juliasmoothoptimizers/nlpmodelsjump.jl

Create NLPModels with JuMP

NLPModelsJuMP

NLPModelsJuMP.jl provides conversion from JuMP.jl / MathOptInterface.jl models to NLPModels.

How to Cite

If you use NLPModelsJuMP.jl in your work, please cite using the format given in CITATION.bib.

Documentation	Linux/macOS/Windows/FreeBSD	Coverage	DOI

See the documentation on NLPModels for the description of its API. Here, we focus on the use of JuMP to create MathOptNLPModel and MathOptNLSModel.

Disclaimer: NLPModelsJuMP is not developed or maintained by the JuMP developers.

Bug reports and discussions

If you think you found a bug, feel free to open an issue. Focused suggestions and requests can also be opened as issues. Before opening a pull request, start an issue or a discussion on the topic, please.

If you want to ask a question not suited for a bug report, feel free to start a discussion here. This forum is for general discussion about this repository and the JuliaSmoothOptimizers organization, so questions about any of our packages are welcome.

nlpmodelsjump.jl's Issues

Consider using multi objective from Jump for NLS

See package MultiObjectiveAlgorithms.jl

MOI Optimizer

I am wondering whether it would be possible to make something like this work

model = Model(NLPModelsJuMP.Optimizer)
set_attribute(model, "solver", percival)
# ...
optimize!(model)

For this, we would need to parse the whole model at the MOI level but looking at the code, it doesn't seem to hard to change this.

• #153
• #155

Let me know what you think.

Track allocations in a MathOptNLP(NLS)Model

I ran the following script tracking allocations:

using Pkg
Pkg.activate(".")
using JuMP, LinearAlgebra, NLPModels, NLPModelsJuMP, NLPModelsTest # main version of NLPModelsTest

nlp_problems = setdiff(NLPModelsTest.nlp_problems, ["MGH01Feas"])
nls_problems = NLPModelsTest.nls_problems
extra_nls_problems = ["HS30", "HS43", "MGH07"]

for problem in lowercase.(nlp_problems)
  include(joinpath("nlp_problems", "$problem.jl"))
end

for problem in lowercase.(nls_problems ∪ extra_nls_problems)
  include(joinpath("nls_problems", "$problem.jl"))
end

for prob in Symbol.(lowercase.(nlp_problems))
  prob_fn = eval(prob)
  nlp = MathOptNLPModel(prob_fn(), hessian = (prob != :nohesspb), name = string(prob))
  print_nlp_allocations(nlp, test_allocs_nlpmodels(nlp))
end

for prob in Symbol.(lowercase.(nls_problems ∪ extra_nls_problems))
  prob_fn = eval(prob)
  nls = prob_fn()
  print_nlp_allocations(nls, test_allocs_nlpmodels(nls))
  print_nlp_allocations(nls, test_allocs_nlsmodels(nls))
end

and got this report:

  Problem name: brownden
                        obj: ████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 16.0  
                hess_coord!: ████████████████████ 96.0  
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0   
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0   
                     hprod!: █████████████████⋅⋅⋅ 80.0  
              hess_op_prod!: █████████████████⋅⋅⋅ 80.0  

  Problem name: hs5
                        obj: ████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 16.0
                hess_coord!: ████████████████████ 96.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: █████████████████⋅⋅⋅ 80.0
              hess_op_prod!: █████████████████⋅⋅⋅ 80.0

  Problem name: hs6
                 hprod_lag!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: ███⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 16.0
     jac_op_transpose_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ███████████████⋅⋅⋅⋅⋅ 80.0
               jac_op_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_lag_coord!: ████████████████████ 112.0
          hess_lag_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ███████████████⋅⋅⋅⋅⋅ 80.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ███████████████⋅⋅⋅⋅⋅ 80.0

