Refactor feed manager (#14761)

8 files changed, 207 insertions, 63 deletions

diff --git a/app/lib/feed_manager.rb b/app/lib/feed_manager.rb
index 785009b52..0876d107b 100644
--- a/app/lib/feed_manager.rb
+++ b/app/lib/feed_manager.rb
@@ -6,31 +6,54 @@ class FeedManager
   include Singleton
   include Redisable
 
+  # Maximum number of items stored in a single feed
   MAX_ITEMS = 400
 
-  # Must be <= MAX_ITEMS or the tracking sets will grow forever
+  # Number of items in the feed since last reblog of status
+  # before the new reblog will be inserted. Must be <= MAX_ITEMS
+  # or the tracking sets will grow forever
   REBLOG_FALLOFF = 40
 
+  # Execute block for every active account
+  # @yield [Account]
+  # @return [void]
   def with_active_accounts(&block)
     Account.joins(:user).where('users.current_sign_in_at > ?', User::ACTIVE_DURATION.ago).find_each(&block)
   end
 
+  # Redis key of a feed
+  # @param [Symbol] type
+  # @param [Integer] id
+  # @param [Symbol] subtype
+  # @return [String]
   def key(type, id, subtype = nil)
     return "feed:#{type}:#{id}" unless subtype
 
     "feed:#{type}:#{id}:#{subtype}"
   end
 
-  def filter?(timeline_type, status, receiver_id)
-    if timeline_type == :home
-      filter_from_home?(status, receiver_id, build_crutches(receiver_id, [status]))
-    elsif timeline_type == :mentions
-      filter_from_mentions?(status, receiver_id)
+  # Check if the status should not be added to a feed
+  # @param [Symbol] timeline_type
+  # @param [Status] status
+  # @param [Account|List] receiver
+  # @return [Boolean]
+  def filter?(timeline_type, status, receiver)
+    case timeline_type
+    when :home
+      filter_from_home?(status, receiver.id, build_crutches(receiver.id, [status]))
+    when :list
+      filter_from_list?(status, receiver) || filter_from_home?(status, receiver.account_id, build_crutches(receiver.account_id, [status]))
+    when :mentions
+      filter_from_mentions?(status, receiver.id)
     else
       false
     end
   end
 
+  # Add a status to a home feed and send a streaming API update
+  # @param [Account] account
+  # @param [Status] status
+  # @return [Boolean]
   def push_to_home(account, status)
     return false unless add_to_feed(:home, account.id, status, account.user&.aggregates_reblogs?)
 
@@ -39,6 +62,10 @@ class FeedManager
     true
   end
 
+  # Remove a status from a home feed and send a streaming API update
+  # @param [Account] account
+  # @param [Status] status
+  # @return [Boolean]
   def unpush_from_home(account, status)
     return false unless remove_from_feed(:home, account.id, status, account.user&.aggregates_reblogs?)
 
@@ -46,21 +73,22 @@ class FeedManager
     true
   end
 
+  # Add a status to a list feed and send a streaming API update
+  # @param [List] list
+  # @param [Status] status
+  # @return [Boolean]
   def push_to_list(list, status)
-    if status.reply? && status.in_reply_to_account_id != status.account_id
-      should_filter = status.in_reply_to_account_id != list.account_id
-      should_filter &&= !list.show_all_replies?
-      should_filter &&= !(list.show_list_replies? && ListAccount.where(list_id: list.id, account_id: status.in_reply_to_account_id).exists?)
-      return false if should_filter
-    end
-
-    return false unless add_to_feed(:list, list.id, status, list.account.user&.aggregates_reblogs?)
+    return false if filter_from_list?(status, list) || !add_to_feed(:list, list.id, status, list.account.user&.aggregates_reblogs?)
 
     trim(:list, list.id)
     PushUpdateWorker.perform_async(list.account_id, status.id, "timeline:list:#{list.id}") if push_update_required?("timeline:list:#{list.id}")
     true
   end
 
+  # Remove a status from a list feed and send a streaming API update
+  # @param [List] list
+  # @param [Status] status
+  # @return [Boolean]
   def unpush_from_list(list, status)
     return false unless remove_from_feed(:list, list.id, status, list.account.user&.aggregates_reblogs?)
 
