JuMP.NLPEvaluator output format



I am trying to evaluate my constraints from my nonlinear program using IPOPT. From this discussion (https://groups.google.com/forum/#!searchin/julia-opt/evaluating$20constraints%7Csort:date/julia-opt/m1ObRwgrvGE/Mo8bjsZ1AwAJ), I am trying to use MathProgBase.eval_g. The first line in calling this is

d = JuMP.NLPEvaluator(m)

I cannot find resources that describe how to interpret the output of this function call. First it lists the number of constraints and variables. After that, it shows a lot of numbers, #undef, and true/false statements–this is the part I am confused about. Is there a detailed description or examples somewhere of interpreting this?

The versions I am using are: Julia v0.6.4 and JuMP v0.18.2.


Please see http://www.juliaopt.org/JuMP.jl/0.18/nlp.html#querying-derivatives-from-a-jump-model. JuMP.NLPEvaluator() returns an opaque object that implements an interface (described in the JuMP and MathProgBase docs). There’s not much to be gained by trying to look inside the object.