  Problem name: hs10
                 hprod_lag!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: ████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 32.0
     jac_op_transpose_prod!: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
            hess_lag_coord!: ████████████████████ 176.0
          hess_lag_op_prod!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: hs11
                 hprod_lag!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: ██████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 32.0
     jac_op_transpose_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ██████████████████⋅⋅ 96.0
            hess_lag_coord!: ████████████████████ 112.0
          hess_lag_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: hs13
                 hprod_lag!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: ███⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 16.0
     jac_op_transpose_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ███████████████⋅⋅⋅⋅⋅ 80.0
               jac_op_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_lag_coord!: ████████████████████ 112.0
          hess_lag_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ███████████████⋅⋅⋅⋅⋅ 80.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ███████████████⋅⋅⋅⋅⋅ 80.0

  Problem name: hs14
                 hprod_lag!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: ████████████████████ 128.0
                     jprod!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 32.0
     jac_op_transpose_prod!: ████████████████████ 128.0
                      cons!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ███████████████⋅⋅⋅⋅⋅ 96.0
            hess_lag_coord!: ██████████████████⋅⋅ 112.0
          hess_lag_op_prod!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: lincon
                 hprod_lag!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     jprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                        obj: ███⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 16.0
     jac_op_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      cons!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                hess_coord!: ███████████████⋅⋅⋅⋅⋅ 80.0
               jac_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_lag_coord!: ████████████████████ 112.0
          hess_lag_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ███████████████⋅⋅⋅⋅⋅ 80.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ███████████████⋅⋅⋅⋅⋅ 80.0

  Problem name: linsv
                 hprod_lag!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                    jtprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     jprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                        obj: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 32.0
     jac_op_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      cons!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      grad!: ████████████████████ 64.0
            hess_lag_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
          hess_lag_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                 jac_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: lls
                 hprod_lag!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                    jtprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     jprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                        obj: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
     jac_op_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      cons!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_lag_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
          hess_lag_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                 jac_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: lls
     hess_op_residual_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
       hess_coord_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
    jac_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            jprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
        jac_coord_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
      jac_op_residual_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
jac_op_residual_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
           jtprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                  residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: mgh01
                        obj: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ████████████████████ 192.0
                      grad!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 96.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 80.0
              hess_op_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 80.0

  Problem name: mgh01
     hess_op_residual_prod!: ███████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
       hess_coord_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
    jac_structure_residual!: ████████████████████ 208.0
            jprod_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
        jac_coord_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
      jac_op_residual_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hprod_residual!: ███████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
jac_op_residual_transpose_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
           jtprod_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                  residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: nlshs20
                 hprod_lag!: ███████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
     jac_op_transpose_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ████████████████████ 208.0
               jac_op_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 96.0
            hess_lag_coord!: █████████████████⋅⋅⋅ 176.0
          hess_lag_op_prod!: ███████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 96.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 96.0

  Problem name: nlshs20
     hess_op_residual_prod!: ███████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
       hess_coord_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
    jac_structure_residual!: ████████████████████ 208.0
            jprod_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
        jac_coord_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
      jac_op_residual_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hprod_residual!: ███████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
jac_op_residual_transpose_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
           jtprod_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                  residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: nlslc
                 hprod_lag!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     jprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                        obj: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
     jac_op_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      cons!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                hess_coord!: ████████████████████ 128.0
               jac_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                      grad!: ███████████████⋅⋅⋅⋅⋅ 96.0
            hess_lag_coord!: ██████████████████⋅⋅ 112.0
          hess_lag_op_prod!: ██████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: █████████████⋅⋅⋅⋅⋅⋅⋅ 80.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: █████████████⋅⋅⋅⋅⋅⋅⋅ 80.0

