mirror of
https://github.com/LemmyNet/lemmy.git
synced 2024-11-22 14:45:10 +00:00
stuff
This commit is contained in:
parent
fdebc85753
commit
9b9314a1d6
|
@ -172,6 +172,7 @@ impl Object for ApubComment {
|
|||
deleted: Some(false),
|
||||
ap_id: Some(note.id.into()),
|
||||
distinguished: note.distinguished,
|
||||
path: None,
|
||||
local: Some(false),
|
||||
language_id,
|
||||
};
|
||||
|
|
|
@ -166,16 +166,18 @@ impl SiteLanguage {
|
|||
.execute(conn)
|
||||
.await?;
|
||||
|
||||
for l in lang_ids {
|
||||
let form = SiteLanguageForm {
|
||||
let forms = lang_ids
|
||||
.into_iter()
|
||||
.map(|l| SiteLanguageForm {
|
||||
site_id: for_site_id,
|
||||
language_id: l,
|
||||
};
|
||||
insert_into(site_language)
|
||||
.values(form)
|
||||
.get_result::<Self>(conn)
|
||||
.await?;
|
||||
}
|
||||
})
|
||||
.collect::<Vec<_>>();
|
||||
|
||||
insert_into(site_language)
|
||||
.values(forms)
|
||||
.get_result::<Self>(conn)
|
||||
.await?;
|
||||
|
||||
CommunityLanguage::limit_languages(conn, instance_id).await?;
|
||||
|
||||
|
|
|
@ -1,6 +1,16 @@
|
|||
use crate::{
|
||||
newtypes::{CommentId, DbUrl, PersonId},
|
||||
schema::comment::dsl::{ap_id, comment, content, creator_id, deleted, path, removed, updated},
|
||||
schema::comment::dsl::{
|
||||
ap_id,
|
||||
comment,
|
||||
content,
|
||||
creator_id,
|
||||
deleted,
|
||||
id,
|
||||
path,
|
||||
removed,
|
||||
updated,
|
||||
},
|
||||
source::comment::{
|
||||
Comment,
|
||||
CommentInsertForm,
|
||||
|
@ -11,16 +21,24 @@ use crate::{
|
|||
CommentUpdateForm,
|
||||
},
|
||||
traits::{Crud, Likeable, Saveable},
|
||||
utils::{get_conn, naive_now, DbPool, DELETED_REPLACEMENT_TEXT},
|
||||
utils::{functions::AsText, get_conn, naive_now, DbPool, DELETED_REPLACEMENT_TEXT},
|
||||
};
|
||||
use diesel::{
|
||||
dsl::{insert_into, sql_query},
|
||||
result::Error,
|
||||
ExpressionMethods,
|
||||
QueryDsl,
|
||||
TextExpressionMethods,
|
||||
};
|
||||
use diesel_async::RunQueryDsl;
|
||||
use diesel_ltree::Ltree;
|
||||
use diesel_ltree::{
|
||||
functions::{ltree2text, text2ltree},
|
||||
Ltree,
|
||||
};
|
||||
use futures_util::{
|
||||
future::TryFutureExt,
|
||||
stream::{self, StreamExt, TryStreamExt},
|
||||
};
|
||||
use url::Url;
|
||||
|
||||
impl Comment {
|
||||
|
@ -57,12 +75,16 @@ impl Comment {
|
|||
comment_form: &CommentInsertForm,
|
||||
parent_path: Option<&Ltree>,
|
||||
) -> Result<Comment, Error> {
|
||||
Comment::create_batch(pool, &[(comment_form, parent_path)]).await?.into_iter().next().ok_or(Error::NotFound)
|
||||
Comment::create_batch(pool, &[(comment_form.clone(), parent_path.cloned())])
|
||||
.await?
