Gurobi interface gives incorrect primal for constraints

[cdiener]

In the following model (zipped JSON) you have a constraint "test" whose primal will have the wrong sign after optimizing.

File: bad.zip

Basically you get stuff like this when having constraints with only a lower bound:

In [95]: mod.solver.optimize()
'optimal'

In [96]: print(mod.constraints.test)
test: 0.0 <= -1.0*Biomass_Ecoli_core_reverse_2cdba + 1.0*Biomass_Ecoli_core

In [97]: mod.constraints.test.primal
Out[97]: -0.8739215069684309

I think the problem is that optlang returns the slack variable from the associated optlang interface, however those change their sign depending on the type of constraint (negative for >= and positive for <=)...

[cdiener]

Actually, it never gives the correct primals...

For instance:

In [1]: import optlang

In [2]: gurobi = optlang.gurobi_interface

In [3]: mod = gurobi.Model()
Academic license - for non-commercial use only

In [4]: x = gurobi.Variable(lb=3, ub=3, name="x")

In [5]: const = gurobi.Constraint(x, ub=5)

In [6]: mod.add([x, const])

In [7]: mod.optimize()
Out[7]: 'optimal'

In [8]: const.primal
Out[8]: 2.0

In [9]: x.primal
Out[9]: 3.0

In [10]: const.lb = 3

In [11]: const.ub = 3

In [12]: mod.optimize()
Out[12]: 'optimal'

In [13]: const.primal
Out[13]: 0.0

In both cases the correct primal for the constraint is 3, but optlang gives 2 and 0.