  Problem name: nlslc
     hess_op_residual_prod!: ████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 176.0
       hess_coord_residual!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
    jac_structure_residual!: ████████████████████ 880.0
            jprod_residual!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
        jac_coord_residual!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
      jac_op_residual_prod!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hprod_residual!: ████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 176.0
jac_op_residual_transpose_prod!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
           jtprod_residual!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                  residual!: ██⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: hs30
                 hprod_lag!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
     jac_op_transpose_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_lag_coord!: ████████████████████ 112.0
          hess_lag_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: hs30
     hess_op_residual_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
       hess_coord_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
    jac_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            jprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
        jac_coord_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
      jac_op_residual_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
jac_op_residual_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
           jtprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                  residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: hs43
                 hprod_lag!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                    jtprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                     jprod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                        obj: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
     jac_op_transpose_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      cons!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
               jac_op_prod!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                      grad!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hess_lag_coord!: ████████████████████ 112.0
          hess_lag_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 64.0
                 jac_coord!: █████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
             jac_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
              hess_op_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: hs43
     hess_op_residual_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
       hess_coord_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
    jac_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            jprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
        jac_coord_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
      jac_op_residual_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
            hprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
jac_op_residual_transpose_prod!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
           jtprod_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                  residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

  Problem name: mgh07
                        obj: ████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                hess_coord!: ████████████████████ 256.0
                      grad!: ████████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 96.0
            hess_structure!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0
                     hprod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 144.0
              hess_op_prod!: ████████████⋅⋅⋅⋅⋅⋅⋅⋅ 144.0

  Problem name: mgh07
     hess_op_residual_prod!: ██████████████⋅⋅⋅⋅⋅⋅ 144.0
       hess_coord_residual!: ██████████████⋅⋅⋅⋅⋅⋅ 144.0
    jac_structure_residual!: ████████████████████ 208.0
            jprod_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
        jac_coord_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
      jac_op_residual_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
            hprod_residual!: ██████████████⋅⋅⋅⋅⋅⋅ 144.0
jac_op_residual_transpose_prod!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
           jtprod_residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
                  residual!: █████⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 48.0
   hess_structure_residual!: ⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅ 0.0

Bug in hessian structure

Oh, I think the problem comes from the model actually. I did the following (a bit nervous) test:

julia> using NLSProblems, NLPModels

julia> nlp = hs06()

julia> hess_structure(nlp)
([1, 267043088], [1, 167116808])

julia> hess_structure(nlp)
([1, 331019536], [1, 269359968])

julia> hess_structure(nlp)
([1, 331019536], [1, 167149984])

julia> hess_structure(nlp)
([1, 331019536], [1, 167150048])

julia> hess_structure(nlp)
([1, 5124065472], [1, 5124065472])

Originally posted by @tmigot in JuliaSmoothOptimizers/CaNNOLeS.jl#43 (comment)

Hessian should be zero/Jacobian coord?

Long post, sorry. I'm having a very specific issue related to JuliaSmoothOptimizers/NLPModels.jl#146 these lines

NLPModelsJuMP.jl/src/nlp_to_mpb.jl

Lines 47 to 56 in 2613241

	function MathProgBase.hesslag_structure(d::NLPModelEvaluator)
	rows, cols, _ = hess_coord(d.nlp, [0.317i for i = 1:d.nlp.meta.nvar],
	y=[0.618i for i = 1:d.nlp.meta.ncon])
	return rows, cols
	end
	function MathProgBase.eval_hesslag(d::NLPModelEvaluator, H, x, σ, μ)
	rows, cols, vals = hess_coord(d.nlp, x, y=μ, obj_weight=σ)
	copyto!(H, vals)
	end

Essentially, my model computes hess_coord with an error that sometimes vanishes. When creating the model, the structure is stored with 4 nonzeros, but during execution hess_coord returns 5 nonzeros. The problem is that the scruture has 4 elements and hess_coord now wants to store 5.

Now with more details. I have the NLSModel mgh04 and the NLS to constrained LS model. The Hessian of F3(x) = x[1] * x[2] - 2 should be [0 1; 1 0], but any of these zeros can turn into an error of order eps().
This can be recreated without NLS:

using JuMP, NLPModels, NLPModelsJuMP
model = Model()
@variable(model, x[1:2])
@NLobjective(model, Min, 0.0)
@NLconstraint(model, x[1] * x[2] == 2)
nlp = MathProgNLPModel(model)
hess_coord(nlp, [0.659; 0.702], y=[1.0])

the output is ([1, 2, 2], [1, 2, 1], [0.0, 1.11022e-16, 1.0]), which means some of the elements that should be 0 is actually close to 1e-16. Now, this doesn't produce an error, because the structure is such that these "null" elements are stored.