|
||||
.into_iter()
|
||||
.next()
|
||||
.ok_or(Error::NotFound)
|
||||
}
|
||||
|
||||
|
||||
pub async fn create_batch(
|
||||
pool: &mut DbPool<'_>,
|
||||
items: &[(&CommentInsertForm, Option<&Ltree>)],
|
||||
items: &[(CommentInsertForm, Option<Ltree>)],
|
||||
) -> Result<Vec<Comment>, Error> {
|
||||
let conn = &mut get_conn(pool).await?;
|
||||
|
||||
|
@ -72,43 +94,44 @@ impl Comment {
|
|||
Box::pin(async move {
|
||||
let forms = items
|
||||
.iter()
|
||||
.map(|&(form, _)| form)
|
||||
.collect::<Vec<_>>();
|
||||
.map(|(comment_form, parent_path)| CommentInsertForm {
|
||||
path: Some(parent_path.clone().unwrap_or(Ltree("0".to_owned()))),
|
||||
..comment_form.clone()
|
||||
});
|
||||
|
||||
// Insert, to get the ids
|
||||
let inserted_comments = insert_into(comment)
|
||||
.values(&forms)
|
||||
/*.on_conflict(ap_id)
|
||||
.do_update()
|
||||
.set()*/
|
||||
.get_result::<Self>(conn)
|
||||
.values(forms.clone().collect::<Vec<_>>())
|
||||
.load::<Self>(conn)
|
||||
.or_else(|_| {
|
||||
// `ap_id` unique constraint violation is handled individually for each row
|
||||
// because batched upsert requires having the same `set` argument for all rows
|
||||
stream::iter(forms)
|
||||
.then(|form| {
|
||||
insert_into(comment)
|
||||
.values(form.clone())
|
||||
.on_conflict(ap_id)
|
||||
.do_update()
|
||||
.set(form)
|
||||
.get_result::<Self>(conn)
|
||||
})
|
||||
.try_collect::<Vec<_>>()
|
||||
})
|
||||
.await?;
|
||||
|
||||
// `ap_id` unique constraint violation is handled individually for each row
|
||||
// because batch upsert requires having the same `set` argument for all rows
|
||||
|
||||
|
||||
let comment_id = inserted_comment.id;
|
||||
|
||||
// You need to update the ltree column
|
||||
let ltree = Ltree(if let Some(parent_path) = parent_path {
|
||||
// The previous parent will already have 0 in it
|
||||
// Append this comment id
|
||||
format!("{}.{}", parent_path.0, comment_id)
|
||||
} else {
|
||||
// '0' is always the first path, append to that
|
||||
format!("{}.{}", 0, comment_id)
|
||||
});
|
||||
|
||||
let updated_comment = diesel::update(comment.find(comment_id))
|
||||
.set(path.eq(ltree))
|
||||
.get_result::<Self>(conn)
|
||||
.await;
|
||||
// For each comment, append its id to its path
|
||||
let updated_comments = diesel::update(comment)
|
||||
.filter(id.eq_any(inserted_comments.into_iter().map(|c| c.id)))
|
||||
.set(path.eq(text2ltree(
|
||||
ltree2text(path).concat(".").concat(AsText::new(id)),
|
||||
)))
|
||||
.load::<Self>(conn)
|
||||
.await?;
|
||||
|
||||
// Update the child count for the parent comment_aggregates
|
||||
// You could do this with a trigger, but since you have to do this manually anyway,
|
||||
// you can just have it here
|
||||
if let Some(parent_path) = parent_path {
|
||||
for parent_path in items.iter().filter_map(|(_, p)| p.as_ref()) {
|
||||
// You have to update counts for all parents, not just the immediate one
|
||||
// TODO if the performance of this is terrible, it might be better to do this as part of a
|
||||
// scheduled query... although the counts would often be wrong.
