Possible fix for #290

- Adds an upgrade command `edrn_nukephotos` which should be run once at upgrade time to nuke all the PI photos that have been duplicated over time
- Adjusts site ingest so that it checks for an existing photo first and re-uses it
- Changes name of photo to include not just surname but also ID number so we can better re-use photos
- Adds update steps to cbiit-deploy, cbiit-deploy-prod, and devrebuild

src/eke.knowledge/src/eke/knowledge/management/commands/edrn_nukephotos.py

@@ -0,0 +1,40 @@
+# encoding: utf-8
+
+'''😌 EDRN Site Content: nuke photos.'''
+
+from django.conf import settings
+from django.core.management.base import BaseCommand
+from wagtail.images.models import Image
+
+
+class Command(BaseCommand):
+    help = 'Delete specific instances of duplicate photos'
+
+    _to_nuke = (
+        'Photo of Chinnaiyan',
+        'Photo of Feng',
+        'Photo of Hanash',
+        'Photo of Semmes',
+        'Photo of Srivastava',
+        'Photo of Stass',
+    )
+
+    def nuke_photos(self):
+        for title in self._to_nuke:
+            count = Image.objects.filter(title=title).count()
+            if count > 0:
+                self.stdout.write(f'Deleting {count} instance(s) of {title}')
+                Image.objects.filter(title=title).delete()
+
+    def handle(self, *args, **options):
+        self.stdout.write('Deleting specific instances of duplicate photos')
+
+        old = getattr(settings, 'WAGTAILREDIRECTS_AUTO_CREATE', True)
+        try:
+            settings.WAGTAILREDIRECTS_AUTO_CREATE = False
+            settings.WAGTAILSEARCH_BACKENDS['default']['AUTO_UPDATE'] = False
+            self.nuke_photos()
+        finally:
+            settings.WAGTAILREDIRECTS_AUTO_CREATE = old
+            settings.WAGTAILSEARCH_BACKENDS['default']['AUTO_UPDATE'] = True
+            self.stdout.write("Job's done!")

src/eke.knowledge/src/eke/knowledge/migrations/0013_alter_person_photo.py

@@ -0,0 +1,20 @@
+# Generated by Django 4.1.9 on 2023-10-12 20:18
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('wagtailimages', '0025_alter_image_file_alter_rendition_file'),
+        ('ekeknowledge', '0012_datacollectionindex_preamble'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='person',
+            name='photo',
+            field=models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.CASCADE, related_name='person_photograph', to='wagtailimages.image'),
+        ),
+    ]

src/eke.knowledge/src/eke/knowledge/migrations/0014_alter_person_photo.py

@@ -0,0 +1,20 @@
+# Generated by Django 4.1.9 on 2023-10-12 21:17
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('wagtailimages', '0025_alter_image_file_alter_rendition_file'),
+        ('ekeknowledge', '0013_alter_person_photo'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='person',
+            name='photo',
+            field=models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.SET_NULL, related_name='person_photograph', to='wagtailimages.image'),
+        ),
+    ]

src/eke.knowledge/src/eke/knowledge/sites.py

@@ -96,23 +96,29 @@ def _get_file_name(self, url: str) -> str:
 
     def _get_surname(self, predicates: dict) -> str:
         name = str(predicates.get(_surname_predicate_uri, [''])[0])
-        if not name:
-            name = str(predicates.get(rdflib.term.URIRef('urn:internal:id'), [''])[0])
-            if not name:
-                name = '«unknown»'
-        return name
+        int_id = str(predicates.get(rdflib.term.URIRef('urn:internal:id'), [''])[0])
+        name = name if name else '«unknown name»'
+        int_id = int_id if int_id else '«unknown id»'
+        return f'{name} ({int_id})'
 
     def compute_new_value(self, modelField: Field, value: str, predicates: dict) -> object:
-        # Curiously, we can't pass a URL stream to ImageFile, since Django's ImageFile expects
-        # to be able to do seek() operations on it. So we have to download to a temporary file:
         name = self._get_surname(predicates)
+        image_title = f'Photograph of {name}'
+
+        # #290 check if the photo already exists and use it
+        existing_image = Image.objects.filter(title=image_title).first()
+        if existing_image: return existing_image
+
+        # Not found, so create it from scratch given the URL in ``value``
         try:
+            # Curiously, we can't pass a URL stream to ImageFile, since Django's ImageFile expects
+            # to be able to do seek() operations on it. So we have to download to a temporary file.
             with tempfile.TemporaryFile() as out_file:
                 with urlopen(value) as image_stream:
                     out_file.write(image_stream.read())
                 image_file = ImageFile(out_file, name=self._get_file_name(value))
                 image = Image(
-                    title=f'Photo of {name}', file=image_file,
+                    title=image_title, file=image_file,
                     # These values only make sense for the photos from the DMCC:
                     focal_point_x=66, focal_point_y=52, focal_point_height=76, focal_point_width=57
                 )

support/cbiit-deploy-prod.sh

@@ -137,7 +137,13 @@ docker compose --project-name edrn start portal" || exit 1
 echo ""
 echo "🆙 Applying upgrades"
 ssh -q $USER@$WEBSERVER "cd $WEBROOT ; \
-docker compose --project-name edrn exec portal django-admin edrn_explorer" || exit 1
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos" || exit 1
 
 echo ""
 echo "🤷‍♀️ Final portal restart and restart of search engine"

support/cbiit-deploy.sh

@@ -142,7 +142,13 @@ docker compose --project-name edrn start portal" || exit 1
 echo ""
 echo "🆙 Applying upgrades"
 ssh -q $USER@$WEBSERVER "cd $WEBROOT ; \
-docker compose --project-name edrn exec portal django-admin edrn_explorer" || exit 1
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos ||\
+docker compose --project-name edrn exec portal django-admin edrn_nukephotos" || exit 1
 
 echo ""
 echo "🤷‍♀️ Final portal restart and restart of search engine"

support/devrebuild.sh

@@ -6,7 +6,7 @@
 # Download the latest production database, apply migrations, and get ready to rock and roll.
 
 
-jpl_sys_ipv4=172.16.16.18
+jpl_sys_ipv4=172.16.16.25
 
 
 # Argument check
@@ -57,8 +57,7 @@ bzip2 --decompress --stdout edrn.sql.bz2 | psql --dbname=edrn --echo-errors --qu
 ./manage.sh edrndevreset
 
 # Add additional upgrade steps here:
-./manage.sh edrn_explorer
-
+./manage.sh edrn_nukephotos  # 6.7
 
 # This may be optional if you want to save time:
 ./manage.sh rdfingest