Show less digits for compare

[aloctavodia]

We should have some rounding rules for compare similar to those we already have for summary. @sethaxen do you have any suggestions?

[sethaxen]

@sethaxen do you have any suggestions?

I haven't checked yet if compare here includes any additional columns to what we use in PosteriorStats, but there we use the same rules for elpd[_diff] and se_elpd[_diff] ans for p and se_p as we do for <col> and se_<col> in summarize. For weights, we use this rule:
https://github.com/arviz-devs/PosteriorStats.jl/blob/ea08a287bfa86c5e4561758dc0f2a4a3b4b72178/src/compare.jl#L145C22-L145C61

This is how many digits it shows after the decimal point

julia> f(wmax) = ceil(Int, -log10(wmax)) + 1;

julia> map(x -> x => f(x), [0.09, 0.1, 0.5, 0.99, 1.0])
5-element Vector{Pair{Float64, Int64}}:
 0.09 => 3
  0.1 => 2
  0.5 => 2
 0.99 => 2
  1.0 => 1

So it basically ensures that we show 2 significant digits for the largest weight. Any weights that are 1/20th of this weight or less would be rounded to 0.

e.g.

julia> using ArviZExampleData, PosteriorStats

julia> models = (
           centered=load_example_data("centered_eight"),
           non_centered=load_example_data("non_centered_eight"),
       );

julia> mc = compare(models)
┌ Warning: 1 parameters had Pareto shape values 0.7 < k ≤ 1. Resulting importance sampling estimates are likely to be unstable.
└ @ PSIS ~/.julia/packages/PSIS/...
ModelComparisonResult with Stacking weights
               rank  elpd  se_elpd  elpd_diff  se_elpd_diff  weight    p  se_p
 non_centered     1   -31      1.5       0            0.0       1.0  0.9  0.32
 centered         2   -31      1.4       0.03         0.061     0.0  0.9  0.33