However, when I convert a MathProgNLSModel using NLStoCons, it loses that information and returns only the nonzero elements. This becomes a problem when I pass it to NLPtoMPB as explained above.

Now, a possibility of workaround is to change the code for eval_hesslag in nlp_to_mpb.jl and return only the vals for which rows and cols are defined in the structure.

Continuous integration

Set up TravisCI and Appveyor.

Support for JuMP's new Nonlinear interface

It would be great if this supported the new NL interface in JuMP. Right now it seems to ignore all constraints provided in that way.

CC @odow

TagBot trigger issue

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!

Define MathOptNLSModel when Hessian is not available

The following problem

using JuMP                  
using NLPModelsJuMP         
using JSOSolvers             
model = Model()
g(x...) = 10*length(x) + sum(x[i].^2 .- 10*cos.(2π*x[i]) for i in 1:length(x))
register(model, :g, 5, g, autodiff = true)
x₀ = [1.0, 0.1, 0.2, -0.5, 1.0]
@variable(model, x[i=1:5], start = x₀[i])
@NLexpression(model, res[i in 1:5], g(x...))
nls = MathOptNLSModel(model, res, name = "NL")

returns an error LoadError: Unsupported feature Hess. Providing a Hessian through register does not work as JuMP does not support Hessians for multivariate functions. I've seen in #68 that this issue was solved for MathOptNLPModel.

Has this been considered for MathOptNLSModel?

Unsupported features should error, not warn

@ccoffrin was running a few examples and getting incorrect results because quadratic constraints are not supported (#50).

Shouldn't this be an error instead of a warning?

NLPModelsJuMP.jl/src/utils.jl

Lines 216 to 219 in 12d8f23

	for (F, S) in contypes
	F == VI && continue
	F <: AF || @warn("Function $F is not supported.")
	S <: LS || @warn("Set $S is not supported.")

Upcoming refactoring of JuMP's nonlinear API

The upcoming release of JuMP v1.2 will break Complementarity. Read more here: https://discourse.julialang.org/t/ann-upcoming-refactoring-of-jumps-nonlinear-api/83052

The breakage looks pretty minor, and is mostly hacks that were added to get constraint bounds

NLPModelsJuMP.jl/src/moi_nlp_model.jl

Lines 23 to 26 in 2cbf771

	nl_lcon = nnln == 0 ? Float64[] : map(nl_con -> nl_con.lb, jmodel.nlp_data.nlconstr)
	nl_ucon = nnln == 0 ? Float64[] : map(nl_con -> nl_con.ub, jmodel.nlp_data.nlconstr)
	eval = jmodel.nlp_data == nothing ? nothing : NLPEvaluator(jmodel)

x-ref: jump-dev/JuMP.jl#2955

Please ping me if you have questions. I'll try to get time before JuMP 1.2 is released to update this to only use the public API.

You probably need something like

data = MOI.get(model, MOI.NLPBlock())
map(b -> b.lb, block.constraint_bounds)

Add support for quadratic constraints

It could be useful to treat separately quadratic and other nonlinear constraints.
An open question is how to store all hessian of the quadratic constraints?

• List of 2D COO
• 3D COO for third order tensor (rows, cols, tubes, vals)
• ...

Add CITATION.cff

I am reviewing the CITATION usage in the repos and in particular moving to CFF files. The bib can still be generated using Github menu.
Here are a couple of suggestions to adapt this repo:

• use CITATION.cff instead of CITATION.bib ; a good example is https://github.com/JuliaSmoothOptimizers/CaNNOLeS.jl/blob/main/CITATION.cff and/or you can use this tool https://citation-file-format.github.io/cff-initializer-javascript/#/
• update the link in the README.md after that

Repo name

I think it would make more sense for this repository to be named JuMPNLPModels, or simply JuMPModels. Opinions?

