Improvements to polarized data reduction in Cosmos
Reported by Thomas Saerbeck on 6th June 2018.
The function cosmos_anal_group_pol is not used!
The integer peak position of the reflected beam can be slightly different in the numors corresponding to the different polarizabilities. This creates a shift when using the coherent method, as the pixel used as the reference to do the 'Q-matching' is not the same. Need to do a similar thing as what it was already done to store the values of Q, dQ, and wavelength and other parameters using the 1st numor as the reference and then employing the same values for the other numors.
It can be tested using table exp_5-54-244/e-fe2o3reduction_win and comparing lines 58, 61 and 7.