Revisiting CounteRGAN for Counterfactual Explainability of Graphs