Add flag to preserve cached media on cleanup (#36200)

Co-authored-by: Daniel King <git@kin.gy>

app/models/media_attachment.rb

@@ -214,6 +214,14 @@ class MediaAttachment < ApplicationRecord
   scope :remote, -> { where.not(remote_url: '') }
   scope :unattached, -> { where(status_id: nil, scheduled_status_id: nil) }
   scope :updated_before, ->(value) { where(arel_table[:updated_at].lt(value)) }
+  scope :without_local_interaction, lambda {
+    where.not(Favourite.joins(:account).merge(Account.local).where(Favourite.arel_table[:status_id].eq(MediaAttachment.arel_table[:status_id])).select(1).arel.exists)
+         .where.not(Bookmark.where(Bookmark.arel_table[:status_id].eq(MediaAttachment.arel_table[:status_id])).select(1).arel.exists)
+         .where.not(Status.local.where(Status.arel_table[:in_reply_to_id].eq(MediaAttachment.arel_table[:status_id])).select(1).arel.exists)
+         .where.not(Status.local.where(Status.arel_table[:reblog_of_id].eq(MediaAttachment.arel_table[:status_id])).select(1).arel.exists)
+         .where.not(Quote.joins(:status).merge(Status.local).where(Quote.arel_table[:quoted_status_id].eq(MediaAttachment.arel_table[:status_id])).select(1).arel.exists)
+         .where.not(Quote.joins(:quoted_status).merge(Status.local).where(Quote.arel_table[:status_id].eq(MediaAttachment.arel_table[:status_id])).select(1).arel.exists)
+  }
 
   attr_accessor :skip_download
 

app/models/status.rb

@@ -90,6 +90,7 @@ class Status < ApplicationRecord
   has_many :local_favorited, -> { merge(Account.local) }, through: :favourites, source: :account
   has_many :local_reblogged, -> { merge(Account.local) }, through: :reblogs, source: :account
   has_many :local_bookmarked, -> { merge(Account.local) }, through: :bookmarks, source: :account
+  has_many :local_replied, -> { merge(Account.local) }, through: :replies, source: :account
 
   has_and_belongs_to_many :tags # rubocop:disable Rails/HasAndBelongsToMany
 

lib/mastodon/cli/media.rb

@@ -17,6 +17,7 @@ module Mastodon::CLI
     option :concurrency, type: :numeric, default: 5, aliases: [:c]
     option :verbose, type: :boolean, default: false, aliases: [:v]
     option :dry_run, type: :boolean, default: false
+    option :keep_interacted, type: :boolean, default: false
     desc 'remove', 'Remove remote media files, headers or avatars'
     long_desc <<-DESC
       Removes locally cached copies of media attachments (and optionally profile
@@ -26,6 +27,9 @@ module Mastodon::CLI
       they are removed. In case of avatars and headers, it specifies how old
       the last webfinger request and update to the user has to be before they
       are pruned. It defaults to 7 days.
+      If --keep-interacted is specified, any media attached to a status that
+      was favourited, bookmarked, quoted, replied to, or reblogged by a local
+      account will be preserved.
       If --prune-profiles is specified, only avatars and headers are removed.
       If --remove-headers is specified, only headers are removed.
       If --include-follows is specified along with --prune-profiles or
@@ -61,7 +65,11 @@ module Mastodon::CLI
       end
 
       unless options[:prune_profiles] || options[:remove_headers]
-        processed, aggregate = parallelize_with_progress(MediaAttachment.cached.remote.where(created_at: ..time_ago)) do |media_attachment|
+        attachment_scope = MediaAttachment.cached.remote.where(created_at: ..time_ago)
+
+        attachment_scope = attachment_scope.without_local_interaction if options[:keep_interacted]
+
+        processed, aggregate = parallelize_with_progress(attachment_scope) do |media_attachment|
           next if media_attachment.file.blank?
 
           size = (media_attachment.file_file_size || 0) + (media_attachment.thumbnail_file_size || 0)

spec/lib/mastodon/cli/media_spec.rb

@@ -73,6 +73,66 @@ RSpec.describe Mastodon::CLI::Media do
           expect(media_attachment.reload.thumbnail).to be_blank
         end
       end
+
+      context 'with --keep-interacted' do
+        let(:options) { { keep_interacted: true } }
+
+        let!(:favourited_media) { Fabricate(:media_attachment, created_at: 1.month.ago, remote_url: 'https://example.com/image.jpg') }
+        let!(:bookmarked_media) { Fabricate(:media_attachment, created_at: 1.month.ago, remote_url: 'https://example.com/image.jpg') }
+        let!(:replied_to_media) { Fabricate(:media_attachment, created_at: 1.month.ago, remote_url: 'https://example.com/image.jpg') }
+        let!(:reblogged_media) { Fabricate(:media_attachment, created_at: 1.month.ago, remote_url: 'https://example.com/image.jpg') }
+        let!(:remote_quoted_media) { Fabricate(:media_attachment, created_at: 1.month.ago, remote_url: 'https://example.com/image.jpg') }
+        let!(:remote_quoting_media) { Fabricate(:media_attachment, created_at: 1.month.ago, remote_url: 'https://example.com/image.jpg') }
+
+        before do
+          local_account = Fabricate(:account, username: 'alice')
+          remote_account = Fabricate(:account, username: 'bob', domain: 'example.com')
+
+          favourited_status = Fabricate(:status, account: remote_account)
+          bookmarked_status = Fabricate(:status, account: remote_account)
+          replied_to_status = Fabricate(:status, account: remote_account)
+          reblogged_status = Fabricate(:status, account: remote_account)
+
+          favourited_media.update!(status: favourited_status)
+          bookmarked_media.update!(status: bookmarked_status)
+          replied_to_media.update!(status: replied_to_status)
+          reblogged_media.update!(status: reblogged_status)
+
+          local_quoting_status = Fabricate(:status, account: local_account)
+          remote_quoted_status = Fabricate(:status, account: remote_account)
+          local_status_being_quoted = Fabricate(:status, account: local_account)
+          remote_quoting_status = Fabricate(:status, account: remote_account)
+
+          remote_quoted_media.update!(status: remote_quoted_status)
+          remote_quoting_media.update!(status: remote_quoting_status)
+
+          non_interacted_status = Fabricate(:status, account: remote_account)
+
+          media_attachment.update(status: non_interacted_status)
+
+          Fabricate(:favourite, account: local_account, status: favourited_status)
+          Fabricate(:bookmark, account: local_account, status: bookmarked_status)
+          Fabricate(:status, account: local_account, in_reply_to_id: replied_to_status.id)
+          Fabricate(:status, account: local_account, reblog: reblogged_status)
+          Fabricate(:quote, account: local_account, status: local_quoting_status, quoted_status: remote_quoted_status)
+          Fabricate(:quote, account: remote_account, status: remote_quoting_status, quoted_status: local_status_being_quoted)
+        end
+
+        it 'keeps media associated with statuses that have been favourited, bookmarked, replied to, or reblogged by a local account' do
+          expect { subject }
+            .to output_results('Removed 1')
+
+          expect(favourited_media.reload.file).to be_present
+          expect(bookmarked_media.reload.file).to be_present
+          expect(replied_to_media.reload.file).to be_present
+          expect(reblogged_media.reload.file).to be_present
+          expect(remote_quoted_media.reload.file).to be_present
+          expect(remote_quoting_media.reload.file).to be_present
+
+          expect(media_attachment.reload.file).to be_blank
+          expect(media_attachment.reload.thumbnail).to be_blank
+        end
+      end
     end
   end