changes in hyperparameters and defaults

src/f3dasm/_src/experimentdata/experimentdata.py

@@ -1517,17 +1517,17 @@ def _iterate(self, optimizer: Optimizer, data_generator: DataGenerator,
         if isinstance(x0_selection, str):
             if x0_selection == 'new':
 
-                if iterations < optimizer.hyperparameters.population:
+                if iterations < optimizer.hyperparameters["population"]:
                     raise ValueError(
                         f'For creating new samples, the total number of '
                         f'requested iterations ({iterations}) cannot be '
                         f'smaller than the population size '
-                        f'({optimizer.hyperparameters.population})')
+                        f'({optimizer.hyperparameters["population"]})')
 
                 init_samples = ExperimentData.from_sampling(
                     domain=self.domain,
                     sampler=sampler,
-                    n_samples=optimizer.hyperparameters.population,
+                    n_samples=optimizer.hyperparameters["population"],
                     seed=optimizer.seed)
 
                 init_samples.evaluate(
@@ -1548,10 +1548,10 @@ def _iterate(self, optimizer: Optimizer, data_generator: DataGenerator,
                     self.add_experiments(init_samples)
 
                 x0_selection = 'last'
-                iterations -= optimizer.hyperparameters.population
+                iterations -= optimizer.hyperparameters["population"]
 
         x0 = x0_factory(experiment_data=self, mode=x0_selection,
-                        n_samples=optimizer.hyperparameters.population)
+                        n_samples=optimizer.hyperparameters["population"])
         optimizer.set_data(x0)
 
         optimizer._check_number_of_datapoints()
@@ -1560,7 +1560,7 @@ def _iterate(self, optimizer: Optimizer, data_generator: DataGenerator,
 
         for _ in range(number_of_updates(
                 iterations,
-                population=optimizer.hyperparameters.population)):
+                population=optimizer.hyperparameters["population"])):
             new_samples = optimizer.update_step(data_generator)
 
             # If new_samples is a tuple of input_data and output_data
@@ -1591,7 +1591,8 @@ def _iterate(self, optimizer: Optimizer, data_generator: DataGenerator,
         if not overwrite:
             # Remove overiterations
             self.remove_rows_bottom(number_of_overiterations(
-                iterations, population=optimizer.hyperparameters.population))
+                iterations,
+                population=optimizer.hyperparameters["population"]))
 
         # Reset the optimizer
         optimizer.reset(ExperimentData(domain=self.domain))
@@ -1657,17 +1658,17 @@ def _iterate_scipy(self, optimizer: Optimizer,
         if isinstance(x0_selection, str):
             if x0_selection == 'new':
 
-                if iterations < optimizer.hyperparameters.population:
+                if iterations < optimizer.hyperparameters["population"]:
                     raise ValueError(
                         f'For creating new samples, the total number of '
                         f'requested iterations ({iterations}) cannot be '
                         f'smaller than the population size '
-                        f'({optimizer.hyperparameters.population})')
+                        f'({optimizer.hyperparameters["population"]})')
 
                 init_samples = ExperimentData.from_sampling(
                     domain=self.domain,
                     sampler=sampler,
-                    n_samples=optimizer.hyperparameters.population,
+                    n_samples=optimizer.hyperparameters["population"],
                     seed=optimizer.seed)
 
                 init_samples.evaluate(
@@ -1690,7 +1691,7 @@ def _iterate_scipy(self, optimizer: Optimizer,
                 x0_selection = 'last'
 
         x0 = x0_factory(experiment_data=self, mode=x0_selection,
-                        n_samples=optimizer.hyperparameters.population)
+                        n_samples=optimizer.hyperparameters["population"])
         optimizer.set_data(x0)
 
         optimizer._check_number_of_datapoints()

src/f3dasm/_src/optimization/adapters/scipy_implementations.py

@@ -57,15 +57,15 @@ def fun(x):
             _, y = sample.to_numpy()
             return float(y)
 
-        self.hyperparameters.maxiter = iterations
+        self.hyperparameters['maxiter'] = iterations
 
         minimize(
             fun=fun,
             method=self.method,
             jac=data_generator.dfdx,
             x0=self.data.get_n_best_output(1).to_numpy()[0].ravel(),
             callback=self._callback,
-            options=self.hyperparameters.__dict__,
+            options=self.hyperparameters,
             bounds=self.domain.get_bounds(),
             tol=0.0,
         )

src/f3dasm/_src/optimization/numpy_implementations.py

@@ -6,15 +6,14 @@
 # =============================================================================
 
 # Standard
-from dataclasses import dataclass
 from typing import List, Tuple
 
 # Third-party core
-import autograd.numpy as np
+import numpy as np
 
 # Locals
 from ..datageneration.datagenerator import DataGenerator
-from .optimizer import Optimizer, OptimizerParameters
+from .optimizer import Optimizer
 
 #                                                          Authorship & Credits
 # =============================================================================
@@ -26,27 +25,19 @@
 # =============================================================================
 
 
-@dataclass
-class RandomSearch_Parameters(OptimizerParameters):
-    """Hyperparameters for RandomSearch optimizer"""
-
-    pass
-
-
 class RandomSearch(Optimizer):
     """Naive random search"""
     require_gradients: bool = False
-    hyperparameters: RandomSearch_Parameters = RandomSearch_Parameters()
 
