It would be nice to have an option for byte count instead of humanised sizes in the output of `pacgraph -c` making counting easier with tools like awk. Post-processing in a pipeline could easily humanise the values.