[Bug]: parameter_constraints does not support RangeParameterConfig

[AlbanMor]

What happened?

Trying:
client.configure_experiment(parameters=[x1,x2,x3,x4,x5], parameter_constraints=[upper_constraint, lower_constraint])

I get the following error:

in validate_constraint_parameters(parameters)
for parameter in parameters:
if not isinstance(parameter, RangeParameter):
raise ValueError(
"All parameters in a parameter constraint must be RangeParameters."
)
Log parameters require a non-linear transformation, and Ax
models only support linear constraints.
if isinstance(parameter, RangeParameter) and parameter.log_scale is True:

ValueError: All parameters in a parameter constraint must be RangeParameters.

I believe this is due to the fact that in Ax 1.2.1 we use RangeParameterConfig and not RangeParameter

Please provide a minimal, reproducible example of the unexpected behavior.

X1 = RangeParameterConfig(name="X1", parameter_type="float", bounds=(80, 175))
X2 = RangeParameterConfig(name="X2", parameter_type="float", bounds=(9,12),step_size=1)
X3 = RangeParameterConfig(name="X3", parameter_type="float", bounds=(600, 800))
X4 = RangeParameterConfig(name="X4", parameter_type="float", bounds=(180, 196))
X5 = RangeParameterConfig(name="X5", parameter_type="float", bounds=(1300, 1620), step_size=20)

upper_constraint = 'X1 - 25.667X2 <= - 133'
lower_constraint = 'X1 - 20X2 >= - 100'

#Initialize Client
client = Client()
#Configure Experiment and Optimization
client.configure_experiment(parameters=[X1, X2, X3, X4, X5],
parameter_constraints=[upper_constraint, lower_constraint])

Please paste any relevant traceback/logs produced by the example provided.

Ax Version

1.2.1

Python Version

3.13

Operating System

MS

(Optional) Describe any potential fixes you've considered to the issue outlined above.

No response

Pull Request

None

Code of Conduct

• I agree to follow Ax's Code of Conduct