@@ -68,36 +96,39 @@ class FeedManager
     true
   end
 
-  def trim(type, account_id)
-    timeline_key = key(type, account_id)
-    reblog_key   = key(type, account_id, 'reblogs')
+  # Fill a home feed with an account's statuses
+  # @param [Account] from_account
+  # @param [Account] into_account
+  # @return [void]
+  def merge_into_home(from_account, into_account)
+    timeline_key = key(:home, into_account.id)
+    aggregate    = into_account.user&.aggregates_reblogs?
+    query        = from_account.statuses.where(visibility: [:public, :unlisted, :private]).includes(:preloadable_poll, reblog: :account).limit(FeedManager::MAX_ITEMS / 4)
 
-    # Remove any items past the MAX_ITEMS'th entry in our feed
-    redis.zremrangebyrank(timeline_key, 0, -(FeedManager::MAX_ITEMS + 1))
+    if redis.zcard(timeline_key) >= FeedManager::MAX_ITEMS / 4
+      oldest_home_score = redis.zrange(timeline_key, 0, 0, with_scores: true).first.last.to_i
+      query = query.where('id > ?', oldest_home_score)
+    end
 
-    # Get the score of the REBLOG_FALLOFF'th item in our feed, and stop
-    # tracking anything after it for deduplication purposes.
-    falloff_rank  = FeedManager::REBLOG_FALLOFF
-    falloff_range = redis.zrevrange(timeline_key, falloff_rank, falloff_rank, with_scores: true)
-    falloff_score = falloff_range&.first&.last&.to_i
+    statuses = query.to_a
+    crutches = build_crutches(into_account.id, statuses)
 
-    return if falloff_score.nil?
+    statuses.each do |status|
+      next if filter_from_home?(status, into_account.id, crutches)
 
-    # Get any reblogs we might have to clean up after.
-    redis.zrangebyscore(reblog_key, 0, falloff_score).each do |reblogged_id|
-      # Remove it from the set of reblogs we're tracking *first* to avoid races.
-      redis.zrem(reblog_key, reblogged_id)
-      # Just drop any set we might have created to track additional reblogs.
-      # This means that if this reblog is deleted, we won't automatically insert
-      # another reblog, but also that any new reblog can be inserted into the
-      # feed.
-      redis.del(key(type, account_id, "reblogs:#{reblogged_id}"))
+      add_to_feed(:home, into_account.id, status, aggregate)
     end
+
+    trim(:home, into_account.id)
   end
 
-  def merge_into_timeline(from_account, into_account)
-    timeline_key = key(:home, into_account.id)
-    aggregate    = into_account.user&.aggregates_reblogs?
+  # Fill a list feed with an account's statuses
+  # @param [Account] from_account
+  # @param [List] list
+  # @return [void]
+  def merge_into_list(from_account, list)
+    timeline_key = key(:list, list.id)
+    aggregate    = list.account.user&.aggregates_reblogs?
     query        = from_account.statuses.where(visibility: [:public, :unlisted, :private]).includes(:preloadable_poll, reblog: :account).limit(FeedManager::MAX_ITEMS / 4)
 
     if redis.zcard(timeline_key) >= FeedManager::MAX_ITEMS / 4
@@ -106,18 +137,22 @@ class FeedManager
     end
 
     statuses = query.to_a
-    crutches = build_crutches(into_account.id, statuses)
+    crutches = build_crutches(list.account_id, statuses)
 
     statuses.each do |status|
-      next if filter_from_home?(status, into_account.id, crutches)
+      next if filter_from_home?(status, list.account_id, crutches) || filter_from_list?(status, list)
 
-      add_to_feed(:home, into_account.id, status, aggregate)
+      add_to_feed(:list, list.id, status, aggregate)
     end
 
-    trim(:home, into_account.id)
+    trim(:list, list.id)
   end
 
-  def unmerge_from_timeline(from_account, into_account)
+  # Remove an account's statuses from a home feed
+  # @param [Account] from_account
+  # @param [Account] into_account
+  # @return [void]
+  def unmerge_from_home(from_account, into_account)
     timeline_key      = key(:home, into_account.id)
     oldest_home_score = redis.zrange(timeline_key, 0, 0, with_scores: true)&.first&.last&.to_i || 0
 
