From 74f69a3813b95142c3ee14e2baacbebd9700af6b Mon Sep 17 00:00:00 2001
From: Andrew Godwin <andrew@aeracode.org>
Date: Sun, 12 Nov 2023 18:01:01 -0700
Subject: [PATCH] Add identity pruning, improve post pruning

---
 activities/management/commands/pruneposts.py | 18 +++++--
 takahe/settings.py                           |  6 +--
 users/management/__init__.py                 |  0
 users/management/commands/__init__.py        |  0
 users/management/commands/pruneidentities.py | 52 ++++++++++++++++++++
 5 files changed, 70 insertions(+), 6 deletions(-)
 create mode 100644 users/management/__init__.py
 create mode 100644 users/management/commands/__init__.py
 create mode 100644 users/management/commands/pruneidentities.py

diff --git a/activities/management/commands/pruneposts.py b/activities/management/commands/pruneposts.py
index c405c0c..8c3667c 100644
--- a/activities/management/commands/pruneposts.py
+++ b/activities/management/commands/pruneposts.py
@@ -3,6 +3,7 @@ import sys
 
 from django.conf import settings
 from django.core.management.base import BaseCommand
+from django.db.models import Q
 from django.utils import timezone
 
 from activities.models import Post
@@ -21,13 +22,21 @@ class Command(BaseCommand):
         )
 
     def handle(self, number: int, *args, **options):
+        if not settings.SETUP.REMOTE_PRUNE_HORIZON:
+            print("Pruning has been disabled as REMOTE_PRUNE_HORIZON=0")
+            sys.exit(2)
         # Find a set of posts that match the initial criteria
         print(f"Running query to find up to {number} old posts...")
         posts = Post.objects.filter(
             local=False,
             created__lt=timezone.now()
             - datetime.timedelta(days=settings.SETUP.REMOTE_PRUNE_HORIZON),
-        ).exclude(interactions__identity__local=True)[:number]
+        ).exclude(
+            Q(interactions__identity__local=True)
+            | Q(visibility=Post.Visibilities.mentioned)
+        )[
+            :number
+        ]
         post_ids_and_uris = dict(posts.values_list("object_uri", "id"))
         print(f"  found {len(post_ids_and_uris)}")
 
@@ -43,9 +52,12 @@ class Command(BaseCommand):
 
         # Delete them
         print(f"  down to {len(post_ids_and_uris)} to delete")
-        number_deleted, _ = Post.objects.filter(
+        print("Deleting...")
+        number_deleted, deleted = Post.objects.filter(
             id__in=post_ids_and_uris.values()
         ).delete()
-        print(f"Deleted {number_deleted} posts and dependencies")
+        print("Deleted:")
+        for model, model_deleted in deleted.items():
+            print(f"  {model}: {model_deleted}")
         if number_deleted == 0:
             sys.exit(1)
diff --git a/takahe/settings.py b/takahe/settings.py
index f9a422d..a2724f1 100644
--- a/takahe/settings.py
+++ b/takahe/settings.py
@@ -143,9 +143,9 @@ class Settings(BaseSettings):
     CACHES_DEFAULT: CacheBackendUrl | None = None
 
     # How long to wait, in days, until remote posts/profiles are pruned from
-    # our database if nobody local has interacted with them. Must be in rough
-    # multiples of two weeks. Set to zero to disable.
-    REMOTE_PRUNE_HORIZON: int = 0
+    # our database if nobody local has interacted with them.
+    # Set to zero to disable.
+    REMOTE_PRUNE_HORIZON: int = 90
 
     # Stator tuning
     STATOR_CONCURRENCY: int = 50
diff --git a/users/management/__init__.py b/users/management/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/users/management/commands/__init__.py b/users/management/commands/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/users/management/commands/pruneidentities.py b/users/management/commands/pruneidentities.py
new file mode 100644
index 0000000..7bf0a55
--- /dev/null
+++ b/users/management/commands/pruneidentities.py
@@ -0,0 +1,52 @@
+import sys
+
+from django.conf import settings
+from django.core.management.base import BaseCommand
+from django.db.models import Q
+from django.utils import timezone
+
+from users.models import Identity
+
+
+class Command(BaseCommand):
+    help = "Prunes identities that have no local interaction"
+
+    def add_arguments(self, parser):
+        parser.add_argument(
+            "--number",
+            "-n",
+            type=int,
+            default=1000,
+            help="The maximum number of identities to prune at once",
+        )
+
+    def handle(self, number: int, *args, **options):
+        if not settings.SETUP.REMOTE_PRUNE_HORIZON:
+            print("Pruning has been disabled as REMOTE_PRUNE_HORIZON=0")
+            sys.exit(2)
+        # Find a set of identities that match the initial criteria
+        print(f"Running query to find up to {number} unused identities...")
+        identities = Identity.objects.filter(
+            local=False,
+            created__lt=timezone.now(),
+        ).exclude(
+            Q(interactions__post__local=True)
+            | Q(posts__isnull=False)
+            | Q(outbound_follows__isnull=False)
+            | Q(inbound_follows__isnull=False)
+            | Q(outbound_blocks__isnull=False)
+            | Q(inbound_blocks__isnull=False)
+        )[
+            :number
+        ]
+        identity_ids = identities.values_list("id", flat=True)
+        print(f"  found {len(identity_ids)}")
+
+        # Delete them
+        print("Deleting...")
+        number_deleted, deleted = Identity.objects.filter(id__in=identity_ids).delete()
+        print("Deleted:")
+        for model, model_deleted in deleted.items():
+            print(f"  {model}: {model_deleted}")
+        if number_deleted == 0:
+            sys.exit(1)