Initial attempt at running surrogate assisted optimization

.gitignore

@@ -153,4 +153,7 @@ reference_solutions.json
 
 
 # Scratch
-scratch/
+scratch/
+
+# Surrogate models
+surrogatemodels/

desdeo/emo/operators/evaluator.py

@@ -4,7 +4,7 @@
 
 import polars as pl
 
-from desdeo.problem import PolarsEvaluator, Problem
+from desdeo.problem import Evaluator, Problem
 from desdeo.tools.message import (
     EvaluatorMessageTopics,
     GenericMessage,
@@ -51,7 +51,7 @@ def __init__(
         super().__init__(**kwargs)
         self.problem = problem
         # TODO(@light-weaver, @gialmisi): This can be so much more efficient.
-        self.evaluator = lambda x: PolarsEvaluator(problem)._polars_evaluate_flat(
+        self.evaluator = lambda x: Evaluator(problem).evaluate(
             {name.symbol: x[name.symbol].to_list() for name in problem.get_flattened_variables()}
         )
         self.variable_symbols = [name.symbol for name in problem.variables]

desdeo/problem/simulator_evaluator.py

@@ -9,7 +9,7 @@
 import joblib
 import numpy as np
 import polars as pl
-#import skops.io as sio
+# import skops.io as sio
 
 from desdeo.problem import (
     ObjectiveTypeEnum,
@@ -22,13 +22,12 @@
 class EvaluatorError(Exception):
     """Error raised when exceptions are encountered in an Evaluator."""
 
+
 class Evaluator:
     """A class for creating evaluators for simulator based and surrogate based objectives, constraints and extras."""
+
     def __init__(
-        self,
-        problem: Problem,
-        params: dict[str, dict] | None = None,
-        surrogate_paths: dict[str, Path] | None = None
+        self, problem: Problem, params: dict[str, dict] | None = None, surrogate_paths: dict[str, Path] | None = None
     ):
         """Creating an evaluator for simulator based and surrogate based objectives, constraints and extras.
 
@@ -45,24 +44,20 @@ def __init__(
         self.problem = problem
         # store the symbol and min or max multiplier as well (symbol, min/max multiplier [1 | -1])
         self.objective_mix_max_mult = [
-            (objective.symbol, -1 if objective.maximize else 1)
-            for objective in problem.objectives
+            (objective.symbol, -1 if objective.maximize else 1) for objective in problem.objectives
         ]
         # Gather symbols objectives of different types into their own lists
         self.analytical_symbols = [
-            obj.symbol for obj in list(
-                filter(lambda x: x.objective_type == ObjectiveTypeEnum.analytical, problem.objectives)
-            )
+            obj.symbol
+            for obj in list(filter(lambda x: x.objective_type == ObjectiveTypeEnum.analytical, problem.objectives))
         ]
         self.simulator_symbols = [
-            obj.symbol for obj in list(
-                filter(lambda x: x.objective_type == ObjectiveTypeEnum.simulator, problem.objectives)
-            )
+            obj.symbol
+            for obj in list(filter(lambda x: x.objective_type == ObjectiveTypeEnum.simulator, problem.objectives))
         ]
         self.surrogate_symbols = [
-            obj.symbol for obj in list(
-                filter(lambda x: x.objective_type == ObjectiveTypeEnum.surrogate, problem.objectives)
-            )
+            obj.symbol
+            for obj in list(filter(lambda x: x.objective_type == ObjectiveTypeEnum.surrogate, problem.objectives))
         ]
         # Gather any constraints' symbols
         if problem.constraints is not None:
@@ -138,9 +133,7 @@ def _evaluate_simulator(self, xs: dict[str, list[int | float]]) -> pl.DataFrame:
             params = self.params.get(sim.name, {})
             # call the simulator with the decision variable values and parameters as dicts
             res = subprocess.run(
-                [sys.executable, sim.file, '-d', str(xs), '-p', str(params)],
-                capture_output=True,
-                text=True
+                [sys.executable, sim.file, "-d", str(xs), "-p", str(params)], capture_output=True, text=True
             )
             if res.returncode == 0:
                 # gather the simulation results (a dict) into the results dataframe
@@ -172,7 +165,7 @@ def _evaluate_surrogates(self, xs: dict[str, list[int | float]]) -> pl.DataFrame
                 uncertainty predictions, then they are set as NaN.
         """
         res = pl.DataFrame()
-        var = np.array([value for _, value in xs.items()]).T # has to be transpose (at least for sklearn models)
+        var = np.array([value for _, value in xs.items()]).T  # has to be transpose (at least for sklearn models)
         for symbol in self.surrogates:
             # get a list of args accepted by the model's predict function
             accepted_args = getfullargspec(self.surrogates[symbol].predict).args
@@ -234,7 +227,7 @@ def _load_surrogates(self, surrogate_paths: dict[str, Path] | None = None):
                         else: # TODO: if there are unknown types they should be checked
                             self.surrogates[obj.symbol] = sio.load(file, unknown_types)
                             #raise EvaluatorError(f"Untrusted types found in the model of {obj.symbol}: {unknown_types}")"""
-            for con in self.problem.constraints:
+            for con in self.problem.constraints or []:  # if there are no constraints, an empty list is used
                 if con.surrogates is not None:
                     with Path.open(f"{con.surrogates[0]}", "rb") as file:
                         self.surrogates[con.symbol] = joblib.load(file)
@@ -244,7 +237,7 @@ def _load_surrogates(self, surrogate_paths: dict[str, Path] | None = None):
                         else: # TODO: if there are unknown types they should be checked
                             self.surrogates[con.symbol] = sio.load(file, unknown_types)
                             #raise EvaluatorError(f"Untrusted types found in the model of {obj.symbol}: {unknown_types}")"""
-            for extra in self.problem.extra_funcs:
+            for extra in self.problem.extra_funcs or []:  # if there are no extra functions, an empty list is used
                 if extra.surrogates is not None:
                     with Path.open(f"{extra.surrogates[0]}", "rb") as file:
                         self.surrogates[extra.symbol] = joblib.load(file)