Performance issue in Dulmage-Mendelsohn

`dulmage_mendelsohn.jl` uses the following:
```julia
filter = cat(vec1, vec2, ...; dims = 1)
other = [n for n in nodes if !(n in filter)]
```
As `filter` is a vector, this is quadratic time and will become slow if `nodes`, the vector of nodes in the bipartite graph, becomes large. As we do expect this set of nodes to become large, we should implement `filter` as a set.