Preprocess: Iterative Glitch + Jump Detection + HWPSS Estimation

[msilvafe]

We want to extend our processing to include at least one iteration of HWPSS estimation between glitch + jump detection. The idea being that the template estimation will be improved once glitches are flagged and filled and smaller glitches are more easily caught after the HWPSS is subtracted. So 1) flag the biggest glitches/jumps 2) fill them 3) estimate + subtract hwpss 4) flag + fill glitches and jumps again 5) add back in hwpss template 6) re-estimate hwpss template with the.

Currently we're not doing this because 1) we expect the HWPSS template bias is fairly small and 2) the overall processing time is slow so this was an obvious cut for now until we speed things up as raised in issue #981