post_process changes, and model closes again.

[1ozturkbe]

now tests fail because of some other issue @bqpd, when solving using mosek_conif. Here is what I see:

Solving took 0.333 seconds.
Traceback (most recent call last):
  File "SPaircraft.py", line 97, in <module>
    sol = test()
  File "SPaircraft.py", line 89, in test
    sol = optimize_aircraft(m, substitutions, fixedBPR, pRatOpt, mutategparg)
  File "SPaircraft.py", line 65, in optimize_aircraft
    sol = m_relax.localsolve(solver='mosek_conif', verbosity=2, iteration_limit=200, reltol=0.01, mutategp=mutategparg)#, x0 = x0)
  File "c:\users\berk\dropbox (mit)\mit graduate school\code\gpkit\gpkit\constraints\prog_factories.py", line 124, in solvefn
    result = progsolve(solver, verbosity, **kwargs)
  File "c:\users\berk\dropbox (mit)\mit graduate school\code\gpkit\gpkit\constraints\sgp.py", line 169, in localsolve
    self.result = gp.generate_result(solver_out, verbosity)
  File "c:\users\berk\dropbox (mit)\mit graduate school\code\gpkit\gpkit\constraints\gp.py", line 211, in generate_result
    result = self._compile_result(solver_out)  # NOTE: SIDE EFFECTS
  File "c:\users\berk\dropbox (mit)\mit graduate school\code\gpkit\gpkit\constraints\gp.py", line 298, in _compile_result
    % cost.varkeys.keys())
ValueError: cost contains unsolved variables dict_keys([Mission.FlightState.Relax.\gamma[3,0], Mission.FlightState.Relax.\gamma[4,0], Mission.FlightState.Atmosphere.Relax.C_1[0,0], Mission.FlightState.Atmosphere.Relax.p_{sl}[2,0], Mission.FlightState.Relax.R[1,0], Mission.FlightState.Atmosphere.Relax.M_{atm}[3,0], Mission.FlightState.Atmosphere.Relax.C_1[4,0], Mission.FlightState.Atmosphere.Relax.R_{atm}[2,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{t}}[3,0], Mission.FlightState.Atmosphere.Relax.M_{atm}[4,0], Mission.FlightSegment.FlightP.AircraftP.Relax.W_{Load_{max}}[2,0], Mission.FlightState.Atmosphere.Relax.T_{sl}[3,0], Mission.FlightSegment.FlightP.Relax.RC_{min}[0,0], Mission.FlightState.Atmosphere.Relax.p_{sl}[4,0], Mission.Aircraft.Engine.EnginePerformance.CombustorPerformance.Relax.M_{4a}[0,0], Mission.FlightSegment.FlightP.AircraftP.Relax.V_{stall}[2,0], Mission.FlightSegment.FlightP.AircraftP.Relax.W_{Load_{max}}[4,0], Mission.FlightState.Atmosphere.Relax.T_{sl}[2,0], Mission.FlightSegment.FlightP.AircraftP.Relax.V_{stall}[3,0], Mission.Aircraft.Engine.EnginePerformance.CombustorPerformance.Relax.M_{4a}[4,0], Mission.FlightState.Atmosphere.Relax.M_{atm}[1,0], Mission.FlightState.Atmosphere.Relax.T_s[1,0], Mission.FlightState.Atmosphere.Relax.R_{atm}[4,0], Mission.Aircraft.Engine.EnginePerformance.CombustorPerformance.Relax.M_{4a}[2,0], Mission.FlightState.Atmosphere.Relax.L_{atm}[3,0], Mission.FlightState.Atmosphere.Relax.T_s[0,0], Mission.FlightState.Atmosphere.Relax.T_s[4,0], Mission.FlightState.Atmosphere.Relax.C_1[2,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{t}}[0,0], Mission.FlightSegment.FlightP.AircraftP.Relax.V_{stall}[0,0], Mission.FlightSegment.FlightP.AircraftP.Relax.W_{Load_{max}}[0,0], Mission.FlightState.Atmosphere.Relax.p_{sl}[3,0], Mission