|
||||
|
@ -137,7 +160,7 @@ where ca.comment_id = c.id"
|
|||
sql_query(update_child_count_stmt).execute(conn).await?;
|
||||
}
|
||||
}
|
||||
updated_comment
|
||||
Ok(updated_comments)
|
||||
}) as _
|
||||
})
|
||||
.await
|
||||
|
|
|
@ -67,6 +67,10 @@ pub struct CommentInsertForm {
|
|||
pub deleted: Option<bool>,
|
||||
pub ap_id: Option<DbUrl>,
|
||||
pub local: Option<bool>,
|
||||
#[cfg(feature = "full")]
|
||||
pub path: Option<Ltree>,
|
||||
#[cfg(not(feature = "full"))]
|
||||
pub path: Option<String>,
|
||||
pub distinguished: Option<bool>,
|
||||
pub language_id: Option<LanguageId>,
|
||||
}
|
||||
|
|
|
@ -13,13 +13,15 @@ use diesel::{
|
|||
deserialize::FromSql,
|
||||
helper_types::AsExprOf,
|
||||
pg::Pg,
|
||||
query_builder::{Query, QueryFragment},
|
||||
query_dsl::methods::LimitDsl,
|
||||
result::{ConnectionError, ConnectionResult, Error as DieselError, Error::QueryBuilderError},
|
||||
serialize::{Output, ToSql},
|
||||
sql_query,
|
||||
sql_types::{Text, Timestamptz},
|
||||
IntoSql,
|
||||
PgConnection,
|
||||
RunQueryDsl, Insertable, Table, Column, AsChangeset, Expression, SelectableExpression, expression::NonAggregate, query_builder::QueryFragment,
|
||||
RunQueryDsl,
|
||||
};
|
||||
use diesel_async::{
|
||||
pg::AsyncPgConnection,
|
||||
|
@ -153,17 +155,65 @@ macro_rules! try_join_with_pool {
|
|||
}};
|
||||
}
|
||||
|
||||
pub async fn batch_upsert<T, U, Target, R>(conn: &mut AsyncPgConnection, target: T, records: U, conflict_target: Target) -> Result<Vec<R>, DieselError>
|
||||
where
|
||||
T: Table,
|
||||
T::AllColumns: Expression + SelectableExpression<T> + NonAggregate + QueryFragment<Pg>,
|
||||
U: IntoIterator + Clone,
|
||||
Vec<U::Item>: Insertable<T>,
|
||||
U::Item: Insertable<T> + AsChangeset<Target = T>,
|
||||
Target: Column<Table = T>,
|
||||
{
|
||||
let result = diesel::insert_into(target).values(records.clone().into_iter().collect::<Vec<_>>()).load::<R>(conn).await;
|
||||
|
||||
/// Includes an SQL comment before `T`, which can be used to label auto_explain output
|
||||
#[derive(QueryId)]
|
||||
pub struct Commented<T> {
|
||||
comment: String,
|
||||
inner: T,
|
||||
}
|
||||
|
||||
impl<T> Commented<T> {
|
||||
pub fn new(inner: T) -> Self {
|
||||
Commented {
|
||||
comment: String::new(),
|
||||
inner,
|
||||
}
|
||||
}
|
||||
|
||||
/// Adds `text` to the comment if `condition` is true
|
||||
pub fn text_if(mut self, text: &str, condition: bool) -> Self {
|
||||
if condition {
|
||||
if !self.comment.is_empty() {
|
||||
self.comment.push_str(", ");
|
||||
}
|
||||
self.comment.push_str(text);
|
||||
}
|
||||
self
|
||||
}
|
||||
|
||||
/// Adds `text` to the comment
|
||||
pub fn text(self, text: &str) -> Self {
|
||||
self.text_if(text, true)
|
||||
}
|
||||
}
|
||||
|
||||
impl<T: Query> Query for Commented<T> {
|
||||
type SqlType = T::SqlType;
|
||||
}
|
||||
|
||||
impl<T: QueryFragment<Pg>> QueryFragment<Pg> for Commented<T> {
|
||||
fn walk_ast<'b>(
|
||||
&'b self,
|
||||
mut out: diesel::query_builder::AstPass<'_, 'b, Pg>,
|
||||
) -> Result<(), DieselError> {
|
||||
for line in self.comment.lines() {
|
||||
out.push_sql("\n-- ");
|
||||
out.push_sql(line);
|
||||
}
|
||||
out.push_sql("\n");
|
||||
self.inner.walk_ast(out.reborrow())
|
||||
}
|
||||
}
|
||||
|
||||
impl<T: LimitDsl> LimitDsl for Commented<T> {
|
||||
type Output = Commented<T::Output>;
|
||||
|
||||
fn limit(self, limit: i64) -> Self::Output {
|
||||
Commented {
|
||||
comment: self.comment,
|
||||
inner: self.inner.limit(limit),
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
pub fn fuzzy_search(q: &str) -> String {
|
||||
|
@ -368,7 +418,10 @@ static EMAIL_REGEX: Lazy<Regex> = Lazy::new(|| {
|
|||
});
|
||||
|
||||
pub mod functions {
|
||||
