Adding a scaled sort, to boost smaller communities.

crates/apub/src/activities/create_or_update/post.rs

@@ -150,7 +150,7 @@ impl ActivityHandler for CreateOrUpdatePage {
     PostLike::like(&mut context.pool(), &like_form).await?;
 
     // Calculate initial hot_rank for post
-    PostAggregates::update_hot_rank(&mut context.pool(), post.id).await?;
+    PostAggregates::update_ranks(&mut context.pool(), post.id).await?;
 
     Ok(())
   }

crates/db_schema/src/aggregates/post_aggregates.rs

@@ -1,10 +1,14 @@
 use crate::{
   aggregates::structs::PostAggregates,
   newtypes::PostId,
-  schema::post_aggregates,
-  utils::{functions::hot_rank, get_conn, DbPool},
+  schema::{community_aggregates, post, post_aggregates},
+  utils::{
+    functions::{hot_rank, scaled_rank},
+    get_conn,
+    DbPool,
+  },
 };
-use diesel::{result::Error, ExpressionMethods, QueryDsl};
+use diesel::{result::Error, ExpressionMethods, JoinOnDsl, QueryDsl};
 use diesel_async::RunQueryDsl;
 
 impl PostAggregates {
@@ -16,9 +20,19 @@ impl PostAggregates {
       .await
   }
 
-  pub async fn update_hot_rank(pool: &mut DbPool<'_>, post_id: PostId) -> Result<Self, Error> {
+  pub async fn update_ranks(pool: &mut DbPool<'_>, post_id: PostId) -> Result<Self, Error> {
     let conn = &mut get_conn(pool).await?;
 
+    // Diesel can't update based on a join, which is necessary for the scaled_rank
+    // https://github.com/diesel-rs/diesel/issues/1478
+    // Just select the users_active_month manually for now, since its a single post anyway
+    let users_active_month = community_aggregates::table
+      .select(community_aggregates::users_active_month)
+      .inner_join(post::table.on(community_aggregates::community_id.eq(post::community_id)))
+      .filter(post::id.eq(post_id))
+      .first::<i64>(conn)
+      .await?;
+
     diesel::update(post_aggregates::table)
       .filter(post_aggregates::post_id.eq(post_id))
       .set((
@@ -27,6 +41,11 @@ impl PostAggregates {
           post_aggregates::score,
           post_aggregates::newest_comment_time_necro,
         )),
+        post_aggregates::scaled_rank.eq(scaled_rank(
+          post_aggregates::score,
+          post_aggregates::published,
+          users_active_month,
+        )),
       ))
       .get_result::<Self>(conn)
       .await

crates/db_schema/src/aggregates/structs.rs

@@ -27,11 +27,11 @@ pub struct CommentAggregates {
   pub published: DateTime<Utc>,
   /// The total number of children in this comment branch.
   pub child_count: i32,
-  pub hot_rank: i32,
+  pub hot_rank: f64,
   pub controversy_rank: f64,
 }
 
-#[derive(PartialEq, Eq, Debug, Serialize, Deserialize, Clone)]
+#[derive(PartialEq, Debug, Serialize, Deserialize, Clone)]
 #[cfg_attr(feature = "full", derive(Queryable, Associations, Identifiable, TS))]
 #[cfg_attr(feature = "full", diesel(table_name = community_aggregates))]
 #[cfg_attr(
@@ -55,7 +55,7 @@ pub struct CommunityAggregates {
   pub users_active_month: i64,
   /// The number of users with any activity in the last year.
   pub users_active_half_year: i64,
-  pub hot_rank: i32,
+  pub hot_rank: f64,
 }
 
 #[derive(PartialEq, Eq, Debug, Serialize, Deserialize, Clone, Default)]
@@ -95,11 +95,13 @@ pub struct PostAggregates {
   pub featured_community: bool,
   /// If the post is featured on the site / to local.
   pub featured_local: bool,
-  pub hot_rank: i32,
-  pub hot_rank_active: i32,
+  pub hot_rank: f64,
+  pub hot_rank_active: f64,
   pub community_id: CommunityId,
   pub creator_id: PersonId,
   pub controversy_rank: f64,
+  /// A rank that amplifies smaller communities
+  pub scaled_rank: f64,
 }
 
