EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition