How to pass options to objective in cyipopt

I'm trying to pass options to a NLP I'm solving in cyipopt.

These options would for affect the objective in the same way in each iteration. For example, the tutorial problem is to minimize

x_1 * x_4 * (x_1 + x_2 + _3) + x_3

subject to some constraints (see https://pythonhosted.org/ipopt/tutorial.html).

I'd like to solve the related problem

scale * x_1 * x_4 * (x_1 + x_2 + _3) + x_3

where scale is a parameter that is set before optimization. The below code shows how to set up the problem in pyipopt, but the scale is hardcoded as 2. How can I set it as an option so that it can be changed flexibly?

import ipopt
import numpy as np

class hs071(object):
    def __init__(self):
        pass

    def objective(self, x, scale):
        # The callback for calculating the objective
        scale = 2
        
        return scale * x[0] * x[3] * np.sum(x[0:3]) + x[2]

    def gradient(self, x, scale):
        # The callback for calculating the gradient
        scale = 2
        
        return np.array([
                    scale * x[0] * x[3] + scale * x[3] * np.sum(x[0:3]),
                    scale * x[0] * x[3],
                    scale * x[0] * x[3] + 1.0,
                    scale * x[0] * np.sum(x[0:3])
                    ])

    def constraints(self, x):
        # The callback for calculating the constraints
        return np.array((np.prod(x), np.dot(x, x)))

    def jacobian(self, x):
        # The callback for calculating the Jacobian
        return np.concatenate((np.prod(x) / x, 2*x))

x0 = [1.0, 5.0, 5.0, 1.0]

lb = [1.0, 1.0, 1.0, 1.0]
ub = [5.0, 5.0, 5.0, 5.0]

cl = [25.0, 40.0]
cu = [2.0e19, 40.0]

nlp = ipopt.problem(
            n=len(x0),
            m=len(cl),
            problem_obj=hs071(),
            lb=lb,
            ub=ub,
            cl=cl,
            cu=cu
            )

x, info = nlp.solve(x0)

NB: defining globals works but is sloppy. There must be a cleaner way to do this, since this is the way that you would add data to an optimization problem.

Add them to the class itself:

import ipopt
import numpy as np

class hs071(object):
    def __init__(self):
        pass

    def objective(self, x):
        # The callback for calculating the objective
        scale = self.scale
        
        return scale * x[0] * x[3] * np.sum(x[0:3]) + x[2]

    def gradient(self, x):
        # The callback for calculating the gradient
        scale = self.scale
        
        return np.array([
                    scale * x[0] * x[3] + scale * x[3] * np.sum(x[0:3]),
                    scale * x[0] * x[3],
                    scale * x[0] * x[3] + 1.0,
                    scale * x[0] * np.sum(x[0:3])
                    ])

    def constraints(self, x):
        # The callback for calculating the constraints
        return np.array((np.prod(x), np.dot(x, x)))

    def jacobian(self, x):
        # The callback for calculating the Jacobian
        return np.concatenate((np.prod(x) / x, 2*x))

x0 = [1.0, 5.0, 5.0, 1.0]

lb = [1.0, 1.0, 1.0, 1.0]
ub = [5.0, 5.0, 5.0, 5.0]

cl = [25.0, 40.0]
cu = [2.0e19, 40.0]

model = hs071()
model.scale = 2

nlp = ipopt.problem(
            n=len(x0),
            m=len(cl),
            problem_obj=model,
            lb=lb,
            ub=ub,
            cl=cl,
            cu=cu
            )

x, info = nlp.solve(x0)