 #[derive(PartialEq, Eq, Debug, Serialize, Deserialize, Clone)]

crates/db_schema/src/lib.rs

@@ -54,6 +54,7 @@ use ts_rs::TS;
 )]
 #[cfg_attr(feature = "full", DbValueStyle = "verbatim")]
 #[cfg_attr(feature = "full", ts(export))]
+// TODO add the controversial and scaled rankings to the doc below
 /// The post sort types. See here for descriptions: https://join-lemmy.org/docs/en/users/03-votes-and-ranking.html
 pub enum SortType {
   #[default]
@@ -75,6 +76,7 @@ pub enum SortType {
   TopSixMonths,
   TopNineMonths,
   Controversial,
+  Scaled,
 }
 
 #[derive(EnumString, Display, Debug, Serialize, Deserialize, Clone, Copy)]

crates/db_schema/src/schema.rs

@@ -100,7 +100,7 @@ diesel::table! {
         downvotes -> Int8,
         published -> Timestamptz,
         child_count -> Int4,
-        hot_rank -> Int4,
+        hot_rank -> Float8,
         controversy_rank -> Float8,
     }
 }
@@ -198,7 +198,7 @@ diesel::table! {
         users_active_week -> Int8,
         users_active_month -> Int8,
         users_active_half_year -> Int8,
-        hot_rank -> Int4,
+        hot_rank -> Float8,
     }
 }
 
@@ -299,6 +299,16 @@ diesel::table! {
     }
 }
 
+diesel::table! {
+    image_upload (id) {
+        id -> Int4,
+        local_user_id -> Int4,
+        pictrs_alias -> Text,
+        pictrs_delete_token -> Text,
+        published -> Timestamptz,
+    }
+}
+
 diesel::table! {
     instance (id) {
         id -> Int4,
@@ -683,11 +693,12 @@ diesel::table! {
         newest_comment_time -> Timestamptz,
         featured_community -> Bool,
         featured_local -> Bool,
-        hot_rank -> Int4,
-        hot_rank_active -> Int4,
+        hot_rank -> Float8,
+        hot_rank_active -> Float8,
         community_id -> Int4,
         creator_id -> Int4,
         controversy_rank -> Float8,
+        scaled_rank -> Float8,
     }
 }
 
@@ -893,6 +904,7 @@ diesel::joinable!(custom_emoji_keyword -> custom_emoji (custom_emoji_id));
 diesel::joinable!(email_verification -> local_user (local_user_id));
 diesel::joinable!(federation_allowlist -> instance (instance_id));
 diesel::joinable!(federation_blocklist -> instance (instance_id));
+diesel::joinable!(image_upload -> local_user (local_user_id));
 diesel::joinable!(local_site -> site (site_id));
 diesel::joinable!(local_site_rate_limit -> local_site (local_site_id));
 diesel::joinable!(local_user -> person (person_id));
@@ -967,6 +979,7 @@ diesel::allow_tables_to_appear_in_same_query!(
     email_verification,
     federation_allowlist,
     federation_blocklist,
+    image_upload,
     instance,
     language,
     local_site,

crates/db_schema/src/utils.rs

@@ -347,7 +347,7 @@ pub fn naive_now() -> DateTime<Utc> {
 
 pub fn post_to_comment_sort_type(sort: SortType) -> CommentSortType {
   match sort {
-    SortType::Active | SortType::Hot => CommentSortType::Hot,
+    SortType::Active | SortType::Hot | SortType::Scaled => CommentSortType::Hot,
     SortType::New | SortType::NewComments | SortType::MostComments => CommentSortType::New,
     SortType::Old => CommentSortType::Old,
     SortType::Controversial => CommentSortType::Controversial,
@@ -384,7 +384,11 @@ pub mod functions {
   use diesel::sql_types::{BigInt, Text, Timestamptz};
 
   sql_function! {
-    fn hot_rank(score: BigInt, time: Timestamptz) -> Integer;
+    fn hot_rank(score: BigInt, time: Timestamptz) -> Double;
+  }
+
+  sql_function! {
+    fn scaled_rank(score: BigInt, time: Timestamptz, users_active_month: BigInt) -> Double;
   }
 
   sql_function! {

crates/db_views/src/comment_report_view.rs

@@ -432,7 +432,7 @@ mod tests {
         downvotes: 0,
         published: agg.published,
         child_count: 0,
-        hot_rank: 1728,
+        hot_rank: 0.1728,
         controversy_rank: 0.0,
       },
       my_vote: None,

crates/db_views/src/comment_view.rs

@@ -886,7 +886,7 @@ mod tests {
         downvotes: 0,
         published: agg.published,
         child_count: 5,
-        hot_rank: 1728,
+        hot_rank: 0.1728,
         controversy_rank: 0.0,
       },
     }