.FlightState.Atmosphere.Relax.C_1[3,0], Mission.FlightState.Atmosphere.Relax.L_{atm}[2,0], Mission.FlightSegment.FlightP.Relax.RC_{min}[4,0], Mission.FlightState.Relax.R[2,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{o}}[1,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{o}}[2,0], Mission.FlightState.Atmosphere.Relax.R_{atm}[3,0], Mission.FlightState.Atmosphere.Relax.R_{atm}[0,0], Mission.FlightSegment.FlightP.Relax.RC_{min}[3,0], Mission.FlightState.Atmosphere.Relax.T_{sl}[1,0], Mission.FlightSegment.FlightP.AircraftP.Relax.W_{Load_{max}}[1,0], Mission.FlightState.Atmosphere.Relax.T_s[3,0], Mission.FlightState.Relax.R[3,0], Mission.FlightState.Atmosphere.Relax.L_{atm}[1,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{o}}[0,0], Mission.Aircraft.Engine.EnginePerformance.CombustorPerformance.Relax.M_{4a}[1,0], Mission.FlightState.Atmosphere.Relax.L_{atm}[0,0], Mission.FlightSegment.FlightP.Relax.RC_{min}[1,0], Mission.FlightState.Atmosphere.Relax.M_{atm}[2,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{t}}[1,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{o}}[4,0], Mission.FlightState.Relax.R[0,0], Mission.FlightState.Atmosphere.Relax.R_{atm}[1,0], Mission.FlightState.Atmosphere.Relax.M_{atm}[0,0], Mission.FlightSegment.FlightP.AircraftP.Relax.W_{Load_{max}}[3,0], Mission.FlightSegment.FlightP.AircraftP.Relax.V_{stall}[4,0], Mission.FlightSegment.FlightP.Relax.RC_{min}[2,0], Mission.FlightState.Atmosphere.Relax.T_{sl}[4,0], Mission.FlightState.Atmosphere.Relax.p_{sl}[0,0], Mission.FlightState.Atmosphere.Relax.C_1[1,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{t}}[2,0], Mission.FlightSegment.FlightP.AircraftP.Relax.V_{stall}[1,0], Mission.FlightState.Relax.\gamma[1,0], Mission.FlightState.Relax.\gamma[0,0], Mission.FlightState.Relax.R[4,0], Mission.FlightState.Atmosphere.Relax.L_{atm}[4,0], Mission.FlightState.Atmosphere.Relax.T_s[2,0], Mission.FlightState.Relax.\gamma[2,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{t}}[4,0], Mission.FlightState.Atmosphere.Relax.p_{sl}[1,0], Mission.FlightSegment.FlightP.AircraftP.WingPerformance.Relax.f_{L_{o}}[3,0], Mission.FlightState.Atmosphere.Relax.T_{sl}[0,0], Mission.Aircraft.Engine.EnginePerformance.CombustorPerformance.Relax.M_{4a}[3,0], Mission.Aircraft.Fuselage.Relax.n_{pass}[0]])

unsolved variables? Haven't seen this one before.

[bqpd]

huh, why are variables in the cost that are not in constraints?

[1ozturkbe]

@bqpd for some reason GPkit does weird things sometimes. With other solvers it does not throw this error. Those slack variables are actually in the models.

[bqpd]

I don't think it's GPkit rolling the dice just for kicks; the error indicates that not all free variables in the cost are in the results.

[1ozturkbe]

Well, I can't figure out why that occurs with mosek_conif and not mosek_cli... There is zero difference in the formulation.

[1ozturkbe]

Alright, model now closes again on the working machines (all but Windows). Turns out this was an issue with substitutions in turbofan somehow falling through when it was integrated into SPaircraft.