@@ -126,14 +161,31 @@ class FeedManager
     end
   end
 
-  def clear_from_timeline(account, target_account)
-    # Clear from timeline all statuses from or mentionning target_account
+  # Remove an account's statuses from a list feed
+  # @param [Account] from_account
+  # @param [List] list
+  # @return [void]
+  def unmerge_from_list(from_account, list)
+    timeline_key      = key(:list, list.id)
+    oldest_list_score = redis.zrange(timeline_key, 0, 0, with_scores: true)&.first&.last&.to_i || 0
+
+    from_account.statuses.select('id, reblog_of_id').where('id > ?', oldest_list_score).reorder(nil).find_each do |status|
+      remove_from_feed(:list, list.id, status, list.account.user&.aggregates_reblogs?)
+    end
+  end
+
+  # Clear all statuses from or mentioning target_account from a home feed
+  # @param [Account] account
+  # @param [Account] target_account
+  # @return [void]
+  def clear_from_home(account, target_account)
     timeline_key        = key(:home, account.id)
     timeline_status_ids = redis.zrange(timeline_key, 0, -1)
     statuses            = Status.where(id: timeline_status_ids).select(:id, :reblog_of_id, :account_id).to_a
     reblogged_ids       = Status.where(id: statuses.map(&:reblog_of_id).compact, account: target_account).pluck(:id)
     with_mentions_ids   = Mention.active.where(status_id: statuses.flat_map { |s| [s.id, s.reblog_of_id] }.compact, account: target_account).pluck(:status_id)
-    target_statuses     = statuses.filter do |status|
+
+    target_statuses = statuses.select do |status|
       status.account_id == target_account.id || reblogged_ids.include?(status.reblog_of_id) || with_mentions_ids.include?(status.id) || with_mentions_ids.include?(status.reblog_of_id)
     end
 
@@ -142,7 +194,10 @@ class FeedManager
     end
   end
 
-  def populate_feed(account)
+  # Populate home feed of account from scratch
+  # @param [Account] account
+  # @return [void]
+  def populate_home(account)
     limit        = FeedManager::MAX_ITEMS / 2
     aggregate    = account.user&.aggregates_reblogs?
     timeline_key = key(:home, account.id)
@@ -177,15 +232,59 @@ class FeedManager
 
   private
 
-  def push_update_required?(timeline_id)
-    redis.exists?("subscribed:#{timeline_id}")
+  # Trim a feed to maximum size by removing older items
+  # @param [Symbol] type
+  # @param [Integer] timeline_id
+  # @return [void]
+  def trim(type, timeline_id)
+    timeline_key = key(type, timeline_id)
+    reblog_key   = key(type, timeline_id, 'reblogs')
+
+    # Remove any items past the MAX_ITEMS'th entry in our feed
+    redis.zremrangebyrank(timeline_key, 0, -(FeedManager::MAX_ITEMS + 1))
+
+    # Get the score of the REBLOG_FALLOFF'th item in our feed, and stop
+    # tracking anything after it for deduplication purposes.
+    falloff_rank  = FeedManager::REBLOG_FALLOFF
+    falloff_range = redis.zrevrange(timeline_key, falloff_rank, falloff_rank, with_scores: true)
+    falloff_score = falloff_range&.first&.last&.to_i
+
+    return if falloff_score.nil?
+
+    # Get any reblogs we might have to clean up after.
+    redis.zrangebyscore(reblog_key, 0, falloff_score).each do |reblogged_id|
+      # Remove it from the set of reblogs we're tracking *first* to avoid races.
+      redis.zrem(reblog_key, reblogged_id)
+      # Just drop any set we might have created to track additional reblogs.
+      # This means that if this reblog is deleted, we won't automatically insert
+      # another reblog, but also that any new reblog can be inserted into the
+      # feed.
+      redis.del(key(type, timeline_id, "reblogs:#{reblogged_id}"))
+    end
   end
 
