From 45ebdb72ca1678eb30e6087f61bd019c7ea903f4 Mon Sep 17 00:00:00 2001 From: Eugen Rochko Date: Sat, 8 Oct 2022 16:45:40 +0200 Subject: Add support for language preferences for trending statuses and links (#18288) --- app/models/trends/base.rb | 4 ++ app/models/trends/links.rb | 98 ++++++++++++++++++++++---------- app/models/trends/preview_card_filter.rb | 25 +++++--- app/models/trends/status_filter.rb | 25 +++++--- app/models/trends/statuses.rb | 82 ++++++++++++++------------ 5 files changed, 151 insertions(+), 83 deletions(-) (limited to 'app/models/trends') diff --git a/app/models/trends/base.rb b/app/models/trends/base.rb index 047111248..a189f11f2 100644 --- a/app/models/trends/base.rb +++ b/app/models/trends/base.rb @@ -98,4 +98,8 @@ class Trends::Base pipeline.rename(from_key, to_key) end end + + def skip_review? + Setting.trendable_by_default + end end diff --git a/app/models/trends/links.rb b/app/models/trends/links.rb index 604894cd6..8808b3ab6 100644 --- a/app/models/trends/links.rb +++ b/app/models/trends/links.rb @@ -11,6 +11,40 @@ class Trends::Links < Trends::Base decay_threshold: 1, } + class Query < Trends::Query + def filtered_for!(account) + @account = account + self + end + + def filtered_for(account) + clone.filtered_for!(account) + end + + def to_arel + scope = PreviewCard.joins(:trend).reorder(score: :desc) + scope = scope.reorder(language_order_clause.desc, score: :desc) if preferred_languages.present? + scope = scope.merge(PreviewCardTrend.allowed) if @allowed + scope = scope.offset(@offset) if @offset.present? + scope = scope.limit(@limit) if @limit.present? + scope + end + + private + + def language_order_clause + Arel::Nodes::Case.new.when(PreviewCardTrend.arel_table[:language].in(preferred_languages)).then(1).else(0) + end + + def preferred_languages + if @account&.chosen_languages.present? + @account.chosen_languages + else + @locale + end + end + end + def register(status, at_time = Time.now.utc) original_status = status.proper @@ -28,24 +62,33 @@ class Trends::Links < Trends::Base record_used_id(preview_card.id, at_time) end + def query + Query.new(key_prefix, klass) + end + def refresh(at_time = Time.now.utc) - preview_cards = PreviewCard.where(id: (recently_used_ids(at_time) + currently_trending_ids(false, -1)).uniq) + preview_cards = PreviewCard.where(id: (recently_used_ids(at_time) + PreviewCardTrend.pluck(:preview_card_id)).uniq) calculate_scores(preview_cards, at_time) end def request_review - preview_cards = PreviewCard.where(id: currently_trending_ids(false, -1)) + PreviewCardTrend.pluck('distinct language').flat_map do |language| + score_at_threshold = PreviewCardTrend.where(language: language, allowed: true).order(rank: :desc).where('rank <= ?', options[:review_threshold]).first&.score || 0 + preview_card_trends = PreviewCardTrend.where(language: language, allowed: false).joins(:preview_card) - preview_cards.filter_map do |preview_card| - next unless would_be_trending?(preview_card.id) && !preview_card.trendable? && preview_card.requires_review_notification? + preview_card_trends.filter_map do |trend| + preview_card = trend.preview_card - if preview_card.provider.nil? - preview_card.provider = PreviewCardProvider.create(domain: preview_card.domain, requested_review_at: Time.now.utc) - else - preview_card.provider.touch(:requested_review_at) - end + next unless trend.score > score_at_threshold && !preview_card.trendable? && preview_card.requires_review_notification? + + if preview_card.provider.nil? + preview_card.provider = PreviewCardProvider.create(domain: preview_card.domain, requested_review_at: Time.now.utc) + else + preview_card.provider.touch(:requested_review_at) + end - preview_card + preview_card + end end end @@ -62,10 +105,7 @@ class Trends::Links < Trends::Base private def calculate_scores(preview_cards, at_time) - global_items = [] - locale_items = Hash.new { |h, key| h[key] = [] } - - preview_cards.each do |preview_card| + items = preview_cards.map do |preview_card| expected = preview_card.history.get(at_time - 1.day).accounts.to_f expected = 1.0 if expected.zero? observed = preview_card.history.get(at_time).accounts.to_f @@ -89,26 +129,24 @@ class Trends::Links < Trends::Base preview_card.update_columns(max_score: max_score, max_score_at: max_time) end - decaying_score = max_score * (0.5**((at_time.to_f - max_time.to_f) / options[:max_score_halflife].to_f)) - - next unless decaying_score >= options[:decay_threshold] + decaying_score = begin + if max_score.zero? || !valid_locale?