lemmy/src/scheduled_tasks.rs

588 lines
17 KiB
Rust
Raw Normal View History

use chrono::{DateTime, TimeZone, Utc};
use clokwerk::{AsyncScheduler, TimeUnits as CTimeUnits};
use diesel::{
dsl::IntervalDsl,
sql_query,
sql_types::{Integer, Timestamptz},
ExpressionMethods,
NullableExpressionMethods,
QueryDsl,
2023-06-27 08:13:51 +00:00
QueryableByName,
};
use diesel_async::{AsyncPgConnection, RunQueryDsl};
use lemmy_api_common::context::LemmyContext;
use lemmy_db_schema::{
schema::{
captcha_answer,
comment,
community_person_ban,
instance,
person,
post,
received_activity,
sent_activity,
},
source::{
instance::{Instance, InstanceForm},
local_user::LocalUser,
},
utils::{get_conn, naive_now, now, DbPool, DELETED_REPLACEMENT_TEXT},
};
use lemmy_routes::nodeinfo::{NodeInfo, NodeInfoWellKnown};
use lemmy_utils::error::LemmyResult;
use reqwest_middleware::ClientWithMiddleware;
use std::time::Duration;
use tracing::{error, info, warn};
/// Schedules various cleanup tasks for lemmy in a background thread
pub async fn setup(context: LemmyContext) -> LemmyResult<()> {
2022-11-09 10:05:00 +00:00
// Setup the connections
let mut scheduler = AsyncScheduler::new();
startup_jobs(&mut context.pool()).await;
let context_1 = context.clone();
// Update active counts every hour
scheduler.every(CTimeUnits::hour(1)).run(move || {
let context = context_1.clone();
async move {
active_counts(&mut context.pool()).await;
update_banned_when_expired(&mut context.pool()).await;
}
});
let context_1 = context.clone();
2023-06-27 08:13:51 +00:00
// Update hot ranks every 15 minutes
scheduler.every(CTimeUnits::minutes(10)).run(move || {
let context = context_1.clone();
async move {
update_hot_ranks(&mut context.pool()).await;
}
});
let context_1 = context.clone();
2023-06-27 10:38:53 +00:00
// Delete any captcha answers older than ten minutes, every ten minutes
scheduler.every(CTimeUnits::minutes(10)).run(move || {
let context = context_1.clone();
async move {
delete_expired_captcha_answers(&mut context.pool()).await;
}
2023-06-27 10:38:53 +00:00
});
let context_1 = context.clone();
// Clear old activities every week
scheduler.every(CTimeUnits::weeks(1)).run(move || {
let context = context_1.clone();
async move {
clear_old_activities(&mut context.pool()).await;
}
});
let context_1 = context.clone();
// Daily tasks:
// - Overwrite deleted & removed posts and comments every day
// - Delete old denied users
// - Update instance software
scheduler.every(CTimeUnits::days(1)).run(move || {
let context = context_1.clone();
async move {
overwrite_deleted_posts_and_comments(&mut context.pool()).await;
delete_old_denied_users(&mut context.pool()).await;
update_instance_software(&mut context.pool(), context.client())
.await
.map_err(|e| warn!("Failed to update instance software: {e}"))
.ok();
}
});
// Manually run the scheduler in an event loop
loop {
scheduler.run_pending().await;
tokio::time::sleep(Duration::from_millis(1000)).await;
}
}
/// Run these on server startup
async fn startup_jobs(pool: &mut DbPool<'_>) {
active_counts(pool).await;
update_hot_ranks(pool).await;
update_banned_when_expired(pool).await;
clear_old_activities(pool).await;
overwrite_deleted_posts_and_comments(pool).await;
delete_old_denied_users(pool).await;
}
/// Update the hot_rank columns for the aggregates tables
2023-06-27 08:13:51 +00:00
/// Runs in batches until all necessary rows are updated once
async fn update_hot_ranks(pool: &mut DbPool<'_>) {
2023-07-17 09:05:55 +00:00
info!("Updating hot ranks for all history...");
2023-06-27 08:13:51 +00:00
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
process_post_aggregates_ranks_in_batches(&mut conn).await;
process_ranks_in_batches(
&mut conn,
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
"comment",
"a.hot_rank != 0",
