uniform_mix_llike_wrapper: Wrapper for 'uniform_mix_llike', mostly for the SQUAREM...
In dcgerard/vicar: Various Ideas for Confounder Adjustment in Regression

This returns the negative log-likelihood for use in SQUAREM.

uniform_mix_llike_wrapper(
  pizxi_vec,
  betahat_ols,
  S_diag,
  alpha_tilde,
  a_seq,
  b_seq,
  lambda_seq,
  degrees_freedom,
  scale_var = TRUE
)

`pizxi_vec`	A vector of numerics. The first M of which are `pi_vals`, the next k of which are `z2`, the last element is `xi`.
`betahat_ols`	A vector of numerics. The OLS regression coefficients.
`S_diag`	A vector of positive numerics. The standard errors.
`alpha_tilde`	A matrix of numerics. The estimates of the coefficients of the confounders.
`a_seq`	A vector of numerics (usually non-positive). The lower bounds of the uniform mixing densities.
`b_seq`	A vector of numerics (usually non-negative). The upper bounds of the uniform mixing densities.
`lambda_seq`	A vector of numerics greater than 1. The penalties on the mixing proportions.
`degrees_freedom`	A positive scalar or a vector of positive scalars the length of `betahat_ols`. The degrees of freedom of the t-distribution.
`scale_var`	A logical. Should we update the variance inflation parameter (`TRUE`) or not (`FALSE`)