Add getter to access variables and constraints with their `JuMP`-name

Use test/nls_consistency.jl from NLPModels.jl

Like test/consistency.jl, we should reuse test/nls_consistency.jl from NLPModels.jl. Currently, we have too much repetition.
May need some modifications in NLPModels.jl.

Can't use views?

julia> using OptimizationProblems
julia> using NLPModelsJuMP
julia> model = MathProgNLPModel(hs6());
julia> xx = rand(5);
julia> x = @view xx[2:3];
julia> obj(model, x)
ERROR: MethodError: no method matching forward_eval(::Array{Float64,1}, ::Array{Float64,1}, ::Array{ReverseDiffSparse.NodeData,1}, ::SparseMatrixCSC{Bool,Int64}, ::Array{Float64,1}, ::Array{Float64,1}, ::SubArray{Float64,1,Array{Float64,1},Tuple{UnitRange{Int64}},true}, ::Array{Float64,1}, ::Array{Float64,1}, ::Array{Float64,1}; user_operators=ReverseDiffSparse.UserOperatorRegistry(Dict{Symbol,Int64}(), MathProgBase.SolverInterface.AbstractNLPEvaluator[], Dict{Symbol,Int64}(), Any[], Any[], Any[]))
Closest candidates are:
  forward_eval(::Array{T,1}, ::AbstractArray{T,1}, ::AbstractArray{ReverseDiffSparse.NodeData,1}, ::Any, ::Any, ::Any, ::Array{T,1}, ::Any, ::Any, ::Any; user_operators) where T at /Users/dpo/.julia/v0.6/ReverseDiffSparse/src/forward.jl:20
  forward_eval(::Array{T,1}, ::AbstractArray{T,1}, ::AbstractArray{ReverseDiffSparse.NodeData,1}, ::Any, ::Any, ::Any, ::Array{T,1}, ::Any, ::Any) where T at /Users/dpo/.julia/v0.6/ReverseDiffSparse/src/forward.jl:20 got unsupported keyword argument "user_operators"
  forward_eval(::Array{T,1}, ::AbstractArray{T,1}, ::AbstractArray{ReverseDiffSparse.NodeData,1}, ::Any, ::Any, ::Any, ::Array{T,1}, ::Any) where T at /Users/dpo/.julia/v0.6/ReverseDiffSparse/src/forward.jl:20 got unsupported keyword argument "user_operators"
Stacktrace:
 [1] forward_eval_all(::JuMP.NLPEvaluator, ::SubArray{Float64,1,Array{Float64,1},Tuple{UnitRange{Int64}},true}) at /Users/dpo/.julia/v0.6/JuMP/src/nlp.jl:445
 [2] eval_f(::JuMP.NLPEvaluator, ::SubArray{Float64,1,Array{Float64,1},Tuple{UnitRange{Int64}},true}) at /Users/dpo/.julia/v0.6/JuMP/src/nlp.jl:477
 [3] obj(::NLPModelsJuMP.MathProgNLPModel, ::SubArray{Float64,1,Array{Float64,1},Tuple{UnitRange{Int64}},true}) at /Users/dpo/.julia/v0.6/NLPModelsJuMP/src/mpb_model.jl:131

MOI optimizer broke

It seems the constructor NLPModelsJuMP.Optimizer() broke with an update of one of the dependencies. When running the tests, I get

Got exception outside of a @test
  MethodError: no method matching SolverCore.GenericExecutionStats{Float64, Vector{Float64}, Vector{Float64}, Any}()
  
  Closest candidates are:
    (::Type{SolverCore.GenericExecutionStats{T, S, V, Tsp}} where {T, S, V, Tsp})(::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any)
     @ SolverCore ~/.julia/packages/SolverCore/iOwNg/src/stats.jl:109
  
  Stacktrace:
    [1] NLPModelsJuMP.Optimizer()
      @ NLPModelsJuMP ~/.julia/dev/NLPModelsJuMP/src/MOI_wrapper.jl:10

