Functions for initial module assignment accept a two-dimensional array as input, and return a Modular object. Function for modularity optimization accept a Modular object as input and return this object after the optimization was applied.

• partition_random, attributes all nodes to a module at random (good default, especially for networks with less than 50/50 species)
• partition_lp, uses asynchronous label propagation to estimate the starting partition (good only for large networks)
• partition_single, each node is its own label (good default if you assume a lot of modules)