Move SQL triggers from migrations into reusable sql file (#4333) * stuff * stuff including batch_upsert function * stuff * do things * stuff * different timestamps * stuff * Revert changes to comment.rs * Update comment.rs * Update comment.rs * Update post_view.rs * Update utils.rs * Update up.sql * Update up.sql * Update down.sql * Update up.sql * Update main.rs * use anyhow macro * Create down.sql * Create up.sql * Create replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update utils.rs * Update .woodpecker.yml * Update sql_format_check.sh * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Create dump_schema.sh * Update start_dev_db.sh * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * stuff * Update replaceable_schema.sql * Update .pg_format * fmt * stuff * stuff (#21) * Update replaceable_schema.sql * Update up.sql * Update replaceable_schema.sql * fmt * update cargo.lock * stuff * Update replaceable_schema.sql * Remove truncate trigger because truncate is already restricted by foreign keys * Update replaceable_schema.sql * fix some things * Update replaceable_schema.sql * Update replaceable_schema.sql * Update .woodpecker.yml * stuff * fix TG_OP * Psql env vars * try to fix combine_transition_tables parse error * Revert "try to fix combine_transition_tables parse error" This reverts commit 75d00a46266fbb49b7fab1b149c79fa1c31ee84a. * refactor combine_transition_tables * try to fix create_triggers * fix some things * try to fix combined_transition_tables * fix sql errors * update comment count in post trigger * fmt * Revert "fmt" This reverts commit a5bcd0834bb91a63b66bf63a848caa078f193940. * Revert "update comment count in post trigger" This reverts commit 0066a4b42b3472c088eed945605a2cf0bfcc1362. * fix everything * Update replaceable_schema.sql * actually fix everything * refactor create_triggers * fix * add semicolons * add is_counted function and fix incorrect bool operator in update_comment_count_from_post * refactor comment trigger * refactor post trigger * fix * Delete crates/db_schema/src/utils/series.rs * subscribers_local * edit migrations * move migrations * remove utils::series module declaration * fix everything * stuff * Move sql to schema_setup dir * utils.sql * delete .pg_format * Update .woodpecker.yml * Update sql_format_check.sh * Update .woodpecker.yml * Merge remote-tracking branch 'upstream/main' into bliss * fmt * Create main.rs * Update lib.rs * Update main.rs * Update .woodpecker.yml * Update main.rs * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update triggers.sql * YAY * Update mod.rs * Update Cargo.toml * a * Update Cargo.toml * Update Cargo.toml * Delete crates/db_schema/src/main.rs * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update utils.sql * Update utils.sql * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update down.sql * Update up.sql * Update triggers.sql * Update .woodpecker.yml * Update .woodpecker.yml * Update triggers.sql * Update down.sql * Update .woodpecker.yml * Update Cargo.toml * Update .woodpecker.yml * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update mod.rs * Update Cargo.toml * Update mod.rs * make dump_schema.sh executable * fix dump_schema.sh * defer * diff dumps * fmt * Update utils.sql * Update .woodpecker.yml * use correct version for pg_dump * Update .woodpecker.yml * Update .woodpecker.yml * change migration date * atomic site_aggregates insert * temporarily repeat tests in CI * drop r schema in CI migration check * show ReceivedActivity::create error * move check_diesel_migration CI step * Update .woodpecker.yml * Update scheduled_tasks.rs * Update scheduled_tasks.rs * update cargo.lock * move sql files * move rank functions * filter post_aggregates update * fmt * cargo fmt * replace post_id with id * update cargo.lock * avoid locking rows that need no change in up.sql * only run replaceable_schema if migrations were run * debug ci test failure * make replaceable_schema work in CI * Update .woodpecker.yml * remove println * Use migration revert and git checkout * Update schema_setup.rs * Fix * Update schema_setup.rs * Update schema_setup.rs * Update .woodpecker.yml --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com>
2024-04-18 00:58:44 +00:00
"SET hot_rank = r.hot_rank(a.score, a.published)",
)
.await;
process_ranks_in_batches(
&mut conn,
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
"community",
"a.hot_rank != 0",
