Extra dim/dimnames attributes in columns of fortify.network output

Hello,

at least in the CRAN version, the data.frame produced by `fortify.network` has `"dim"` and `"dimnames"` attributes set for columns `x, y, xend, yend`. This is an issue for `dplyr::filter`, which complains that the columns are not one-dimensional vectors. A simple turnaround that I use is to first call:
`as.data.frame(lapply(X=df, FUN=as.vector))`
on the output (here named `df`).

Given the increasing popularity of `dplyr` and other `tidyverse` packages, I was wondering if there is a compelling reason to keep these extra attributes in the output?

Thanks a lot for an otherwise excellent package!