Integrate high frequency roaming lines analysis and decide on appliation of TDCF values

Analysis of high frequency roaming lines for systematic error estimation needs to be integrated into the new pyDARM framework.

@jeffrey-kissel has pointed out that there are inconsistent usages of TDCF correction factors between different analyses:

Also note that we're compensating for time-dependence differently when processing the high-frequency roaming lines (only correcting for \kappa_C and f_cc and using front-end computed values instead GDS computed values or sweep MCMC computed values), so whatever we decide should be compatible with correcting these transfer functions too, because they're now folded in to the same estimate of unknown systematic error.