Restrict ipopt to only accept minimization problem

Author: metab0t

include/pyoptinterface/ipopt_model.hpp

@@ -106,8 +106,16 @@ struct IpoptModel
 	}
 
 	template <typename T>
-	void set_objective(const T &expr, bool clear_nl = false)
+	void set_objective(const T &expr, ObjectiveSense sense = ObjectiveSense::Minimize,
+	                   bool clear_nl = false)
 	{
+		if (sense != ObjectiveSense::Minimize)
+		{
+			throw std::runtime_error(
+			    "Currently Ipopt only supports ObjectiveSense::Minimize, please negate the "
+			    "objective manually if you intend to maximize the objective");
+		}
+
 		m_lq_model.set_objective(expr);
 		if (clear_nl)
 		{

lib/ipopt_model_ext.cpp

@@ -131,15 +131,15 @@ NB_MODULE(ipopt_model_ext, m)
 	    .def("add_objective", &IpoptModel::add_objective<double>)
 
 	    .def("set_objective", &IpoptModel::set_objective<ExprBuilder>, nb::arg("expr"),
-	         nb::arg("clear_nl") = false)
+	         nb::arg("sense") = ObjectiveSense::Minimize, nb::arg("clear_nl") = false)
 	    .def("set_objective", &IpoptModel::set_objective<ScalarQuadraticFunction>, nb::arg("expr"),
-	         nb::arg("clear_nl") = false)
+	         nb::arg("sense") = ObjectiveSense::Minimize, nb::arg("clear_nl") = false)
 	    .def("set_objective", &IpoptModel::set_objective<ScalarAffineFunction>, nb::arg("expr"),
-	         nb::arg("clear_nl") = false)
+	         nb::arg("sense") = ObjectiveSense::Minimize, nb::arg("clear_nl") = false)
 	    .def("set_objective", &IpoptModel::set_objective<VariableIndex>, nb::arg("expr"),
-	         nb::arg("clear_nl") = false)
+	         nb::arg("sense") = ObjectiveSense::Minimize, nb::arg("clear_nl") = false)
 	    .def("set_objective", &IpoptModel::set_objective<double>, nb::arg("expr"),
-	         nb::arg("clear_nl") = false)
+	         nb::arg("sense") = ObjectiveSense::Minimize, nb::arg("clear_nl") = false)
 
 	    .def("_add_nl_objective", &IpoptModel::_add_nl_objective)