how to add constraint which act also like a blackbox function without explicit function?

[WeiYUN13]

how to add constraint which act also like a blackbox function without explicit function?
I only see the form like x[:,1] + x[:,2] but no form like a defined function to call in the constraints.

Thank you so much

[ekalosak]

One approach to this might be to add the black-box "constraint" behavior to the objective function: the constraints are explicitly defined, but the objective function already contains the black-box behavior of the optimization target.

To be more concrete, consider the following pseudocode:
bad (notice the function call in the constraint)

def fun(x): 
    return (6*x-2)**2*np.sin(12*x-4)
domain = [{'name': 'var_1', 'type': 'continuous', 'domain': (0,1)}]
constraint = [{'name': 'constr_1', 'constraint': 'fun(x) - 6.0'}]

good

PENALTY_VAL = 0
def fun(x):
    _r = (6*x-2)**2*np.sin(12*x-4)
    r = _r if (_r - 6 > 0) else PENALTY_VAL
    return r
domain = [{'name': 'var_1', 'type': 'continuous', 'domain': (0,1)}]

There must be a rejection mechanism somewhere in the GPyOpt codebase, but I'm not at leisure to search for it right now - please keep us posted as this develops!!