Move SQL triggers from migrations into reusable sql file (#4333) * stuff * stuff including batch_upsert function * stuff * do things * stuff * different timestamps * stuff * Revert changes to comment.rs * Update comment.rs * Update comment.rs * Update post_view.rs * Update utils.rs * Update up.sql * Update up.sql * Update down.sql * Update up.sql * Update main.rs * use anyhow macro * Create down.sql * Create up.sql * Create replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update utils.rs * Update .woodpecker.yml * Update sql_format_check.sh * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Create dump_schema.sh * Update start_dev_db.sh * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * stuff * Update replaceable_schema.sql * Update .pg_format * fmt * stuff * stuff (#21) * Update replaceable_schema.sql * Update up.sql * Update replaceable_schema.sql * fmt * update cargo.lock * stuff * Update replaceable_schema.sql * Remove truncate trigger because truncate is already restricted by foreign keys * Update replaceable_schema.sql * fix some things * Update replaceable_schema.sql * Update replaceable_schema.sql * Update .woodpecker.yml * stuff * fix TG_OP * Psql env vars * try to fix combine_transition_tables parse error * Revert "try to fix combine_transition_tables parse error" This reverts commit 75d00a46266fbb49b7fab1b149c79fa1c31ee84a. * refactor combine_transition_tables * try to fix create_triggers * fix some things * try to fix combined_transition_tables * fix sql errors * update comment count in post trigger * fmt * Revert "fmt" This reverts commit a5bcd0834bb91a63b66bf63a848caa078f193940. * Revert "update comment count in post trigger" This reverts commit 0066a4b42b3472c088eed945605a2cf0bfcc1362. * fix everything * Update replaceable_schema.sql * actually fix everything * refactor create_triggers * fix * add semicolons * add is_counted function and fix incorrect bool operator in update_comment_count_from_post * refactor comment trigger * refactor post trigger * fix * Delete crates/db_schema/src/utils/series.rs * subscribers_local * edit migrations * move migrations * remove utils::series module declaration * fix everything * stuff * Move sql to schema_setup dir * utils.sql * delete .pg_format * Update .woodpecker.yml * Update sql_format_check.sh * Update .woodpecker.yml * Merge remote-tracking branch 'upstream/main' into bliss * fmt * Create main.rs * Update lib.rs * Update main.rs * Update .woodpecker.yml * Update main.rs * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update triggers.sql * YAY * Update mod.rs * Update Cargo.toml * a * Update Cargo.toml * Update Cargo.toml * Delete crates/db_schema/src/main.rs * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update utils.sql * Update utils.sql * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update down.sql * Update up.sql * Update triggers.sql * Update .woodpecker.yml * Update .woodpecker.yml * Update triggers.sql * Update down.sql * Update .woodpecker.yml * Update Cargo.toml * Update .woodpecker.yml * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update mod.rs * Update Cargo.toml * Update mod.rs * make dump_schema.sh executable * fix dump_schema.sh * defer * diff dumps * fmt * Update utils.sql * Update .woodpecker.yml * use correct version for pg_dump * Update .woodpecker.yml * Update .woodpecker.yml * change migration date * atomic site_aggregates insert * temporarily repeat tests in CI * drop r schema in CI migration check * show ReceivedActivity::create error * move check_diesel_migration CI step * Update .woodpecker.yml * Update scheduled_tasks.rs * Update scheduled_tasks.rs * update cargo.lock * move sql files * move rank functions * filter post_aggregates update * fmt * cargo fmt * replace post_id with id * update cargo.lock * avoid locking rows that need no change in up.sql * only run replaceable_schema if migrations were run * debug ci test failure * make replaceable_schema work in CI * Update .woodpecker.yml * remove println * Use migration revert and git checkout * Update schema_setup.rs * Fix * Update schema_setup.rs * Update schema_setup.rs * Update .woodpecker.yml --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com>
2024-04-18 00:58:44 +00:00
"SET hot_rank = r.hot_rank(a.subscribers, a.published)",
)
.await;
info!("Finished hot ranks update!");
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
2023-06-27 08:13:51 +00:00
}
2023-06-27 08:13:51 +00:00
#[derive(QueryableByName)]
struct HotRanksUpdateResult {
#[diesel(sql_type = Timestamptz)]
published: DateTime<Utc>,
2023-06-27 08:13:51 +00:00
}
2023-07-17 09:05:55 +00:00
/// Runs the hot rank update query in batches until all rows have been processed.
