A note on how to identify redundant constraints using JuMP

Specific Domains Optimization (Mathematical)
,
3

I only intend to remove trivial constraints, mainly those artificial bounds added to ensure the compactness of the feasible region.
e.g., The primal problem is \qquad \min_{y, x} y \ge |x|_1
My code for this might be

# the following 4 lines are the primal problem
JuMP.@variable(primal, y)
JuMP.@variable(primal, x[1:2])
JuMP.@constraint(primal, [y; x] in JuMP.MOI.NormOneCone(3))
JuMP.@objective(primal, Min, y)
# To make the feasible region compact, we artificially add these
JuMP.set_lower_bound.(y, -3); JuMP.set_upper_bound.(y, 3); 
JuMP.set_lower_bound.(x, -4); JuMP.set_upper_bound.(x, 4);
# optimize WITH artificial bounds
JuMP.optimize!(primal); JuMP.assert_is_solved_and_feasible(primal; dual = true, allow_local = false)
S = [JuMP.dual(JuMP.LowerBoundRef(y))    JuMP.dual(JuMP.UpperBoundRef(y));
    JuMP.dual.(JuMP.LowerBoundRef.(x))  JuMP.dual.(JuMP.UpperBoundRef.(x))]
@assert all(S .== 0)
# then we can delete all artificial bounds and reoptimize to retain the same optimal objective_value