I have the following DGP:
"Data generating process (DGP): Assume the true data generating process is the following: x is distributed iid N(0,1). eps_i (the i–th epsilon) is distributed N(0,1+exp(x_i)). That is, the variance of epsilon is (1+exp(x)) for that observation. Then assume that the LHS variable Y is generated according to: y = 2 + 1 * x + epsilon . Make the number of observations (N) in the dataset be 50."

But now, I need to estimate it with FGLS: "go back to the d.g.p. from problem (1a) and see how the properties of the FGLS (Feasible Generalized Least Squares) estimator goes. Figure out how to estimate FGLS in Stata, and describe your method (both in “econometrics” language and in “stata” language). In order to implement FGLS you’ll need a statistical model for the heteroscedasiticy. Use this model: var(eps_i) = gamma_0 + gamma_1 * exp(x_i). Then show how (1) the estimated slope parameter and"

Can anyone provide guidance on how to do this?
Thanks!