Reimplement the `flamegraph.pl` and `difffolded.pl` in Python

The `flamegraph.pl` and `difffolded.pl` scripts have generally an okay performance, however, creating a flamegraph grid (2x2) for large profiles often consumes too much time and memory. We should re-implement both scripts in native Python and implement some performance optimizations when we already have the profile data pre-processed (as is the case in the `showdiff` command.