Phase relative weighting for fitting

Hi Lee, Jenne asked me to do this, so here it is. Takes the SNR_phase_rel option in data2filter and passes this variable into residuals_log_im_scale. Not much work was required to get residuals_log_im_scale to be given as an array. Attached are examples of fitting with and without the phase weighting.

I've noticed that setting a portion of SNR_phase_rel to zeroes leads to the fitting often return a flat filter. Something to be looked into maybe...

Anyway, attached are two fits using the simple4 test data, but I modified the phase behaviour at higher frequencies. Let me know if any changes are required.

SNR_phase_rel = 1e-6 above 1.5 Hz: testing_external_with_1d5Hz_phase_cut.pdf

SNR_phase_rel = 1: testing_external.pdf