForeignKey(