real-ish graph: 2000 nodes, 40000 edges
  in-degree  max=  600 mean=  20.00
  out-degree max=  594 mean=  20.00

degree-preserving null
  n_nodes_match                True
  n_edges_match                True
  in_degree_exact              True
  out_degree_exact             True
  self_loops                   0
  multi_edges                  0
  edges_retained               3369
  fraction_rewired             0.915775
  weight_multiset_preserved    True

Erdos-Renyi null (the weak control)
  n_nodes_match                True
  n_edges_match                True
  in_degree_exact              False
  out_degree_exact             False
  self_loops                   0
  multi_edges                  0
  edges_retained               360
  fraction_rewired             0.991
  weight_multiset_preserved    True

statistic              real    weak null    strict null
reciprocity          0.0104       0.0100     0.0110+-0.0007
clustering           0.0622       0.0201     0.0596+-0.0006
mean_path_ish        2.2717       2.4242     2.2610+-0.0012

clustering gap vs weak null   : 0.0420
clustering gap vs strict null : 0.0025
reproducibility: same seed identical, different seed differs -> ok

all checks passed
