This is an unofficial implementation in python3.6 for the paper: Facetnet: a framework for analyzing communities and their evolutions in dynamic networks
main.py: entrance of the program, containing 3 experiments on synthetic networks, can be run directly.
util.py: helper funtions of the program, including the soft modularity stated in the paper.
synthetic.py: codes for creating the synthetic networks, including the network mentioned in part 4.1.2.
Algorithm for both fixed and dynamic node numbers are implemented, but the one with varied community numbers is not, since I find it insufficient to implement with the information in the paper. Equation (9) and (10) would fail if the number of communities increases.