Fix handling of instance maximum for constrained optimization

climada/util/calibrate/bayesian_optimizer.py

@@ -190,11 +190,11 @@
 class StopEarly(Exception):
     """An exception for stopping an optimization iteration early"""
 
     pass
 
 
 @dataclass(eq=False)
 class BayesianOptimizerController(object):
     """A class for controlling the iterations of a :py:class:`BayesianOptimizer`.
 
     Each iteration in the optimizer consists of a random sampling of the parameter space
@@ -291,7 +291,7 @@
             return -np.inf
         return self._improvements[-1].target
 
     def is_converged(self) -> bool:
         """Check if convergence criteria are met"""
         return True
 
@@ -329,14 +329,16 @@
 
     def _is_new_max(self, instance):
         """Determine if a guessed value is the new maximum"""
-        if instance.max is None:
+        instance_max = instance.max
+        if not instance_max or instance_max.get("target") is None:
             # During constrained optimization, there might not be a maximum
             # value since the optimizer might've not encountered any points
             # that fulfill the constraints.
             return False
 
-        if instance.max["target"] > self._previous_max:
+        if instance_max["target"] > self._previous_max:
             return True
+
         return False
 
     def _maybe_stop_early(self, instance):
@@ -499,7 +501,7 @@
         """Invert the cost function because BayesianOptimization maximizes the target"""
         return -self.input.cost_func(data, predicted)
 
     def run(
         self,
         controller: BayesianOptimizerController,
     ) -> BayesianOptimizerOutput:
@@ -527,7 +529,7 @@
 
         while controller.iterations < controller.max_iterations:
             try:
                 LOGGER.info(f"Optimization iteration: {controller.iterations}")
                 self.optimizer.maximize(**controller.optimizer_params())
             except StopEarly:
                 continue
@@ -570,7 +572,7 @@
 
         super().__post_init__()
 
     def plot_impf_variability(
         self,
         cost_func_diff: float = 0.1,
         p_space_df: Optional[pd.DataFrame] = None,
@@ -621,7 +623,7 @@
 
         # Set plot defaults
         color = plot_impf_kws.pop("color", "tab:blue")
         lw = plot_impf_kws.pop("lw", 2)
         zorder = plot_impf_kws.pop("zorder", 3)
         label = plot_impf_kws.pop("label", "best fit")
 
@@ -630,7 +632,7 @@
         axes = []
 
         for impf_idx in range(n_impf):
             _, ax = plt.subplots()
 
             # Plot best-fit impact function
             best_impf = self.impf_set.get_func(haz_type=haz_type)[impf_idx]