I may be wrong but to me it seems that the tests in /tests are not testing the correctness of results but merely if they compile and run without error.
Am I wrong ?
It would be good to have a few unit test for correctness.
Note that I already wrote some test for bipartite graph as part of my python wrapping testing.
https://github.com/geeklhem/pybip/blob/master/pybip/tests/test_modularity.py
It test module identity on the southern women dataset and role values on simpler graph were I could do the computation by hand.