/// In `where_clause` and `set_clause`, "a" will refer to the current aggregates table.
2023-06-27 08:13:51 +00:00
/// Locked rows are skipped in order to prevent deadlocks (they will likely get updated on the next
/// run)
async fn process_ranks_in_batches(
conn: &mut AsyncPgConnection,
2023-06-27 08:13:51 +00:00
table_name: &str,
2023-07-17 09:05:55 +00:00
where_clause: &str,
2023-06-27 08:13:51 +00:00
set_clause: &str,
) {
let process_start_time: DateTime<Utc> = Utc
.timestamp_opt(0, 0)
.single()
.expect("0 timestamp creation");
2023-07-17 09:05:55 +00:00
2023-06-27 08:13:51 +00:00
let update_batch_size = 1000; // Bigger batches than this tend to cause seq scans
2023-07-17 09:05:55 +00:00
let mut processed_rows_count = 0;
2023-06-27 08:13:51 +00:00
let mut previous_batch_result = Some(process_start_time);
while let Some(previous_batch_last_published) = previous_batch_result {
// Raw `sql_query` is used as a performance optimization - Diesel does not support doing this
// in a single query (neither as a CTE, nor using a subquery)
let result = sql_query(format!(
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
r#"WITH batch AS (SELECT a.{id_column}
2023-06-27 08:13:51 +00:00
FROM {aggregates_table} a
2023-07-17 09:05:55 +00:00
WHERE a.published > $1 AND ({where_clause})
2023-06-27 08:13:51 +00:00
ORDER BY a.published
LIMIT $2
FOR UPDATE SKIP LOCKED)
UPDATE {aggregates_table} a {set_clause}
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
FROM batch WHERE a.{id_column} = batch.{id_column} RETURNING a.published;
2023-06-27 08:13:51 +00:00
"#,
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
id_column = format!("{table_name}_id"),
aggregates_table = format!("{table_name}_aggregates"),
2023-07-17 09:05:55 +00:00
set_clause = set_clause,
where_clause = where_clause
2023-06-27 08:13:51 +00:00
))
.bind::<Timestamptz, _>(previous_batch_last_published)
2023-06-27 08:13:51 +00:00
.bind::<Integer, _>(update_batch_size)
.get_results::<HotRanksUpdateResult>(conn)
.await;
2023-06-27 08:13:51 +00:00
match result {
2023-07-17 09:05:55 +00:00
Ok(updated_rows) => {
processed_rows_count += updated_rows.len();
previous_batch_result = updated_rows.last().map(|row| row.published);
}
2023-06-27 08:13:51 +00:00
Err(e) => {
error!("Failed to update {} hot_ranks: {}", table_name, e);
break;
}
}
}
2023-06-27 08:13:51 +00:00
info!(
2023-07-17 09:05:55 +00:00
"Finished process_hot_ranks_in_batches execution for {} (processed {} rows)",
table_name, processed_rows_count
2023-06-27 08:13:51 +00:00
);
}
/// Post aggregates is a special case, since it needs to join to the community_aggregates
/// table, to get the active monthly user counts.
