General Post-Processing Framework for Fairness Adjustment of Machine Learning Models