Continuous Time q-learning for McKean-Vlasov Control Problems