The versions used are

(NLPModelsJuMP) pkg> st
Project NLPModelsJuMP v0.12.1
Status `~/.julia/dev/NLPModelsJuMP/Project.toml`
  [4076af6c] JuMP v1.15.1
  [b8f27783] MathOptInterface v1.20.1
  [a4795742] NLPModels v0.20.0
  [01435c0c] Percival v0.7.0
  [ff4d7338] SolverCore v0.3.7
  [37e2e46d] LinearAlgebra
  [de0858da] Printf
  [2f01184e] SparseArrays

`Function MathOptInterface.ScalarQuadraticFunction{Float64} is not supported.`

I am suspecting this is related to #50, there is this warning in OptimizationProblems.jl and the resulting NLPModel has no constraints, while the JuMP model seems to be ok.

using OptimizationProblems
using NLPModelsJuMP

prob = :hs61
prob_fn = eval(Meta.parse("OptimizationProblems.PureJuMP.$(prob)"))
nlp_jump = MathOptNLPModel(prob_fn())
nlp_jump.meta.ncon # is == 0

Add documentation about the order of variables and constraints

compare to JuMP.all_variables

Update NLPModelsJuMP to use MathOptInterface

Improve hess_coord! when y is not provided

JuliaSmoothOptimizers/ADNLPModels.jl#130

Function tracing (JuMP 1.15.0) is not supported

A short, not minimal, example:

n = 2
jmp = Model()
σ(t) = 1 / (1 + exp(-t))
@variable(jmp, x[1:n])
@variable(jmp, y[1:n])
@objective(jmp, Min,
	sum(σ(x[i] - 1)^2 for i = 1:n) + sum(y[i]^2 for i = 1:n)
)
@constraint(jmp, [i=1:n,j=i+1:n], x[i] + x[j] == y[i] - y[j])

nlp = MathOptNLPModel(jmp)

Raises

MethodError: Cannot `convert` an object of type

MathOptInterface.ScalarNonlinearFunction to an object of type

Union{MathOptInterface.VariableIndex, MathOptInterface.ScalarAffineFunction{Float64}, MathOptInterface.ScalarQuadraticFunction{Float64}}

Closest candidates are:

convert(::Type{T}, !Matched::T) where T

@ Base Base.jl:64

get(::MathOptInterface.Utilities.ObjectiveContainer{Float64}, ::MathOptInterface.ObjectiveFunction{Union{MathOptInterface.VariableIndex, MathOptInterface.ScalarAffineFunction{Float64}, MathOptInterface.ScalarQuadraticFunction{Float64}}})@objective_container.jl:130
[email protected]:294[inlined]
get(::MathOptInterface.Utilities.UniversalFallback{MathOptInterface.Utilities.Model{Float64}}, ::MathOptInterface.ObjectiveFunction{Union{MathOptInterface.VariableIndex, MathOptInterface.ScalarAffineFunction{Float64}, MathOptInterface.ScalarQuadraticFunction{Float64}}})@universalfallback.jl:550
[email protected]:869[inlined]
parser_objective_MOI(::MathOptInterface.Utilities.CachingOptimizer{MathOptInterface.AbstractOptimizer, MathOptInterface.Utilities.UniversalFallback{MathOptInterface.Utilities.Model{Float64}}}, ::Int64)@utils.jl:328
var"#MathOptNLPModel#15"(::Bool, ::String, ::Type{NLPModelsJuMP.MathOptNLPModel}, ::JuMP.Model)@moi_nlp_model.jl:31
NLPModelsJuMP.MathOptNLPModel(::JuMP.Model)@moi_nlp_model.jl:19
top-level scope@[Local: 12](http://localhost:1234/edit?id=20cefd64-57ee-11ee-0b4e-f19cdb29ea67#)