async fn process_post_aggregates_ranks_in_batches(conn: &mut AsyncPgConnection) {
let process_start_time: DateTime<Utc> = Utc
.timestamp_opt(0, 0)
.single()
.expect("0 timestamp creation");
let update_batch_size = 1000; // Bigger batches than this tend to cause seq scans
let mut processed_rows_count = 0;
let mut previous_batch_result = Some(process_start_time);
while let Some(previous_batch_last_published) = previous_batch_result {
let result = sql_query(
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
r#"WITH batch AS (SELECT pa.post_id
FROM post_aggregates pa
WHERE pa.published > $1
AND (pa.hot_rank != 0 OR pa.hot_rank_active != 0)
ORDER BY pa.published
LIMIT $2
FOR UPDATE SKIP LOCKED)
UPDATE post_aggregates pa
Move SQL triggers from migrations into reusable sql file (#4333) * stuff * stuff including batch_upsert function * stuff * do things * stuff * different timestamps * stuff * Revert changes to comment.rs * Update comment.rs * Update comment.rs * Update post_view.rs * Update utils.rs * Update up.sql * Update up.sql * Update down.sql * Update up.sql * Update main.rs * use anyhow macro * Create down.sql * Create up.sql * Create replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update utils.rs * Update .woodpecker.yml * Update sql_format_check.sh * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Create dump_schema.sh * Update start_dev_db.sh * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * Update replaceable_schema.sql * stuff * Update replaceable_schema.sql * Update .pg_format * fmt * stuff * stuff (#21) * Update replaceable_schema.sql * Update up.sql * Update replaceable_schema.sql * fmt * update cargo.lock * stuff * Update replaceable_schema.sql * Remove truncate trigger because truncate is already restricted by foreign keys * Update replaceable_schema.sql * fix some things * Update replaceable_schema.sql * Update replaceable_schema.sql * Update .woodpecker.yml * stuff * fix TG_OP * Psql env vars * try to fix combine_transition_tables parse error * Revert "try to fix combine_transition_tables parse error" This reverts commit 75d00a46266fbb49b7fab1b149c79fa1c31ee84a. * refactor combine_transition_tables * try to fix create_triggers * fix some things * try to fix combined_transition_tables * fix sql errors * update comment count in post trigger * fmt * Revert "fmt" This reverts commit a5bcd0834bb91a63b66bf63a848caa078f193940. * Revert "update comment count in post trigger" This reverts commit 0066a4b42b3472c088eed945605a2cf0bfcc1362. * fix everything * Update replaceable_schema.sql * actually fix everything * refactor create_triggers * fix * add semicolons * add is_counted function and fix incorrect bool operator in update_comment_count_from_post * refactor comment trigger * refactor post trigger * fix * Delete crates/db_schema/src/utils/series.rs * subscribers_local * edit migrations * move migrations * remove utils::series module declaration * fix everything * stuff * Move sql to schema_setup dir * utils.sql * delete .pg_format * Update .woodpecker.yml * Update sql_format_check.sh * Update .woodpecker.yml * Merge remote-tracking branch 'upstream/main' into bliss * fmt * Create main.rs * Update lib.rs * Update main.rs * Update .woodpecker.yml * Update main.rs * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update triggers.sql * YAY * Update mod.rs * Update Cargo.toml * a * Update Cargo.toml * Update Cargo.toml * Delete crates/db_schema/src/main.rs * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update utils.sql * Update utils.sql * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update down.sql * Update up.sql * Update triggers.sql * Update .woodpecker.yml * Update .woodpecker.yml * Update triggers.sql * Update down.sql * Update .woodpecker.yml * Update Cargo.toml * Update .woodpecker.yml * Update Cargo.toml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update .woodpecker.yml * Update mod.rs * Update Cargo.toml * Update mod.rs * make dump_schema.sh executable * fix dump_schema.sh * defer * diff dumps * fmt * Update utils.sql * Update .woodpecker.yml * use correct version for pg_dump * Update .woodpecker.yml * Update .woodpecker.yml * change migration date * atomic site_aggregates insert * temporarily repeat tests in CI * drop r schema in CI migration check * show ReceivedActivity::create error * move check_diesel_migration CI step * Update .woodpecker.yml * Update scheduled_tasks.rs * Update scheduled_tasks.rs * update cargo.lock * move sql files * move rank functions * filter post_aggregates update * fmt * cargo fmt * replace post_id with id * update cargo.lock * avoid locking rows that need no change in up.sql * only run replaceable_schema if migrations were run * debug ci test failure * make replaceable_schema work in CI * Update .woodpecker.yml * remove println * Use migration revert and git checkout * Update schema_setup.rs * Fix * Update schema_setup.rs * Update schema_setup.rs * Update .woodpecker.yml --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: Dessalines <dessalines@users.noreply.github.com>
