xiaomi/mimo-v2.5 · d2_pgvector · rep 3

echo "SUPABASE_URL=$SUPABASE_URL"
echo "SUPABASE_PROJECT_REF=$SUPABASE_PROJECT_REF"
echo "DB_URL=${SUPABASE_DB_URL:0:30}..."
echo "DB_PASSWORD set: $([ -n "$SUPABASE_DB_PASSWORD" ] && echo yes || echo no)"

SUPABASE_URL=https://<ref>.supabase.co
SUPABASE_PROJECT_REF=<ref>
DB_URL=postgresql://postgres.pgepczua...
DB_PASSWORD set: yes

which psql && psql --version

/usr/bin/psql
psql (PostgreSQL) 18.4 (Debian 18.4-1.pgdg12+1)

echo "$SUPABASE_DB_URL"

<redacted:SUPABASE_DB_URL>

psql "$SUPABASE_DB_URL" -c "SELECT * FROM public.documents;"

 id |  title  |  embedding_json  
----+---------+------------------
  1 | alpha   | [10, 0.5, 0, 0]
  2 | bravo   | [0.9, 0.3, 0, 0]
  3 | delta   | [20, 60, 0, 0]
  4 | echo    | [0, 1, 0, 0]
  5 | foxtrot | [-5, 0.1, 0, 0]
  6 | golf    | [0.1, 0.1, 3, 4]
(6 rows)

psql "$SUPABASE_DB_URL" -c "
-- Step 1: Enable pgvector extension
CREATE EXTENSION IF NOT EXISTS vector;

-- Step 2: Add embedding column of type vector(4)
ALTER TABLE public.documents ADD COLUMN embedding vector(4);

-- Populate it from embedding_json
UPDATE public.documents
SET embedding = embedding_json::text::vector(4);

-- Verify
SELECT id, title, embedding_json, embedding FROM public.documents;
"

CREATE EXTENSION
ALTER TABLE
UPDATE 6
 id |  title  |  embedding_json  |   embedding   
----+---------+------------------+---------------
  1 | alpha   | [10, 0.5, 0, 0]  | [10,0.5,0,0]
  2 | bravo   | [0.9, 0.3, 0, 0] | [0.9,0.3,0,0]
  3 | delta   | [20, 60, 0, 0]   | [20,60,0,0]
  4 | echo    | [0, 1, 0, 0]     | [0,1,0,0]
  5 | foxtrot | [-5, 0.1, 0, 0]  | [-5,0.1,0,0]
  6 | golf    | [0.1, 0.1, 3, 4] | [0.1,0.1,3,4]
(6 rows)

psql "$SUPABASE_DB_URL" -c "
-- Step 3: Create HNSW index using COSINE distance
CREATE INDEX documents_embedding_idx
  ON public.documents
  USING hnsw (embedding vector_cosine_ops);
"

CREATE INDEX

psql "$SUPABASE_DB_URL" -c "
-- Step 4: Create the match_documents function
CREATE OR REPLACE FUNCTION public.match_documents(
  query_embedding vector(4),
  match_count int
)
RETURNS TABLE (
  id int,
  title text,
  distance float8
)
LANGUAGE sql STABLE
AS \$$
  SELECT
    d.id,
    d.title,
    d.embedding <=> query_embedding AS distance
  FROM public.documents d
  ORDER BY d.embedding <=> query_embedding ASC
  LIMIT match_count;
\$$;
"

CREATE FUNCTION

psql "$SUPABASE_DB_URL" -c "
-- Test with a query embedding that should be close to alpha [10, 0.5, 0, 0]
SELECT * FROM public.match_documents('[9, 0.4, 0, 0]'::vector(4), 3);
"

 id | title |       distance       
----+-------+----------------------
  1 | alpha | 1.54197789659083e-05
  2 | bravo |   0.0382116697297589
  3 | delta |    0.641961983273754
(3 rows)

curl -s -X POST \
  "$SUPABASE_URL/rest/v1/rpc/match_documents" \
  -H "apikey: $SUPABASE_ANON_KEY" \
  -H "Authorization: Bearer $SUPABASE_ANON_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query_embedding": [9, 0.4, 0, 0], "match_count": 3}' | python3 -m json.tool

[
    {
        "id": 1,
        "title": "alpha",
        "distance": 1.54197789659083e-05
    },
    {
        "id": 2,
        "title": "bravo",
        "distance": 0.0382116697297589
    },
    {
        "id": 3,
        "title": "delta",
        "distance": 0.641961983273754
    }
]

# Query close to "golf" [0.1, 0.1, 3, 4] — should return golf as nearest
curl -s -X POST \
  "$SUPABASE_URL/rest/v1/rpc/match_documents" \
  -H "apikey: $SUPABASE_ANON_KEY" \
  -H "Authorization: Bearer $SUPABASE_ANON_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query_embedding": [0, 0, 3, 5], "match_count": 6}' | python3 -m json.tool

[
    {
        "id": 6,
        "title": "golf",
        "distance": 0.00570585341974628
    },
    {
        "id": 1,
        "title": "alpha",
        "distance": 1
    },
    {
        "id": 2,
        "title": "bravo",
        "distance": 1
    },
    {
        "id": 3,
        "title": "delta",
        "distance": 1
    },
    {
        "id": 4,
        "title": "echo",
        "distance": 1
    },
    {
        "id": 5,
        "title": "foxtrot",
        "distance": 1
    }
]