FairPy: A Toolkit for Evaluation of Social Biases and their Mitigation in Large Language Models