mig_beta child adjustments

[sarahertog]

Child-woman adjustment requires female pop as input, but this input is not available to mig_beta() when estimating residual net mig for males. Current implementation adjusts boys to net mig of adult males and girls to net mig of adult females. Can yield implausible sex ratios of child net migration, such as when total net mig reflects male labor pattern. Similar issue with constant adjustment -- current implementation is sex specific. I have modified our use of this function for WPP to first use mig_beta() with no child adjustment, and then in a subsequent step adjust that output using adult female pop and net mig (for cwr) and both sexes pop and net mig (for constant).