Incorrect type assumption with int()

[sbromberger]

function lmatrix(g::Graph)
    n_v = nv(g)
    A = int(adjacency_matrix(g))   # note that this creates a sparse matrix of integers
    D = spdiagm(sum(A,2)[:])
    return D - A
end

yields

ERROR  sum: ErrorException("no method found for the specified argument types")
                                         Signature: (Int64,Int64)

Also curiously, it doesn't warn that I'm not using n_v.