mmm interesting. This comment made me realize a way to add the effect of the mute lists on the random walks used for computing pagarank, personalized Pagerank and more.

It's quite simple actually:

The random walk starts from `v`, which has a `mute_list`.

At each step, a `new node` is selected at random from the set:

`out_neighbours(current_node) - mute_list`

This means that the mute list of the starting node has priority over the follows of any other node in the walk.