(preview_card.language) + 0 + else + max_score * (0.5**((at_time.to_f - max_time.to_f) / options[:max_score_halflife].to_f)) + end + end - global_items << { score: decaying_score, item: preview_card } - locale_items[preview_card.language] << { score: decaying_score, item: preview_card } if valid_locale?(preview_card.language) + [decaying_score, preview_card] end - replace_items('', global_items) + to_insert = items.filter { |(score, _)| score >= options[:decay_threshold] } + to_delete = items.filter { |(score, _)| score < options[:decay_threshold] } - Trends.available_locales.each do |locale| - replace_items(":#{locale}", locale_items[locale]) + PreviewCardTrend.transaction do + PreviewCardTrend.upsert_all(to_insert.map { |(score, preview_card)| { preview_card_id: preview_card.id, score: score, language: preview_card.language, allowed: preview_card.trendable? || false } }, unique_by: :preview_card_id) if to_insert.any? + PreviewCardTrend.where(preview_card_id: to_delete.map { |(_, preview_card)| preview_card.id }).delete_all if to_delete.any? + PreviewCardTrend.connection.exec_update('UPDATE preview_card_trends SET rank = t0.calculated_rank FROM (SELECT id, row_number() OVER w AS calculated_rank FROM preview_card_trends WINDOW w AS (PARTITION BY language ORDER BY score DESC)) t0 WHERE preview_card_trends.id = t0.id') end end - - def filter_for_allowed_items(items) - items.select { |item| item[:item].trendable? } - end - - def would_be_trending?(id) - score(id) > score_at_rank(options[:review_threshold] - 1) - end end diff --git a/app/models/trends/preview_card_filter.rb b/app/models/trends/preview_card_filter.rb index 25add58c8..0a81146d4 100644 --- a/app/models/trends/preview_card_filter.rb +++ b/app/models/trends/preview_card_filter.rb @@ -13,10 +13,10 @@ class Trends::PreviewCardFilter end def results - scope = PreviewCard.unscoped + scope = initial_scope params.each do |key, value| - next if %w(page locale).include?(key.to_s) + next if %w(page).include?(key.to_s) scope.merge!(scope_for(key, value.to_s.strip)) if value.present? end @@ -26,21 +26,30 @@ class Trends::PreviewCardFilter private + def initial_scope + PreviewCard.select(PreviewCard.arel_table[Arel.star]) + .joins(:trend) + .eager_load(:trend) + .reorder(score: :desc) + end + def scope_for(key, value) case key.to_s when 'trending' trending_scope(value) + when 'locale' + PreviewCardTrend.where(language: value) else raise "Unknown filter: #{key}" end end def trending_scope(value) - scope = Trends.links.query - - scope = scope.in_locale(@params[:locale].to_s) if @params[:locale].present? - scope = scope.allowed if value == 'allowed' - - scope.to_arel + case value + when 'allowed' + PreviewCardTrend.allowed + else + PreviewCardTrend.all + end end end diff --git a/app/models/trends/status_filter.rb b/app/models/trends/status_filter.rb index 7c453e339..cb0f75d67 100644 --- a/app/models/trends/status_filter.rb +++ b/app/models/trends/status_filter.rb @@ -13,10 +13,10 @@ class Trends::StatusFilter end def results - scope = Status.unscoped.kept + scope = initial_scope params.each do |key, value| - next if %w(page locale).include?(key.to_s) + next if %w(page).include?(key.to_s) scope.merge!(scope_for(key, value.to_s.strip)) if value.present? end @@ -26,21 +26,30 @@ class Trends::StatusFilter private + def initial_scope + Status.select(Status.arel_table[Arel.star]) + .joins(:trend) + .eager_load(:trend) + .reorder(score: :desc) + end + def scope_for(key, value) case key.to_s when 'trending' trending_scope(value) + when 'locale' + StatusTrend.where(language: value) else raise "Unknown filter: #{key}" end end def trending_scope(value) - scope = Trends.statuses.query - - scope = scope.in_locale(@params[:locale].to_s) if @params[:locale].present? - scope = scope.allowed if value == 'allowed' - - scope.to_arel + case value + when 'allowed' + StatusTrend.allowed + else + StatusTrend.all + end end end diff --git a/app/models/trends/statuses.rb b/app/models/trends/statuses.rb index 777065d3e..c9ef7c8f2 100644 --- a/app/models/trends/statuses.rb +++ b/app/models/trends/statuses.rb @@ -20,13 +20,27 @@ class Trends::Statuses < Trends::Base clone.filtered_for!