use diesel::sql_types::{BigInt, Text, Timestamptz};
|
||||
use diesel::{
|
||||
pg::Pg,
|
||||
sql_types::{BigInt, Text, Timestamptz},
|
||||
};
|
||||
|
||||
sql_function! {
|
||||
fn hot_rank(score: BigInt, time: Timestamptz) -> Double;
|
||||
|
@ -386,6 +439,9 @@ pub mod functions {
|
|||
|
||||
// really this function is variadic, this just adds the two-argument version
|
||||
sql_function!(fn coalesce<T: diesel::sql_types::SqlType + diesel::sql_types::SingleValue>(x: diesel::sql_types::Nullable<T>, y: T) -> T);
|
||||
|
||||
// Use `AsText::new`
|
||||
postfix_operator!(AsText, "::text", Text, backend: Pg);
|
||||
}
|
||||
|
||||
pub const DELETED_REPLACEMENT_TEXT: &str = "*Permanently Deleted*";
|
||||
|
|
|
@ -43,6 +43,7 @@ use lemmy_db_schema::{
|
|||
get_conn,
|
||||
limit_and_offset,
|
||||
now,
|
||||
Commented,
|
||||
DbConn,
|
||||
DbPool,
|
||||
ListFn,
|
||||
|
@ -282,7 +283,10 @@ fn queries<'a>() -> Queries<
|
|||
);
|
||||
}
|
||||
|
||||
query.first::<PostView>(&mut conn).await
|
||||
Commented::new(query)
|
||||
.text("PostView::read")
|
||||
.first::<PostView>(&mut conn)
|
||||
.await
|
||||
};
|
||||
|
||||
let list = move |mut conn: DbConn<'a>, options: PostQuery<'a>| async move {
|
||||
|
@ -557,7 +561,14 @@ fn queries<'a>() -> Queries<
|
|||
|
||||
debug!("Post View Query: {:?}", debug_query::<Pg, _>(&query));
|
||||
|
||||
query.load::<PostView>(&mut conn).await
|
||||
Commented::new(query)
|
||||
.text("PostQuery::list")
|
||||
.text_if(
|
||||
"getting upper bound for next query",
|
||||
options.community_id_just_for_prefetch,
|
||||
)
|
||||
.load::<PostView>(&mut conn)
|
||||
.await
|
||||
};
|
||||
|
||||
Queries::new(read, list)
|
||||
|
@ -1457,12 +1468,13 @@ mod tests {
|
|||
.post_id(inserted_post.id)
|
||||
.content("yes".to_owned())
|
||||
.build();
|
||||
let inserted_comment = Comment::create(pool, &comment_form, None).await.unwrap();
|
||||
inserted_comment_ids.push(inserted_comment.id);
|
||||
comment_forms.push((comment_form, None));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Comment::create_batch(pool, &comment_forms).await.unwrap();
|
||||
|
||||
let mut listed_post_ids = vec![];
|
||||
let mut page_after = None;
|
||||
loop {
|
||||
|
|
|
@ -17,24 +17,34 @@ then
|
|||
rm -rf $PGDATA
|
||||
fi
|
||||
|
||||
# Create cluster
|
||||
initdb --username=postgres --auth=trust --no-instructions \
|
||||
config_args=(
|
||||
# Only listen to socket in current directory
|
||||
-c listen_addresses= -c unix_socket_directories=$PWD \
|
||||
-c listen_addresses=
|
||||
-c unix_socket_directories=$PWD
|
||||
|
||||
# Write logs to a file in $PGDATA/log
|
||||
-c logging_collector=on \
|
||||
-c logging_collector=on
|
||||
|
||||
# Log all query plans by default
|
||||
-c session_preload_libraries=auto_explain -c auto_explain.log_min_duration=0
|
||||
-c session_preload_libraries=auto_explain
|
||||
-c auto_explain.log_min_duration=0
|
||||
|
||||
# Include actual row amounts and run times for query plan nodes
|
||||
-c auto_explain.log_analyze=on
|
||||
|
||||
# Avoid sequential scans so query plans show what index scans can be done
|
||||
# (index scan is normally avoided in some cases, such as the table being small enough)
|
||||
-c enable_seqscan=off
|
||||
|
||||
# Don't log parameter values
|
||||
-c auto_explain.log_parameter_max_length=0
|
||||
)
|
||||
|
||||
# Create cluster
|
||||
initdb --username=postgres --auth=trust --no-instructions
|
||||
|
||||
# Start server that only listens to socket in current directory
|
||||
pg_ctl start
|
||||
pg_ctl start --options="${config_args[*]}"
|
||||
|
||||
# Setup database
|
||||
psql -c "CREATE USER lemmy WITH PASSWORD 'password' SUPERUSER;" -U postgres
|
||||
|
|
Loading…
Reference in a new issue