+  # Check if there is a streaming API client connected
+  # for the given feed
+  # @param [String] timeline_key
+  # @return [Boolean]
+  def push_update_required?(timeline_key)
+    redis.exists?("subscribed:#{timeline_key}")
+  end
+
+  # Check if the account is blocking or muting any of the given accounts
+  # @param [Integer] receiver_id
+  # @param [Array<Integer>] account_ids
+  # @param [Symbol] context
   def blocks_or_mutes?(receiver_id, account_ids, context)
     Block.where(account_id: receiver_id, target_account_id: account_ids).any? ||
       (context == :home ? Mute.where(account_id: receiver_id, target_account_id: account_ids).any? : Mute.where(account_id: receiver_id, target_account_id: account_ids, hide_notifications: true).any?)
   end
 
+  # Check if status should not be added to the home feed
+  # @param [Status] status
+  # @param [Integer] receiver_id
+  # @param [Hash] crutches
+  # @return [Boolean]
   def filter_from_home?(status, receiver_id, crutches)
     return false if receiver_id == status.account_id
     return true  if status.reply? && (status.in_reply_to_id.nil? || status.in_reply_to_account_id.nil?)
@@ -218,6 +317,11 @@ class FeedManager
     false
   end
 
+  # Check if status should not be added to the mentions feed
+  # @see NotifyService
+  # @param [Status] status
+  # @param [Integer] receiver_id
+  # @return [Boolean]
   def filter_from_mentions?(status, receiver_id)
     return true if receiver_id == status.account_id
     return true if phrase_filtered?(status, receiver_id, :notifications)
@@ -234,6 +338,27 @@ class FeedManager
     should_filter
   end
 
+  # Check if status should not be added to the list feed
+  # @param [Status] status
+  # @param [List] list
+  # @return [Boolean]
+  def filter_from_list?(status, list)
+    if status.reply? && status.in_reply_to_account_id != status.account_id
+      should_filter = status.in_reply_to_account_id != list.account_id
+      should_filter &&= !list.show_all_replies?
+      should_filter &&= !(list.show_list_replies? && ListAccount.where(list_id: list.id, account_id: status.in_reply_to_account_id).exists?)
+
+      return !!should_filter
+    end
+
+    false
+  end
+
+  # Check if the status hits a phrase filter
+  # @param [Status] status
+  # @param [Integer] receiver_id
+  # @param [Symbol] context
+  # @return [Boolean]
   def phrase_filtered?(status, receiver_id, context)
     active_filters = Rails.cache.fetch("filters:#{receiver_id}") { CustomFilter.where(account_id: receiver_id).active_irreversible.to_a }.to_a
 
@@ -269,6 +394,11 @@ class FeedManager
   # added, and false if it was not added to the feed. Note that this is
   # an internal helper: callers must call trim or push updates if
   # either action is appropriate.
+  # @param [Symbol] timeline_type
+  # @param [Integer] account_id
+  # @param [Status] status
+  # @param [Boolean] aggregate_reblogs
+  # @return [Boolean]
   def add_to_feed(timeline_type, account_id, status, aggregate_reblogs = true)
     timeline_key = key(timeline_type, account_id)
     reblog_key   = key(timeline_type, account_id, 'reblogs')
@@ -312,6 +442,11 @@ class FeedManager
   # with reblogs, and returning true if a status was removed. As with
   # `add_to_feed`, this does not trigger push updates, so callers must
   # do so if appropriate.
+  # @param [Symbol] timeline_type
+  # @param [Integer] account_id
+  # @param [Status] status
+  # @param [Boolean] aggregate_reblogs
+  # @return [Boolean]
   def remove_from_feed(timeline_type, account_id, status, aggregate_reblogs = true)
     timeline_key = key(timeline_type, account_id)
     reblog_key   = key(timeline_type, account_id, 'reblogs')
@@ -346,6 +481,11 @@ class FeedManager
     redis.zrem(timeline_key, status.id)
   end
 
+  # Pre-fetch various objects and relationships for given statuses that
+  # are going to be checked by the filtering methods
+  # @param [Integer] receiver_id
+  # @param [Array<Status>] statuses
+  # @return [Hash]
   def build_crutches(receiver_id, statuses)
     crutches = {}
 
diff --git a/app/services/after_block_service.rb b/app/services/after_block_service.rb
index 2a0e10a79..314919df8 100644
--- a/app/services/after_block_service.rb
+++ b/app/services/after_block_service.rb
@@ -13,7 +13,7 @@ class AfterBlockService < BaseService
   private
 
