SEMDICE: Off-policy State Entropy Maximization via Stationary Distribution Correction Estimation