mirror of
https://github.com/StanGirard/quivr.git
synced 2024-12-04 05:07:48 +03:00
41 lines
1.2 KiB
MySQL
41 lines
1.2 KiB
MySQL
|
create extension if not exists vector;
|
||
|
|
||
|
-- Create a table to store your documents
|
||
|
create table if not exists vectors (
|
||
|
id bigserial primary key,
|
||
|
user_id text, -- new column added here
|
||
|
content text, -- corresponds to Document.pageContent
|
||
|
metadata jsonb, -- corresponds to Document.metadata
|
||
|
embedding vector(1536) -- 1536 works for OpenAI embeddings, change if needed
|
||
|
);
|
||
|
|
||
|
CREATE OR REPLACE FUNCTION match_vectors(query_embedding vector(1536), match_count int, p_user_id text) -- user_id changed to p_user_id here
|
||
|
RETURNS TABLE(
|
||
|
id bigint,
|
||
|
user_id text, -- new column added here
|
||
|
content text,
|
||
|
metadata jsonb,
|
||
|
-- we return matched vectors to enable maximal marginal relevance searches
|
||
|
embedding vector(1536),
|
||
|
similarity float)
|
||
|
LANGUAGE plpgsql
|
||
|
AS $$
|
||
|
# variable_conflict use_column
|
||
|
BEGIN
|
||
|
RETURN query
|
||
|
SELECT
|
||
|
id,
|
||
|
user_id, -- new column added here
|
||
|
content,
|
||
|
metadata,
|
||
|
embedding,
|
||
|
1 -(vectors.embedding <=> query_embedding) AS similarity
|
||
|
FROM
|
||
|
vectors
|
||
|
WHERE vectors.user_id = p_user_id -- filter changed here
|
||
|
ORDER BY
|
||
|
vectors.embedding <=> query_embedding
|
||
|
LIMIT match_count;
|
||
|
END;
|
||
|
$$;
|