Add get_optimal_result logic for objective values

src/everest/everest_storage.py

@@ -20,6 +20,7 @@
 from ropt.plan import BasicOptimizer, Event
 from ropt.results import FunctionResults, GradientResults, convert_to_maximize
 from seba_sqlite import CalculationResult
+from seba_sqlite.sqlite_storage import OptimalResult
 
 from everest.simulator import Simulator
 
@@ -507,6 +508,44 @@ def get_optimal_result(self):
         # then return it
         # print("yo")
 
+        has_merit = "merit_value" in self._dataframes.batches[0].objective.columns
+
+        if has_merit:
+            # Minimize merit
+            sorted_batches = list(self._dataframes.batches)
+            sorted_batches.sort(
+                key=lambda b: b.objective.select(polars.col("merit_value").min()).item()
+            )
+            batch_with_min_merit = sorted_batches[0]
+            print(batch_with_min_merit)
+
+        else:
+            # Maximize objective
+            sorted_batches = list(self._dataframes.batches)
+            sorted_batches.sort(
+                key=lambda b: -b.objective.select(
+                    polars.col("weighted_objective_value").sample(n=1)
+                ).item()
+            )
+            batch = sorted_batches[0]
+
+            return OptimalResult(
+                batch=batch.batch_id,
+                controls={
+                    k: batch.objective[k].sample(1).item()
+                    for k in self._dataframes.initial_values["control_name"]
+                },
+                total_objective=batch.objective.select(
+                    polars.col("weighted_objective_value").sample(n=1)
+                ).item(),
+                expected_objectives={
+                    d["objective_name"]: d["objective_value"]
+                    for d in batch.objective[
+                        ["objective_name", "objective_value"]
+                    ].to_dicts()
+                },
+            )
+
         snapshot = {}  #  SebaSnapshot(self._output_dir)
         optimum = next(
             (