Avoid import of whole util module

fjwillemsen · fjwillemsen · commit a48394a33e68 · 2025-03-20T18:54:05.000+01:00
diff --git a/kernel_tuner/strategies/basinhopping.py b/kernel_tuner/strategies/basinhopping.py
@@ -1,7 +1,7 @@
 """The strategy that uses the basinhopping global optimization method."""
 import scipy.optimize
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc, setup_method_arguments, setup_method_options
@@ -31,7 +31,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     try:
         opt_result = scipy.optimize.basinhopping(cost_func, x0, T=T, stepsize=eps,
                                              minimizer_kwargs=minimizer_kwargs, disp=tuning_options.verbose)
-    except util.StopCriterionReached as e:
+    except StopCriterionReached as e:
         if tuning_options.verbose:
             print(e)
 
diff --git a/kernel_tuner/strategies/bayes_opt.py b/kernel_tuner/strategies/bayes_opt.py
@@ -149,7 +149,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
         if max_fevals - bo.fevals <= 0:
             raise ValueError("No function evaluations left for optimization after sampling")
         bo.optimize(max_fevals)
-    except util.StopCriterionReached as e:
+    except StopCriterionReached as e:
         if tuning_options.verbose:
             print(e)
 
diff --git a/kernel_tuner/strategies/diff_evo.py b/kernel_tuner/strategies/diff_evo.py
@@ -1,7 +1,7 @@
 """The differential evolution strategy that optimizes the search through the parameter space."""
 from scipy.optimize import differential_evolution
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc
@@ -32,7 +32,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     try:
         opt_result = differential_evolution(cost_func, bounds, maxiter=maxiter, popsize=popsize, init=population_enc,
                                         polish=False, strategy=method, disp=tuning_options.verbose)
-    except util.StopCriterionReached as e:
+    except StopCriterionReached as e:
         if tuning_options.verbose:
             print(e)
 
diff --git a/kernel_tuner/strategies/dual_annealing.py b/kernel_tuner/strategies/dual_annealing.py
@@ -1,7 +1,7 @@
 """The strategy that uses the dual annealing optimization method."""
 import scipy.optimize
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc, setup_method_arguments, setup_method_options
@@ -31,7 +31,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     opt_result = None
     try:
         opt_result = scipy.optimize.dual_annealing(cost_func, bounds, minimizer_kwargs=minimizer_kwargs, x0=x0, maxfun=max_fevals)
-    except util.StopCriterionReached as e:
+    except StopCriterionReached as e:
         if tuning_options.verbose:
             print(e)
 
diff --git a/kernel_tuner/strategies/firefly_algorithm.py b/kernel_tuner/strategies/firefly_algorithm.py
@@ -3,7 +3,7 @@
 
 import numpy as np
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc, scale_from_params
@@ -42,7 +42,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     for j in range(num_particles):
         try:
             swarm[j].compute_intensity(cost_func)
-        except util.StopCriterionReached as e:
+        except StopCriterionReached as e:
             if tuning_options.verbose:
                 print(e)
             return cost_func.results
@@ -65,7 +65,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
                     swarm[i].move_towards(swarm[j], beta, alpha)
                     try:
                         swarm[i].compute_intensity(cost_func)
-                    except util.StopCriterionReached as e:
+                    except StopCriterionReached as e:
                         if tuning_options.verbose:
                             print(e)
                         return cost_func.results
diff --git a/kernel_tuner/strategies/genetic_algorithm.py b/kernel_tuner/strategies/genetic_algorithm.py
@@ -4,7 +4,7 @@
 
 import numpy as np
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached, get_best_config
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc
@@ -42,7 +42,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
         for dna in population:
             try:
                 time = cost_func(dna, check_restrictions=False)
-            except util.StopCriterionReached as e:
+            except StopCriterionReached as e:
                 if tuning_options.verbose:
                     print(e)
                 return cost_func.results
@@ -54,7 +54,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
 
         # 'best_score' is used only for printing
         if tuning_options.verbose and cost_func.results:
-            best_score = util.get_best_config(
+            best_score = get_best_config(
                 cost_func.results, tuning_options.objective, tuning_options.objective_higher_is_better
             )[tuning_options.objective]
 
diff --git a/kernel_tuner/strategies/greedy_ils.py b/kernel_tuner/strategies/greedy_ils.py
@@ -1,5 +1,5 @@
 """A simple greedy iterative local search algorithm for parameter search."""
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc
@@ -40,7 +40,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
         try:
             candidate = base_hillclimb(candidate, neighbor, max_fevals, searchspace, tuning_options, cost_func, restart=restart, randomize=True)
             new_score = cost_func(candidate, check_restrictions=False)
-        except util.StopCriterionReached as e:
+        except StopCriterionReached as e:
             if tuning_options.verbose:
                 print(e)
             return cost_func.results
diff --git a/kernel_tuner/strategies/greedy_mls.py b/kernel_tuner/strategies/greedy_mls.py
@@ -1,5 +1,5 @@
 """A greedy multi-start local search algorithm for parameter search."""
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.hillclimbers import base_hillclimb
@@ -30,7 +30,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
 
         try:
             base_hillclimb(candidate, neighbor, max_fevals, searchspace, tuning_options, cost_func, restart=restart, randomize=randomize, order=order)
-        except util.StopCriterionReached as e:
+        except StopCriterionReached as e:
             if tuning_options.verbose:
                 print(e)
             return cost_func.results
diff --git a/kernel_tuner/strategies/minimize.py b/kernel_tuner/strategies/minimize.py
@@ -2,7 +2,7 @@
 
