Update channel_combination.py

[laurajanem]

• Added module RAMRS_SVD_weighting (Receive Array MRS SVD weighting, from the reference in the comments). Performs SVD channel combination on the raw data (prior to averaging). Useful in cases where there might be drift across the averages, which can be corrected after channel combination more easily, since channel combination increases the SNR of the averages.

• Added a few additional outputs to the svd_weighting module to also return just the channel weights (without phase correction). This is useful for cases where channel weights are derived from the unsuppressed H2O signal, and then applied to the suppressed H2O signal. The phase shift term won't be correct for the suppressed H2O signal.

Note: I know the updates to the svd_weighting algorithm will break the way it is used in the channel combination Jupyter notebook, though, so we can discuss alternative ways to implement this if it is too cumbersome. It would be helpful to be able to return the channel weights, minus the phase correction term, though.