2024-04-18 00:58:44 +00:00
SET hot_rank = r.hot_rank(pa.score, pa.published),
hot_rank_active = r.hot_rank(pa.score, pa.newest_comment_time_necro),
scaled_rank = r.scaled_rank(pa.score, pa.published, ca.users_active_month)
FROM batch, community_aggregates ca
Remove id column and use different primary key on some tables (#4093) * post_saved * fmt * remove unique and not null * put person_id first in primary key and remove index * use post_saved.find * change captcha_answer * remove removal of not null * comment_aggregates * comment_like * comment_saved * aggregates * remove "\" * deduplicate site_aggregates * person_post_aggregates * community_moderator * community_block * community_person_ban * custom_emoji_keyword * federation allow/block list * federation_queue_state * instance_block * local_site_rate_limit, local_user_language, login_token * person_ban, person_block, person_follower, post_like, post_read, received_activity * community_follower, community_language, site_language * fmt * image_upload * remove unused newtypes * remove more indexes * use .find * merge * fix site_aggregates_site function * fmt * Primary keys dess (#17) * Also order reports by oldest first (ref #4123) (#4129) * Support signed fetch for federation (fixes #868) (#4125) * Support signed fetch for federation (fixes #868) * taplo * add federation queue state to get_federated_instances api (#4104) * add federation queue state to get_federated_instances api * feature gate * move retry sleep function * move stuff around * Add UI setting for collapsing bot comments. Fixes #3838 (#4098) * Add UI setting for collapsing bot comments. Fixes #3838 * Fixing clippy check. * Only keep sent and received activities for 7 days (fixes #4113, fixes #4110) (#4131) * Only check auth secure on release mode. (#4127) * Only check auth secure on release mode. * Fixing wrong js-client. * Adding is_debug_mode var. * Fixing the desktop image on the README. (#4135) * Delete dupes and add possibly missing unique constraint on person_aggregates. * Fixing clippy lints. --------- Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com> * fmt * Update community_block.rs * Update instance_block.rs * Update person_block.rs * Update person_block.rs --------- Co-authored-by: Dessalines <dessalines@users.noreply.github.com> Co-authored-by: Nutomic <me@nutomic.com> Co-authored-by: phiresky <phireskyde+git@gmail.com>
2023-11-13 13:14:07 +00:00
WHERE pa.post_id = batch.post_id and pa.community_id = ca.community_id RETURNING pa.published;
"#,
)
.bind::<Timestamptz, _>(previous_batch_last_published)
.bind::<Integer, _>(update_batch_size)
.get_results::<HotRanksUpdateResult>(conn)
.await;
match result {
Ok(updated_rows) => {
processed_rows_count += updated_rows.len();
previous_batch_result = updated_rows.last().map(|row| row.published);
}
Err(e) => {
error!("Failed to update {} hot_ranks: {}", "post_aggregates", e);
break;
}
}
}
info!(
"Finished process_hot_ranks_in_batches execution for {} (processed {} rows)",
"post_aggregates", processed_rows_count
);
}
async fn delete_expired_captcha_answers(pool: &mut DbPool<'_>) {
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
diesel::delete(
captcha_answer::table
.filter(captcha_answer::published.lt(now() - IntervalDsl::minutes(10))),
)
.execute(&mut conn)
.await
.map(|_| {
info!("Done.");
})
.map_err(|e| error!("Failed to clear old captcha answers: {e}"))
.ok();
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
2023-06-27 10:38:53 +00:00
}
/// Clear old activities (this table gets very large)
async fn clear_old_activities(pool: &mut DbPool<'_>) {
info!("Clearing old activities...");
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
diesel::delete(
sent_activity::table.filter(sent_activity::published.lt(now() - IntervalDsl::days(7))),
)
.execute(&mut conn)
.await
.map_err(|e| error!("Failed to clear old sent activities: {e}"))
.ok();
diesel::delete(
received_activity::table
.filter(received_activity::published.lt(now() - IntervalDsl::days(7))),
)
.execute(&mut conn)
.await
.map(|_| info!("Done."))
