In the latest release of igraph, version 1.2.11, the test suite broke due to a change in the cluster_louvain function which now has a stochastic implementation (details in PR https://github.com/igraph/igraph/pull/1696). Hence, the modularity is no longer necessarily exactly 0.4432351, but some value that depends on the random number generator. In order to get reproducible results, so that the test suite doesn't fail, you should seed the RNG prior to running cluster_louvain using set.seed.
In the latest release of
igraph
, version 1.2.11, the test suite broke due to a change in thecluster_louvain
function which now has a stochastic implementation (details in PR https://github.com/igraph/igraph/pull/1696). Hence, the modularity is no longer necessarily exactly0.4432351
, but some value that depends on the random number generator. In order to get reproducible results, so that the test suite doesn't fail, you should seed the RNG prior to runningcluster_louvain
usingset.seed
.