remove surfaces from tests

Author: SimonBlanke

tests/local_test_performance/local_test_global_opt.py

@@ -2,8 +2,6 @@
 from tqdm import tqdm
 import numpy as np
 
-from surfaces.test_functions.mathematical import RastriginFunction
-
 from gradient_free_optimizers import (
     RandomSearchOptimizer,
     RandomRestartHillClimbingOptimizer,
@@ -25,7 +23,11 @@
 def test_global_perf(Optimizer):
     ackley_function = RastriginFunction(n_dim=1, metric="score")
 
-    search_space = {"x0": np.arange(-100, 101, 1)}
+    def objective_function(para):
+        score = -para["x1"] * para["x1"]
+        return score
+
+    search_space = {"x1": np.arange(-100, 101, 1)}
     initialize = {"vertices": 2}
 
     n_opts = 33
@@ -35,7 +37,7 @@ def test_global_perf(Optimizer):
     for rnd_st in tqdm(range(n_opts)):
         opt = Optimizer(search_space, initialize=initialize, random_state=rnd_st)
         opt.search(
-            ackley_function.objective_function,
+            objective_function,
             n_iter=n_iter,
             memory=False,
             verbosity=False,

tests/test_early_stop.py

@@ -105,15 +105,15 @@ def objective_function(para):
     n_iter_no_change = 5
     early_stopping = {
         "n_iter_no_change": 5,
-        "tol_abs": 0.1,
+        "tol_abs": 1,
         "tol_rel": None,
     }
 
     start1 = {"x1": 0}
-    start2 = {"x1": 0.1}
-    start3 = {"x1": 0.2}
-    start4 = {"x1": 0.3}
-    start5 = {"x1": 0.4}
+    start2 = {"x1": 1}
+    start3 = {"x1": 2}
+    start4 = {"x1": 3}
+    start5 = {"x1": 4}
 
     warm_start_l = [
         start1,
@@ -162,13 +162,13 @@ def objective_function(para):
     n_iter_no_change = 5
     early_stopping = {
         "n_iter_no_change": n_iter_no_change,
-        "tol_abs": 0.1,
+        "tol_abs": 10,
         "tol_rel": None,
     }
 
     start1 = {"x1": 0}
-    start2 = {"x1": 0.09}
-    start3 = {"x1": 0.20}
+    start2 = {"x1": 9}
+    start3 = {"x1": 20}
 
     warm_start_l = [
         start1,

tests/test_optimizers/test_opt_algos_simple.py

@@ -1,5 +1,6 @@
 from gradient_free_optimizers.optimizers.core_optimizer import search_tracker
 import pytest
+import numpy as np
 
 from gradient_free_optimizers import (
     HillClimbingOptimizer,
@@ -21,8 +22,6 @@
     ForestOptimizer,
 )
 
-from surfaces.test_functions.mathematical import SphereFunction
-
 optimizers = (
     "Optimizer",
     [
@@ -47,13 +46,20 @@
 )
 
 
-sphere_function = SphereFunction(n_dim=2, metric="score")
+def objective_function(para):
+    score = -para["x1"] * para["x1"]
+    return score
+
+
+search_space = {
+    "x1": np.arange(0, 10, 1),
+}
 
 
 @pytest.mark.parametrize(*optimizers)
 def test_opt_algos_0(Optimizer):
-    opt = Optimizer(sphere_function.search_space())
-    opt.search(sphere_function.objective_function, n_iter=15)
+    opt = Optimizer(search_space)
+    opt.search(objective_function, n_iter=15)
 
     _ = opt.best_para
     _ = opt.best_score

tests/test_optimizers/test_random_state.py

@@ -5,11 +5,8 @@
 
 
 from ._parametrize import optimizers_non_deterministic as optimizers
-from surfaces.test_functions.mathematical import AckleyFunction
 from gradient_free_optimizers import DirectAlgorithm
 
-ackkley_function = AckleyFunction()
-
 
 def objective_function(para):
     score = -(para["x0"] * para["x0"] + para["x1"] * para["x1"])
@@ -33,13 +30,13 @@ def objective_function(para):
 def test_random_state_0(Optimizer):
     opt0 = Optimizer(search_space, initialize={"random": n_random}, random_state=1)
     opt0.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
     opt1 = Optimizer(search_space, initialize={"random": n_random}, random_state=1)
     opt1.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
@@ -56,13 +53,13 @@ def test_random_state_0(Optimizer):
 def test_random_state_1(Optimizer):
     opt0 = Optimizer(search_space, initialize={"random": n_random}, random_state=10)
     opt0.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
     opt1 = Optimizer(search_space, initialize={"random": n_random}, random_state=10)
     opt1.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
@@ -76,13 +73,13 @@ def test_random_state_1(Optimizer):
 def test_random_state_2(Optimizer):
     opt0 = Optimizer(search_space, initialize={"random": n_random}, random_state=1)
     opt0.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
     opt1 = Optimizer(search_space, initialize={"random": n_random}, random_state=10)
     opt1.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
@@ -100,15 +97,15 @@ def test_random_state_direct():
         search_space, initialize={"random": n_random}, random_state=1
     )
     opt0.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
     opt1 = DirectAlgorithm(
         search_space, initialize={"random": n_random}, random_state=10
     )
     opt1.search(
-        ackkley_function.objective_function,
+        objective_function,
         n_iter=n_iter,
     )
 
@@ -124,10 +121,10 @@ def test_random_state_direct():
 @pytest.mark.parametrize(*optimizers)
 def test_no_random_state_0(Optimizer):
     opt0 = Optimizer(search_space, initialize={"random": n_random})
-    opt0.search(ackkley_function.objective_function, n_iter=n_iter)
+    opt0.search(objective_function, n_iter=n_iter)
 
     opt1 = Optimizer(search_space, initialize={"random": n_random})
-    opt1.search(ackkley_function.objective_function, n_iter=n_iter)
+    opt1.search(objective_function, n_iter=n_iter)
 
     print("\n opt0.search_data \n", opt0.search_data)
     print("\n opt1.search_data \n", opt1.search_data)

tests/test_parameters/test_simulated_annealing.py

@@ -5,14 +5,17 @@
 import pytest
 import numpy as np
 
-from surfaces.test_functions.mathematical import SphereFunction
-
 from gradient_free_optimizers import SimulatedAnnealingOptimizer
 
 
-sphere_function = SphereFunction(n_dim=2)
-objective_function = sphere_function.objective_function
-search_space = sphere_function.search_space()
+def objective_function(para):
+    score = -para["x1"] * para["x1"]
+    return score
+
+
+search_space = {
+    "x1": np.arange(0, 10, 1),
+}
 
 
 n_iter = 1000

tests/test_parameters/test_stochastic_hill_climbing.py

@@ -5,17 +5,20 @@
 import pytest
 import numpy as np
 
-from surfaces.test_functions.mathematical import SphereFunction
-
 from gradient_free_optimizers import StochasticHillClimbingOptimizer
 
 
-sphere_function = SphereFunction(n_dim=2)
-objective_function = sphere_function.objective_function
-search_space = sphere_function.search_space()
+n_iter = 1000
 
 
-n_iter = 1000
+def objective_function(para):
+    score = -para["x1"] * para["x1"]
+    return score
+
+
+search_space = {
+    "x1": np.arange(0, 10, 1),
+}
 
 
 def test_p_accept():