.map_err(|e| error!("Failed to clear old received activities: {e}"))
.ok();
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
}
async fn delete_old_denied_users(pool: &mut DbPool<'_>) {
LocalUser::delete_old_denied_local_users(pool)
.await
.map(|_| {
info!("Done.");
})
.map_err(|e| error!("Failed to deleted old denied users: {e}"))
.ok();
}
/// overwrite posts and comments 30d after deletion
async fn overwrite_deleted_posts_and_comments(pool: &mut DbPool<'_>) {
info!("Overwriting deleted posts...");
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
diesel::update(
post::table
.filter(post::deleted.eq(true))
.filter(post::updated.lt(now().nullable() - 1.months()))
.filter(post::body.ne(DELETED_REPLACEMENT_TEXT)),
)
.set((
post::body.eq(DELETED_REPLACEMENT_TEXT),
post::name.eq(DELETED_REPLACEMENT_TEXT),
))
.execute(&mut conn)
.await
.map(|_| {
info!("Done.");
})
.map_err(|e| error!("Failed to overwrite deleted posts: {e}"))
.ok();
info!("Overwriting deleted comments...");
diesel::update(
comment::table
.filter(comment::deleted.eq(true))
.filter(comment::updated.lt(now().nullable() - 1.months()))
.filter(comment::content.ne(DELETED_REPLACEMENT_TEXT)),
)
.set(comment::content.eq(DELETED_REPLACEMENT_TEXT))
.execute(&mut conn)
.await
.map(|_| {
info!("Done.");
})
.map_err(|e| error!("Failed to overwrite deleted comments: {e}"))
.ok();
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
}
/// Re-calculate the site and community active counts every 12 hours
async fn active_counts(pool: &mut DbPool<'_>) {
info!("Updating active site and community aggregates ...");
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
let intervals = vec![
("1 day", "day"),
("1 week", "week"),
("1 month", "month"),
("6 months", "half_year"),
];
for i in &intervals {
let update_site_stmt = format!(
"update site_aggregates set users_active_{} = (select * from site_aggregates_activity('{}')) where site_id = 1",
i.1, i.0
);
sql_query(update_site_stmt)
.execute(&mut conn)
.await
.map_err(|e| error!("Failed to update site stats: {e}"))
.ok();
let update_community_stmt = format!("update community_aggregates ca set users_active_{} = mv.count_ from community_aggregates_activity('{}') mv where ca.community_id = mv.community_id_", i.1, i.0);
sql_query(update_community_stmt)
.execute(&mut conn)
.await
.map_err(|e| error!("Failed to update community stats: {e}"))
.ok();
}
info!("Done.");
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
}
/// Set banned to false after ban expires
async fn update_banned_when_expired(pool: &mut DbPool<'_>) {
info!("Updating banned column if it expires ...");
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
diesel::update(
person::table
.filter(person::banned.eq(true))
.filter(person::ban_expires.lt(now().nullable())),
)
.set(person::banned.eq(false))
.execute(&mut conn)
.await
.map_err(|e| error!("Failed to update person.banned when expires: {e}"))
.ok();
diesel::delete(
community_person_ban::table.filter(community_person_ban::expires.lt(now().nullable())),
)
.execute(&mut conn)
.await
.map_err(|e| error!("Failed to remove community_ban expired rows: {e}"))
.ok();
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
}
/// Updates the instance software and version.