 import scipy.optimize
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies.common import (
     CostFunc,
@@ -30,7 +30,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     opt_result = None
     try:
         opt_result = scipy.optimize.minimize(cost_func, x0, method=method, options=options, **kwargs)
-    except util.StopCriterionReached as e:
+    except StopCriterionReached as e:
         if tuning_options.verbose:
             print(e)
 
diff --git a/kernel_tuner/strategies/pso.py b/kernel_tuner/strategies/pso.py
@@ -5,7 +5,7 @@
 
 import numpy as np
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc, scale_from_params
@@ -52,7 +52,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
         for j in range(num_particles):
             try:
                 swarm[j].evaluate(cost_func)
-            except util.StopCriterionReached as e:
+            except StopCriterionReached as e:
                 if tuning_options.verbose:
                     print(e)
                 return cost_func.results
diff --git a/kernel_tuner/strategies/random_sample.py b/kernel_tuner/strategies/random_sample.py
@@ -1,7 +1,7 @@
 """Iterate over a random sample of the parameter space."""
 import numpy as np
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc
@@ -26,7 +26,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
     for sample in samples:
         try:
             cost_func(sample, check_restrictions=False)
-        except util.StopCriterionReached as e:
+        except StopCriterionReached as e:
             if tuning_options.verbose:
                 print(e)
             return cost_func.results
diff --git a/kernel_tuner/strategies/simulated_annealing.py b/kernel_tuner/strategies/simulated_annealing.py
@@ -4,7 +4,7 @@
 
 import numpy as np
 
-from kernel_tuner import util
+from kernel_tuner.util import StopCriterionReached
 from kernel_tuner.searchspace import Searchspace
 from kernel_tuner.strategies import common
 from kernel_tuner.strategies.common import CostFunc
@@ -52,7 +52,7 @@ def tune(searchspace: Searchspace, runner, tuning_options):
             new_pos = neighbor(pos, searchspace)
             try:
                 new_cost = cost_func(new_pos, check_restrictions=False)
-            except util.StopCriterionReached as e:
+            except StopCriterionReached as e:
                 if tuning_options.verbose:
                     print(e)
                 return cost_func.results
diff --git a/test/strategies/test_strategies.py b/test/strategies/test_strategies.py
@@ -4,7 +4,7 @@
 import pytest
 
 import kernel_tuner
-from kernel_tuner import util
+from kernel_tuner.util import InvalidConfig
 from kernel_tuner.interface import strategy_map
 
 from ..context import skip_if_no_bayesopt_botorch, skip_if_no_bayesopt_gpytorch
@@ -75,7 +75,7 @@ def test_strategies(vector_add, strategy):
         unique_results = {}
         for result in results:
             x_int = ",".join([str(v) for k, v in result.items() if k in tune_params])
-            if not isinstance(result["time"], util.InvalidConfig):
+            if not isinstance(result["time"], InvalidConfig):
                 unique_results[x_int] = result["time"]
         assert len(unique_results) <= filter_options["max_fevals"]
 
diff --git a/test/test_compiler_functions.py b/test/test_compiler_functions.py
@@ -13,7 +13,7 @@
 import kernel_tuner
 from kernel_tuner.backends.compiler import CompilerFunctions, Argument, is_cupy_array, get_array_module
 from kernel_tuner.core import KernelSource, KernelInstance
-from kernel_tuner import util
+from kernel_tuner.util import delete_temp_file
 
 from .context import skip_if_no_gfortran, skip_if_no_gcc, skip_if_no_openmp, skip_if_no_cupy
 from .test_runners import env as cuda_env  # noqa: F401
@@ -391,7 +391,7 @@ def test_complies_fortran_function_with_module():
         assert np.isclose(result, 42.0)
 
     finally:
-        util.delete_temp_file("my_fancy_module.mod")
+        delete_temp_file("my_fancy_module.mod")
 
 
 @pytest.fixture
diff --git a/test/test_integration.py b/test/test_integration.py
@@ -5,7 +5,7 @@
 import pytest
 
 from kernel_tuner import integration
-from kernel_tuner import util
+from kernel_tuner.util import delete_temp_file
 from datetime import datetime, timezone
 
 
@@ -71,7 +71,7 @@ def test_store_results(fake_results):
         assert my_gpu_100_data[0]["time"] < 100
 
     finally:
-        util.delete_temp_file(filename)
+        delete_temp_file(filename)
 
 
 def test_setup_device_targets(fake_results):
@@ -136,8 +136,8 @@ def test_setup_device_targets(fake_results):
         assert expected in output_str
 
     finally:
-        util.delete_temp_file(results_filename)
-        util.delete_temp_file(header_filename)
+        delete_temp_file(results_filename)
+        delete_temp_file(header_filename)
 
 
 def test_setup_device_targets_max(fake_results):
@@ -174,5 +174,5 @@ def test_setup_device_targets_max(fake_results):
         assert expected in output_str
 
     finally:
-        util.delete_temp_file(results_filename)
-        util.delete_temp_file(header_filename)
+        delete_temp_file(results_filename)
+        delete_temp_file(header_filename)
diff --git a/test/test_kernelbuilder.py b/test/test_kernelbuilder.py
@@ -3,8 +3,8 @@
 
 import pytest
 from kernel_tuner import kernelbuilder
-from kernel_tuner import util
 from kernel_tuner import integration
+from kernel_tuner.util import delete_temp_file
 
 
 backends = ["cuda", "cupy"]
@@ -59,4 +59,4 @@ def test_PythonKernel_tuned(test_kernel, backend):
         assert np.allclose(reference[0], a+b)
 
     finally:
-        util.delete_temp_file(test_results_file)
+        delete_temp_file(test_results_file)