(account) end + def to_arel + scope = Status.joins(:trend).reorder(score: :desc) + scope = scope.reorder(language_order_clause.desc, score: :desc) if preferred_languages.present? + scope = scope.merge(StatusTrend.allowed) if @allowed + scope = scope.not_excluded_by_account(@account).not_domain_blocked_by_account(@account) if @account.present? + scope = scope.offset(@offset) if @offset.present? + scope = scope.limit(@limit) if @limit.present? + scope + end + private - def apply_scopes(scope) - if @account.nil? - scope + def language_order_clause + Arel::Nodes::Case.new.when(StatusTrend.arel_table[:language].in(preferred_languages)).then(1).else(0) + end + + def preferred_languages + if @account&.chosen_languages.present? + @account.chosen_languages else - scope.not_excluded_by_account(@account).not_domain_blocked_by_account(@account) + @locale end end end @@ -36,9 +50,6 @@ class Trends::Statuses < Trends::Base end def add(status, _account_id, at_time = Time.now.utc) - # We rely on the total reblogs and favourites count, so we - # don't record which account did the what and when here - record_used_id(status.id, at_time) end @@ -47,18 +58,23 @@ class Trends::Statuses < Trends::Base end def refresh(at_time = Time.now.utc) - statuses = Status.where(id: (recently_used_ids(at_time) + currently_trending_ids(false, -1)).uniq).includes(:account, :media_attachments) + statuses = Status.where(id: (recently_used_ids(at_time) + StatusTrend.pluck(:status_id)).uniq).includes(:status_stat, :account) calculate_scores(statuses, at_time) end def request_review - statuses = Status.where(id: currently_trending_ids(false, -1)).includes(:account) + StatusTrend.pluck('distinct language').flat_map do |language| + score_at_threshold = StatusTrend.where(language: language, allowed: true).order(rank: :desc).where('rank <= ?', options[:review_threshold]).first&.score || 0 + status_trends = StatusTrend.where(language: language, allowed: false).joins(:status).includes(status: :account) - statuses.filter_map do |status| - next unless would_be_trending?(status.id) && !status.trendable? && status.requires_review_notification? + status_trends.filter_map do |trend| + status = trend.status - status.account.touch(:requested_review_at) - status + if trend.score > score_at_threshold && !status.trendable? && status.requires_review_notification? + status.account.touch(:requested_review_at) + status + end + end end end @@ -75,14 +91,11 @@ class Trends::Statuses < Trends::Base private def eligible?(status) - status.public_visibility? && status.account.discoverable? && !status.account.silenced? && status.spoiler_text.blank? && !status.sensitive? && !status.reply? + status.public_visibility? && status.account.discoverable? && !status.account.silenced? && status.spoiler_text.blank? && !status.sensitive? && !status.reply? && valid_locale?(status.language) end def calculate_scores(statuses, at_time) - global_items = [] - locale_items = Hash.new { |h, key| h[key] = [] } - - statuses.each do |status| + items = statuses.map do |status| expected = 1.0 observed = (status.reblogs_count + status.favourites_count).to_f @@ -94,29 +107,24 @@ class Trends::Statuses < Trends::Base end end - decaying_score = score * (0.5**((at_time.to_f - status.created_at.to_f) / options[:score_halflife].to_f)) - - next unless decaying_score >= options[:decay_threshold] + decaying_score = begin + if score.zero? || !eligible?(status) + 0 + else + score * (0.5**((at_time.to_f - status.created_at.to_f) / options[:score_halflife].to_f)) + end + end - global_items << { score: decaying_score, item: status } - locale_items[status.language] << { account_id: status.account_id, score: decaying_score, item: status } if valid_locale?(status.language) + [decaying_score, status] end - replace_items('', global_items) + to_insert = items.filter { |(score, _)| score >= options[:decay_threshold] } + to_delete = items.filter { |(score, _)| score < options[:decay_threshold] } - Trends.available_locales.each do |locale| - replace_items(":#{locale}", locale_items[locale]) + StatusTrend.transaction do + StatusTrend.upsert_all(to_insert.map { |(score, status)| { status_id: status.id, account_id: status.account_id, score: score, language: status.language, allowed: status.trendable? || false } }, unique_by: :status_id) if to_insert.any? + StatusTrend.where(status_id: to_delete.map { |(_, status)| status.id }).delete_all if to_delete.any? + StatusTrend.connection.exec_update('UPDATE status_trends SET rank = t0.calculated_rank FROM (SELECT id, row_number() OVER w AS calculated_rank FROM status_trends WINDOW w AS (PARTITION BY language ORDER BY score DESC)) t0 WHERE status_trends.id = t0.id') end end - - def filter_for_allowed_items(items) - # Show only one status per account, pick the one with the highest score - # that's also eligible to trend - - items.group_by { |item| item[:account_id] }.values.filter_map { |account_items| account_items.select { |item| item[:item].trendable? && item[:item].account.discoverable? }.max_by { |item| item[:score] } } - end - - def would_be_trending?(id) - score(id) > score_at_rank(options[:review_threshold] - 1) - end end -- cgit