Merge branch 'feature/object-pruning' into 'develop'

Object pruning

See merge request pleroma/pleroma!1181

1 files changed, 47 insertions, 0 deletions

diff --git a/lib/mix/tasks/pleroma/database.ex b/lib/mix/tasks/pleroma/database.ex
index f650b447d..f9bafb277 100644
--- a/lib/mix/tasks/pleroma/database.ex
+++ b/lib/mix/tasks/pleroma/database.ex
@@ -5,6 +5,7 @@
 defmodule Mix.Tasks.Pleroma.Database do
   alias Mix.Tasks.Pleroma.Common
   alias Pleroma.Conversation
+  alias Pleroma.Object
   alias Pleroma.Repo
   alias Pleroma.User
   require Logger
@@ -23,6 +24,10 @@ defmodule Mix.Tasks.Pleroma.Database do
     Options:
     - `--vacuum` - run `VACUUM FULL` after the embedded objects are replaced with their references
 
+  ## Prune old objects from the database
+
+      mix pleroma.database prune_objects
+
   ## Create a conversation for all existing DMs. Can be safely re-run.
 
       mix pleroma.database bump_all_conversations
@@ -72,4 +77,46 @@ defmodule Mix.Tasks.Pleroma.Database do
     Enum.each(users, &User.remove_duplicated_following/1)
     Enum.each(users, &User.update_follower_count/1)
   end
+
+  def run(["prune_objects" | args]) do
+    import Ecto.Query
+
+    {options, [], []} =
+      OptionParser.parse(
+        args,
+        strict: [
+          vacuum: :boolean
+        ]
+      )
+
+    Common.start_pleroma()
+
+    deadline = Pleroma.Config.get([:instance, :remote_post_retention_days])
+
+    Logger.info("Pruning objects older than #{deadline} days")
+
+    time_deadline =
+      NaiveDateTime.utc_now()
+      |> NaiveDateTime.add(-(deadline * 86_400))
+
+    public = "https://www.w3.org/ns/activitystreams#Public"
+
+    from(o in Object,
+      where: fragment("?->'to' \\? ? OR ?->'cc' \\? ?", o.data, ^public, o.data, ^public),
+      where: o.inserted_at < ^time_deadline,
+      where:
+        fragment("split_part(?->>'actor', '/', 3) != ?", o.data, ^Pleroma.Web.Endpoint.host())
+    )
+    |> Repo.delete_all()
+
+    if Keyword.get(options, :vacuum) do
+      Logger.info("Runnning VACUUM FULL")
+
+      Repo.query!(
+        "vacuum full;",
+        [],
+        timeout: :infinity
+      )
+    end
+  end
 end