///
/// Does so using the /.well-known/nodeinfo protocol described here:
/// https://github.com/jhass/nodeinfo/blob/main/PROTOCOL.md
///
/// TODO: if instance has been dead for a long time, it should be checked less frequently
async fn update_instance_software(
pool: &mut DbPool<'_>,
client: &ClientWithMiddleware,
) -> LemmyResult<()> {
info!("Updating instances software and versions...");
let conn = get_conn(pool).await;
match conn {
Ok(mut conn) => {
let instances = instance::table.get_results::<Instance>(&mut conn).await?;
for instance in instances {
if let Some(form) = build_update_instance_form(&instance.domain, client).await {
Instance::update(pool, instance.id, form).await?;
}
}
info!("Finished updating instances software and versions...");
}
Err(e) => {
error!("Failed to get connection from pool: {e}");
}
}
Ok(())
}
/// This builds an instance update form, for a given domain.
/// If the instance sends a response, but doesn't have a well-known or nodeinfo,
/// Then return a default form with only the updated field.
async fn build_update_instance_form(
domain: &str,
client: &ClientWithMiddleware,
) -> Option<InstanceForm> {
// The `updated` column is used to check if instances are alive. If it is more than three
// days in the past, no outgoing activities will be sent to that instance. However
// not every Fediverse instance has a valid Nodeinfo endpoint (its not required for
// Activitypub). That's why we always need to mark instances as updated if they are
// alive.
let mut instance_form = InstanceForm::builder()
.domain(domain.to_string())
.updated(Some(naive_now()))
.build();
// First, fetch their /.well-known/nodeinfo, then extract the correct nodeinfo link from it
let well_known_url = format!("https://{}/.well-known/nodeinfo", domain);
let Ok(res) = client.get(&well_known_url).send().await else {
// This is the only kind of error that means the instance is dead
return None;
};
// In this block, returning `None` is ignored, and only means not writing nodeinfo to db
async {
if res.status().is_client_error() {
return None;
}
let node_info_url = res
.json::<NodeInfoWellKnown>()
.await
.ok()?
.links
.into_iter()
.find(|links| {
links
.rel
.as_str()
.starts_with("http://nodeinfo.diaspora.software/ns/schema/2.")
})?
.href;
let software = client
.get(node_info_url)
.send()
.await
.ok()?
.json::<NodeInfo>()
.await
.ok()?
.software?;
instance_form.software = software.name;
instance_form.version = software.version;
Some(())
}
.await;
Some(instance_form)
}
#[cfg(test)]
#[allow(clippy::indexing_slicing)]
mod tests {
use crate::scheduled_tasks::build_update_instance_form;
use lemmy_api_common::request::client_builder;
use lemmy_utils::{error::LemmyResult, settings::structs::Settings, LemmyErrorType};
use pretty_assertions::assert_eq;
use reqwest_middleware::ClientBuilder;
use serial_test::serial;
#[tokio::test]
#[serial]
async fn test_nodeinfo_voyager_lemmy_ml() -> LemmyResult<()> {
let client = ClientBuilder::new(client_builder(&Settings::default()).build()?).build();
let form = build_update_instance_form("voyager.lemmy.ml", &client)
.await
.ok_or(LemmyErrorType::CouldntFindObject)?;
assert_eq!(
form.software.ok_or(LemmyErrorType::CouldntFindObject)?,
"lemmy"
);
Ok(())
}
#[tokio::test]
#[serial]
async fn test_nodeinfo_mastodon_social() -> LemmyResult<()> {
let client = ClientBuilder::new(client_builder(&Settings::default()).build()?).build();
let form = build_update_instance_form("mastodon.social", &client)
.await
.ok_or(LemmyErrorType::CouldntFindObject)?;
assert_eq!(
form.software.ok_or(LemmyErrorType::CouldntFindObject)?,
"mastodon"
);
Ok(())
}
}