plot_quality() works reasonably well for single sources but fails for larger set

[mdingemanse]

Running plot_quality() for eyeballing turn durations and transitions works as expected for single sources:

Reprex:

devtools::install_github("elpaco-escience/ifadv")
data <- init(ifadv::ifadv)

plot_quality(data, source = "/dutch2/DVA9M")

But for a full dataset consisting of multiple source it doesn't produce sensible results:

plot_quality(data)

I think what we get here is that transition times are computed across sources, resulting in some extremely long (nonextant) transitions.

Another reason I would much rather work with proper FTO calculations as proposed in #70.

[mdingemanse]

To be fixed in branch 75-fix-qual-plots