A linear constraint with two linear variables still raises an issue if any non-linear variables are present in the system.

[ronald-jaepel]

So

optimization_problem.add_variable('var_0', lb=-2, ub=2, transform="linear")
optimization_problem.add_variable('var_1', lb=-2, ub=2, transform="linear")
optimization_problem.add_variable('var_2', lb=0, ub=2, transform="log")

optimization_problem.add_linear_constraint(['var_0', 'var_1'], [-1, -0.5], 0)

raises

CADETProcessError: Non-linear transform was used in linear constraints.

Is this intended?

The reason is this section in

    @property
    def A_transformed(self):
        """np.ndarray: LHS Matrix of linear inequality constraints in transformed space.

        See Also
        --------
        A
        A_independent_transformed
        A_independent

        """
        A_t = self.A.copy()
        for a in A_t:
            for j, v in enumerate(self.variables):
                t = v.transform
                if isinstance(t, NoTransform):
                    continue

                if not t.is_linear:
                    raise CADETProcessError(
                        "Non-linear transform was used in linear constraints."
                    )

Which raises an error if any variable is not linear as long as there is a inequality constraint.

If this is a bug, my proposed solution would be to change this to

    @property
    def A_transformed(self):
        """np.ndarray: LHS Matrix of linear inequality constraints in transformed space.

        See Also
        --------
        A
        A_independent_transformed
        A_independent

        """
        A_t = self.A.copy()
        for a in A_t:
            for j, v in enumerate(self.variables):
                t = v.transform
                if isinstance(t, NoTransform):
                    continue

                if a[j] != 0 and not t.is_linear:
                    raise CADETProcessError(
                        "Non-linear transform was used in linear constraints."
                    )

So that only variables with a non-zero influence in the inequality constraint are checked.
I've pushed this to PR #116

[ronald-jaepel]

Closed as solved with #116