query — nearest neighbors

query() is the bread-and-butter call. Vector in, ranked hits out.

hits = notes.query(
    query_embeddings=[0.1, 0.2, 0.3, 0.4],
    n_results=5,
    max_distance=0.3,
)
for h in hits:
    print(h.key, h.distance, h.meta)

From docs/cpp/api-reference/collection.md: each Hit carries key, distance, meta, document, vector — all hydrated from the authoritative record store.