Release notes of JuMP 1.15.0: https://jump.dev/blog/1.15.0-release/

Quadratic functions in the constraints are not picked up by MathOptNLPModel

More specifically,

model = Model()
@variable(model, x[1:2])
@objective(model, Min, sum(x .^ 4))
@constraint(model, sum(x .^ 2) == 1)
nlp = MathOptNLPModel(model)

warns with

┌ Warning: Function MathOptInterface.ScalarQuadraticFunction{Float64} is not supported.
└ @ NLPModelsJuMP ~/.julia/packages/NLPModelsJuMP/1KSlI/src/utils.jl:219

And indeed, there are no constraints in the model. Solving the problem via JuMP using NLPModels.Optimizer works.

cc. @blegat, if you can help us again.

Limit NLPModels version to 0.8.1

Linear / quadratic models

If a JuMP model m has linear constraints + linear or quadratic objective, we cannot use NLPEvalaluator(m).constraints because it's an undefined reference. We can detect if we have non linear objective with eval.has_nlobj but I don't know if we have something similar for non linear constraints...

Edit :num_nl_constraints(m) works.

SingleVariable constraints

It's possible to encouter SingleVariable constraints if we define in the same time the variable and the constraint.

m = Model()
@variable(m, t >= 1)
@variable(m, s == 2)

I should remove this line https://github.com/JuliaSmoothOptimizers/NLPModelsJuMP.jl/blob/master/src/utils.jl#L165 and add a function parser_SV.

Combine this package with SparseMatricesCOO

We could probably simplify a few things using https://github.com/JuliaSmoothOptimizers/SparseMatricesCOO.jl

• SparseMatrixCOO instead of COO

  NLPModelsJuMP.jl/src/utils.jl

  Line 28 in c05a30f

  mutable struct COO

• Move coo_sym_dot

  NLPModelsJuMP.jl/src/utils.jl

  Line 74 in c05a30f

  """

  to https://github.com/JuliaSmoothOptimizers/SparseMatricesCOO.jl/blob/main/src/coo_linalg.jl
• Move coo_sym_add_mul!

  NLPModelsJuMP.jl/src/utils.jl

  Line 74 in c05a30f

  """

  to https://github.com/JuliaSmoothOptimizers/SparseMatricesCOO.jl/blob/main/src/coo_linalg.jl

Define MathOptNLPModel when hessian is not available

Trying to convert the following problem in an NLPModel:

using JuMP, NLPModelsJuMP

  model = Model()
  @variable(model, x[1:2])
  g(x::T, y::T) where {T<:Real} = x * y
  function ∇g(g::Vector{T}, x::T, y::T) where {T<:Real}
      g[1] = y
      g[2] = x
      return
  end
  register(model, :g, 2, g, ∇g)
  @NLobjective(model, Min, g(x[1], x[2]))
  nlp = MathOptNLPModel(model)

will return Unsupported feature Hess. From what I understand in MOI, when using register in the model, the hessian is not available. Hence, :Hess and :HessVec are not available for such problems in MOI.initialize.

Any suggestion to handle this case?
Maybe add a try/catch and then assume that the user won't use hess-related function.

Improve MOI optimizer

A todo list to improve the connection with MOI (#152 ):

• increase coverage
• specify feasible point primal #157 (comment)
• specify dual feasible #157 (comment)
• add solver name
• add documentation
• add tutorial
• use SolverCore default solver instead of loading Percival in the tests

cc' @blegat @amontoison

Increase coverage in NLS model

In the nls model the following are apparently not covered by the unit tests:

• NLPModelsJuMP.jl/src/moi_nls_model.jl

  Line 373 in f842b9b

  function NLPModels.hess_structure!(

• NLPModelsJuMP.jl/src/moi_nls_model.jl

  Line 395 in f842b9b

  function NLPModels.hess_coord!(

• NLPModelsJuMP.jl/src/moi_nls_model.jl

  Line 418 in f842b9b

  function NLPModels.hess_coord!(

Retroactively add an upper bound to JuMP to older versions of NLPModelsJuMP

This bound:

JuMP = "1.2 - 1.14.1"

I'm looking into this.