Skip to content

ENG-1769 Optimize semantic search by removing LIMIT from function body #1141

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
mdroidian wants to merge 3 commits into
base: main
Choose a base branch
from
+45 −15
Open

ENG-1769 Optimize semantic search by removing LIMIT from function body #1141

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2fa802f
Optimize semantic search by removing LIMIT from function body [ENG-1769]
cursoragent Jun 22, 2026
68cd04e
Remove space_id parameter - moved to separate ticket
cursoragent Jun 22, 2026
a0b28aa
drop if exists
mdroidian Jun 22, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 6 additions & 5 deletions apps/roam/src/utils/discourseNodeSearchProviders.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -512,11 +512,12 @@ const runSupabaseSemanticSearch = async ({
}

const queryEmbedding = await createEmbedding(trimmedQuery);
const { data, error } = await supabase.rpc("match_content_embeddings", {
query_embedding: JSON.stringify(queryEmbedding),
match_threshold: SUPABASE_MATCH_THRESHOLD,
match_count: SEARCH_TEST_RESULT_LIMIT,
});
const { data, error } = await supabase
.rpc("match_content_embeddings", {
query_embedding: JSON.stringify(queryEmbedding),
match_threshold: SUPABASE_MATCH_THRESHOLD,
})
.limit(SEARCH_TEST_RESULT_LIMIT);

if (error) {
throw new Error(error.message);
Expand Down
11 changes: 6 additions & 5 deletions apps/roam/src/utils/hyde.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -538,11 +538,12 @@ export const findSimilarNodesVectorOnly = async ({

const queryEmbedding = await createEmbedding(text);

const { data, error } = await supabase.rpc("match_content_embeddings", {
query_embedding: JSON.stringify(queryEmbedding),
match_threshold: threshold,
match_count: limit,
});
const { data, error } = await supabase
.rpc("match_content_embeddings", {
query_embedding: JSON.stringify(queryEmbedding),
match_threshold: threshold,
})
.limit(limit);

if (error) {
console.error("Vector search failed:", error);
Expand Down
1 change: 0 additions & 1 deletion packages/database/src/dbTypes.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1722,7 +1722,6 @@ export type Database = {
match_content_embeddings: {
Args: {
current_document_id?: number
match_count: number
match_threshold: number
query_embedding: string
}
Expand Down
32 changes: 32 additions & 0 deletions packages/database/supabase/migrations/20260622040252_optimize_match_content_embeddings.sql
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,32 @@
-- Optimize match_content_embeddings by removing LIMIT from function body
-- This improves query planner performance as the LIMIT parameter was killing the planner

set search_path to public, extensions ;

DROP FUNCTION IF EXISTS public.match_content_embeddings(extensions.vector, double precision, integer, integer) ;

CREATE OR REPLACE FUNCTION public.match_content_embeddings (
Comment thread
mdroidian marked this conversation as resolved.
query_embedding extensions.vector,
match_threshold double precision,
current_document_id integer DEFAULT NULL::integer)
RETURNS TABLE (
content_id bigint,
roam_uid Text,
text_content Text,
similarity double precision)
SET search_path = 'extensions'
LANGUAGE sql STABLE
AS $$
SELECT
c.id AS content_id,
c.source_local_id AS roam_uid,
c.text AS text_content,
1 - (c.vector <=> query_embedding) AS similarity
FROM public.my_contents_with_embedding_openai_text_embedding_3_small_1536 AS c
WHERE 1 - (c.vector <=> query_embedding) > match_threshold
AND (current_document_id IS NULL OR c.document_id = current_document_id)
ORDER BY
c.vector <=> query_embedding ASC;
$$ ;
Comment thread
mdroidian marked this conversation as resolved.

RESET ALL ;
5 changes: 1 addition & 4 deletions packages/database/supabase/schemas/embedding.sql
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,7 +57,6 @@ set search_path to public, extensions ;
CREATE OR REPLACE FUNCTION public.match_content_embeddings (
query_embedding extensions.vector,
match_threshold double precision,
match_count integer,
current_document_id integer DEFAULT NULL::integer)
RETURNS TABLE (
content_id bigint,
Expand All @@ -76,14 +75,12 @@ FROM public.my_contents_with_embedding_openai_text_embedding_3_small_1536 AS c
WHERE 1 - (c.vector <=> query_embedding) > match_threshold
AND (current_document_id IS NULL OR c.document_id = current_document_id)
ORDER BY
c.vector <=> query_embedding ASC
LIMIT match_count;
c.vector <=> query_embedding ASC;
$$ ;

ALTER FUNCTION public.match_content_embeddings (
query_embedding extensions.vector,
match_threshold double precision,
match_count integer,
current_document_id integer) OWNER TO "postgres" ;

CREATE OR REPLACE FUNCTION public.match_embeddings_for_subset_nodes (
Expand Down