Similarity function for embeddings, to use in Supabase

similarity_function.sql

CREATE OR REPLACE FUNCTION public.match_documents(
        query_embedding vector,
        match_threshold double precision,
        match_count integer
    ) RETURNS TABLE(
        id bigint,
        content text,
        similarity double precision,
        metadata jsonb
    ) LANGUAGE sql STABLE AS $function$
select documents.id,
    documents.content,
    1 - (documents.embedding <=> query_embedding) as similarity,
    documents.metadata
from documents
where documents.embedding <=> query_embedding < 1 - match_threshold
order by documents.embedding <=> query_embedding
limit match_count;
$function$;

You should have a table called documents with columns: id, content, metadata, embedding.