Incorrect waterfall indices in dedispersion operation

[0xCoto]

There appears to be somewhat of an inaccurate application of the incoherent dedispersion formula, yielding this error:

Traceback (most recent call last):
  File "quick_obs.py", line 34, in <module>
    obs_file='observation.dat', plot_file='plot.png')
  File "/home/apostolos/.local/lib/python3.5/site-packages/virgo/virgo.py", line 565, in plot
    shift(t_bin, n)
  File "/home/apostolos/.local/lib/python3.5/site-packages/virgo/virgo.py", line 450, in shift
    waterfall[:, phase_num] = np.roll(waterfall[:, phase_num], -n_rows)
IndexError: index 256 is out of bounds for axis 1 with size 256

I've only noticed this error just now with a long observation. I suspect there is a mistakenly flipped order (i.e. mixing waterfall indices, ultimately trying to shift channels instead of time bins).

Lines 449-450 and 557-565 are suspicious, but I couldn't figure out a solution from a quick look. Furhter investigation is necessary.