crates/db_views/src/post_view.rs

@@ -380,6 +380,9 @@ fn queries<'a>() -> Queries<
       SortType::Hot => query
         .then_order_by(post_aggregates::hot_rank.desc())
         .then_order_by(post_aggregates::published.desc()),
+      SortType::Scaled => query
+        .then_order_by(post_aggregates::scaled_rank.desc())
+        .then_order_by(post_aggregates::published.desc()),
       SortType::Controversial => query.then_order_by(post_aggregates::controversy_rank.desc()),
       SortType::New => query.then_order_by(post_aggregates::published.desc()),
       SortType::Old => query.then_order_by(post_aggregates::published.asc()),
@@ -1154,9 +1157,10 @@ mod tests {
         newest_comment_time: inserted_post.published,
         featured_community: false,
         featured_local: false,
-        hot_rank: 1728,
-        hot_rank_active: 1728,
+        hot_rank: 0.1728,
+        hot_rank_active: 0.1728,
         controversy_rank: 0.0,
+        scaled_rank: 0.3621,
         community_id: inserted_post.community_id,
         creator_id: inserted_post.creator_id,
       },

crates/db_views_actor/src/community_view.rs

@@ -105,7 +105,7 @@ fn queries<'a>() -> Queries<
     }
 
     match options.sort.unwrap_or(Hot) {
-      Hot | Active => query = query.order_by(community_aggregates::hot_rank.desc()),
+      Hot | Active | Scaled => query = query.order_by(community_aggregates::hot_rank.desc()),
       NewComments | TopDay | TopTwelveHour | TopSixHour | TopHour => {
         query = query.order_by(community_aggregates::users_active_day.desc())
       }

migrations/2023-08-23-182533_scaled_rank/down.sql

@@ -0,0 +1,87 @@
+DROP FUNCTION scaled_rank;
+
+ALTER TABLE community_aggregates
+    ALTER COLUMN hot_rank TYPE integer,
+    ALTER COLUMN hot_rank SET DEFAULT 1728;
+
+ALTER TABLE comment_aggregates
+    ALTER COLUMN hot_rank TYPE integer,
+    ALTER COLUMN hot_rank SET DEFAULT 1728;
+
+ALTER TABLE post_aggregates
+    ALTER COLUMN hot_rank TYPE integer,
+    ALTER COLUMN hot_rank SET DEFAULT 1728,
+    ALTER COLUMN hot_rank_active TYPE integer,
+    ALTER COLUMN hot_rank_active SET DEFAULT 1728;
+
+-- Change back to integer version
+DROP FUNCTION hot_rank (numeric, published timestamp with time zone);
+
+CREATE OR REPLACE FUNCTION hot_rank (score numeric, published timestamp with time zone)
+    RETURNS integer
+    AS $$
+DECLARE
+    hours_diff numeric := EXTRACT(EPOCH FROM (now() - published)) / 3600;
+BEGIN
+    IF (hours_diff > 0) THEN
+        RETURN floor(10000 * log(greatest (1, score + 3)) / power((hours_diff + 2), 1.8))::integer;
+    ELSE
+        -- if the post is from the future, set hot score to 0. otherwise you can game the post to
+        -- always be on top even with only 1 vote by setting it to the future
+        RETURN 0;
+    END IF;
+END;
+$$
+LANGUAGE plpgsql
+IMMUTABLE PARALLEL SAFE;
+
+ALTER TABLE post_aggregates
+    DROP COLUMN scaled_rank;
+
+-- The following code is necessary because postgres can't remove
+-- a single enum value.
+ALTER TABLE local_user
+    ALTER default_sort_type DROP DEFAULT;
+
+UPDATE
+    local_user
+SET
+    default_sort_type = 'Hot'
+WHERE
+    default_sort_type = 'Scaled';
+
+-- rename the old enum
+ALTER TYPE sort_type_enum RENAME TO sort_type_enum__;
+
+-- create the new enum
+CREATE TYPE sort_type_enum AS ENUM (
+    'Active',
+    'Hot',
+    'New',
+    'Old',
+    'TopDay',
+    'TopWeek',
+    'TopMonth',
+    'TopYear',
+    'TopAll',
+    'MostComments',
+    'NewComments',
+    'TopHour',
+    'TopSixHour',
+    'TopTwelveHour',
+    'TopThreeMonths',
+    'TopSixMonths',
+    'TopNineMonths'
+);
+
+-- alter all your enum columns
+ALTER TABLE local_user
+    ALTER COLUMN default_sort_type TYPE sort_type_enum
+    USING default_sort_type::text::sort_type_enum;
+
+ALTER TABLE local_user
+    ALTER default_sort_type SET DEFAULT 'Active';
+
+-- drop the old enum
+DROP TYPE sort_type_enum__;
+