Bring split TIV calculation in line with gulpy

[hchagani-oasislmf]

Bring split TIV calculation in line with gulpy

With gulcalc allocation rule greater than 0, if the sum of losses per sample exceed the TIV, the TIV is split in proportion to the losses. There is no need to do this for the standard deviation (sidx = -2) nor for the chance of loss (sidx = -4). Therefore, the calculation on these fields has been dropped to bring it in line with that which is currently conducted in gulpy.