LST axis is incorrect in the waterfall plot when there is a discontinuity in time

[hughbg]

Discontinuities in spectrometer data over time can be caused by restarting the spectrometer or correlator. A single h5 file may contain a gap in time. If it does, then the waterfall plot must indicate this using the y-axis. Currently the assumption is that there is a continuous range of LSTs from the first one to the last one, which means the y-axis is incorrect in the case of a discontinuity..

[hughbg]

There is a way to deal with this - add the missing LSTs, filling them with 0s, and masking them. This would be for display purposes in plot_waterfall, and in plot_ds9 these could be put in the FITs file.

[jaycedowell]

Most of the time when I have seen a discontinuity it is from the day rolling over (23:56:04 -> 0:00:00) rather then from gaps in the data. What about "unrolling" the LSTs before sorting and then protecting any function that takes in an LST with a modulo one day?