   def clear_home_feed!
-    FeedManager.instance.clear_from_timeline(@account, @target_account)
+    FeedManager.instance.clear_from_home(@account, @target_account)
   end
 
   def clear_conversations!
diff --git a/app/services/notify_service.rb b/app/services/notify_service.rb
index abd676494..e4ca10eb1 100644
--- a/app/services/notify_service.rb
+++ b/app/services/notify_service.rb
@@ -13,15 +13,13 @@ class NotifyService < BaseService
     push_to_conversation! if direct_message?
     send_email! if email_enabled?
   rescue ActiveRecord::RecordInvalid
-    # rubocop:disable Style/RedundantReturn
-    return
-    # rubocop:enable Style/RedundantReturn
+    nil
   end
 
   private
 
   def blocked_mention?
-    FeedManager.instance.filter?(:mentions, @notification.mention.status, @recipient.id)
+    FeedManager.instance.filter?(:mentions, @notification.mention.status, @recipient)
   end
 
   def blocked_favourite?
diff --git a/app/services/precompute_feed_service.rb b/app/services/precompute_feed_service.rb
index 076dedaca..61f573534 100644
--- a/app/services/precompute_feed_service.rb
+++ b/app/services/precompute_feed_service.rb
@@ -2,7 +2,7 @@
 
 class PrecomputeFeedService < BaseService
   def call(account)
-    FeedManager.instance.populate_feed(account)
+    FeedManager.instance.populate_home(account)
   ensure
     Redis.current.del("account:#{account.id}:regeneration")
   end
diff --git a/app/workers/feed_insert_worker.rb b/app/workers/feed_insert_worker.rb
index 1ae3c877b..633ec91bd 100644
--- a/app/workers/feed_insert_worker.rb
+++ b/app/workers/feed_insert_worker.rb
@@ -27,9 +27,12 @@ class FeedInsertWorker
   end
 
   def feed_filtered?
-    # Note: Lists are a variation of home, so the filtering rules
-    # of home apply to both
-    FeedManager.instance.filter?(:home, @status, @follower.id)
+    case @type
+    when :home
+      FeedManager.instance.filter?(:home, @status, @follower)
+    when :list
+      FeedManager.instance.filter?(:list, @status, @list)
+    end
   end
 
   def perform_push
diff --git a/app/workers/merge_worker.rb b/app/workers/merge_worker.rb
index d745cb99c..74ef7d4da 100644
--- a/app/workers/merge_worker.rb
+++ b/app/workers/merge_worker.rb
@@ -6,6 +6,8 @@ class MergeWorker
   sidekiq_options queue: 'pull'
 
   def perform(from_account_id, into_account_id)
-    FeedManager.instance.merge_into_timeline(Account.find(from_account_id), Account.find(into_account_id))
+    FeedManager.instance.merge_into_home(Account.find(from_account_id), Account.find(into_account_id))
+  rescue ActiveRecord::RecordNotFound
+    true
   end
 end
diff --git a/app/workers/mute_worker.rb b/app/workers/mute_worker.rb
index 7bf0923a5..c74f657cb 100644
--- a/app/workers/mute_worker.rb
+++ b/app/workers/mute_worker.rb
@@ -4,9 +4,8 @@ class MuteWorker
   include Sidekiq::Worker
 
   def perform(account_id, target_account_id)
-    FeedManager.instance.clear_from_timeline(
-      Account.find(account_id),
-      Account.find(target_account_id)
-    )
+    FeedManager.instance.clear_from_home(Account.find(account_id), Account.find(target_account_id))
+  rescue ActiveRecord::RecordNotFound
+    true
   end
 end
diff --git a/app/workers/unmerge_worker.rb b/app/workers/unmerge_worker.rb
index ea6aacebf..1a23faae5 100644
--- a/app/workers/unmerge_worker.rb
+++ b/app/workers/unmerge_worker.rb
@@ -6,6 +6,8 @@ class UnmergeWorker
   sidekiq_options queue: 'pull'
 
   def perform(from_account_id, into_account_id)
-    FeedManager.instance.unmerge_from_timeline(Account.find(from_account_id), Account.find(into_account_id))
+    FeedManager.instance.unmerge_from_home(Account.find(from_account_id), Account.find(into_account_id))
+  rescue ActiveRecord::RecordNotFound
+    true
   end
 end