add variance module

[zhaotianjing]

wiki page: https://github.com/reworkhow/JWAS.jl/wiki/Constraint-on-variance-components

1. change estimateScale to estimate_scale;
2. delete mega_trait because it equals adding constraint on variances
3. move constraint argument in runMCMC() to build_model(). This is to put all residual variance related argument together.
4. add variance structural, and modify related code. add printed info if constraint; add error if constraint in ST

   mutable struct Variance
       val::Union{AbstractFloat, AbstractArray, Bool}   #value of the variance, e.g., single-trait: 1.5, two-trait: [1.3 0.4; 0.4 0.8]
       df::Union{AbstractFloat,Bool}                    #degrees of freedom, e.g., 4.0
       scale::Union{AbstractFloat, AbstractArray, Bool} #scale, e.g., single-trait: 1.0, two-trait: [1.0 0; 0 1.0]
   
       estimate_variance::Bool  #estimate_variance=true means estimate variance at each MCMC iteration
       estimate_scale::Bool     #estimate_scale=true means estimate scale at each MCMC iteration
       constraint::Bool        #constraint=true means in multi-trait analysis, covariance is zero
   end
   

5. make scale in MT be a diagonal matrix everywhere
6. organize arguments in build_model(), get_genotypes(), set_random ()
7. relocate some comments in NNMM
8. relocate initialization of scale_R, df_R to build_model()
9. add R_constraint!() and G_constraint!() functions to change scale and df if constraint