handle missing data in viz's modules

[ahmadtourei]

It'd be great to add support for visualizing a concatenated patch out of a spool with missing patches in between. After merging #375, we can use spool.concatenate(time=None) to merge all patches in the spool, even if there are missing patches.. It'd be useful to visualize the resultant concatenated patch with missing data. Ideally, we do not want to fill the array with zeroes or NaNs beforehand and fill up memory if we can handle it in viz's modules. Also, we probably do not want to alter the axis labels of the dimension where patches are missing (e.g., time when there was on/off periods in IU) or eliminate the gap.

[ahmadtourei]

Current Patch.viz.waterfall() behavior: visualizes concatenated patch that has data gaps without the gaps and with wrong time axis which gets extended from the first to the last time sample with no gaps in between.