-    def set_seed(self):
-        np.random.seed(self.seed)
+    def set_algorithm(self):
+        self.algorithm = np.random.default_rng(self.seed)
 
     def update_step(
             self, data_generator: DataGenerator
     ) -> Tuple[np.ndarray, np.ndarray]:
         x_new = np.atleast_2d(
             [
-                np.random.uniform(
+                self.algorithm.uniform(
                     low=self.domain.get_bounds()[d, 0],
                     high=self.domain.get_bounds()[d, 1])
                 for d in range(len(self.domain))

src/f3dasm/_src/optimization/optimizer.py

@@ -9,7 +9,8 @@
 
 # Standard
 from dataclasses import dataclass
-from typing import ClassVar, Iterable, List, Optional, Protocol, Tuple
+from typing import (Any, ClassVar, Dict, Iterable, List, Optional, Protocol,
+                    Tuple)
 
 # Third-party core
 import numpy as np
@@ -48,27 +49,13 @@ def to_numpy() -> Tuple[np.ndarray, np.ndarray]:
     def select(self, indices: int | slice | Iterable[int]) -> ExperimentData:
         ...
 
-
-@dataclass
-class OptimizerParameters:
-    """Interface of a continuous benchmark function
-
-    Parameters
-    ----------
-    population : int
-        population of the optimizer update step
-    force_bounds : bool
-        force the optimizer to not exceed the boundaries of the domain
-    """
-
-    population: int = 1
-    force_bounds: bool = True
+# =============================================================================
 
 
 class Optimizer:
     type: ClassVar[str] = 'any'
     require_gradients: ClassVar[bool] = False
-    hyperparameters: OptimizerParameters = OptimizerParameters()
+    default_hyperparameters: Dict[str, Any] = None
 
     def __init__(
             self, domain: Domain, seed: Optional[int] = None,
@@ -98,8 +85,17 @@ def __init__(
         if not hyperparameters:
             hyperparameters = {}
 
+        # Set the default hyperparameters to an empty dictionary if not set
+        if not self.default_hyperparameters:
+            self.default_hyperparameters = {}
+
         # Overwrite the default hyperparameters with the given hyperparameters
-        self.hyperparameters.__init__(**hyperparameters)
+        self.hyperparameters = self.default_hyperparameters.copy()
+        self.hyperparameters.update(hyperparameters)
+
+        # Set the default population to 1
+        if 'population' not in self.hyperparameters:
+            self.hyperparameters['population'] = 1
 
         # Set the name of the optimizer to the class name if no name is given
         if name is None:
@@ -134,10 +130,10 @@ def _check_number_of_datapoints(self):
         ValueError
             Raises then the number of datapoints is insufficient
         """
-        if len(self.data) < self.hyperparameters.population:
+        if len(self.data) < self.hyperparameters['population']:
             raise ValueError(
                 f'There are {len(self.data)} datapoints available, \
-                     need {self.hyperparameters.population} for initial \
+                     need {self.hyperparameters["population"]} for initial \
                          population!'
             )
 
@@ -154,9 +150,6 @@ def set_data(self, data: ExperimentData):
         """Set the data attribute to the given data"""
         self.data = data
 
-    def add_experiments(self, experiments: ExperimentData):
-        ...
-
     def get_name(self) -> str:
         """Get the name of the optimizer
 

src/f3dasm/_src/optimization/scipy_implementations.py

@@ -6,12 +6,10 @@
 # =============================================================================
 
 # Standard
-from dataclasses import dataclass
 from typing import List
 
 # Locals
 from .adapters.scipy_implementations import _SciPyOptimizer
-from .optimizer import OptimizerParameters
 
 #                                                          Authorship & Credits
 # =============================================================================
@@ -22,60 +20,45 @@
 #
 # =============================================================================
 
-
-@dataclass
-class CG_Parameters(OptimizerParameters):
-    """CG Parameters"""
-
-    gtol: float = 0.0
+CG_DEFAULTS = {'gtol': 0.0}
 
 
 class CG(_SciPyOptimizer):
     """CG"""
     require_gradients: bool = True
     method: str = "CG"
-    hyperparameters: CG_Parameters = CG_Parameters()
+    default_hyperparameters = CG_DEFAULTS
 
     def get_info(self) -> List[str]:
         return ['Stable', 'First-Order', 'Single-Solution']
 
 # =============================================================================
 
 
-@dataclass
-class LBFGSB_Parameters(OptimizerParameters):
-    """Hyperparameters for LBFGSB optimizer"""
-
-    ftol: float = 0.0
-    gtol: float = 0.0
+LBFGSB_DEFAULTS = {'ftol': 0.0, 'gtol': 0.0}
 
 
 class LBFGSB(_SciPyOptimizer):
     """L-BFGS-B"""
     require_gradients: bool = True
     method: str = "L-BFGS-B"
-    hyperparameters: LBFGSB_Parameters = LBFGSB_Parameters()
+    default_hyperparameters = LBFGSB_DEFAULTS
 
     def get_info(self) -> List[str]:
         return ['Stable', 'First-Order', 'Single-Solution']
 
 # =============================================================================
 
 
-@dataclass
-class NelderMead_Parameters(OptimizerParameters):
-    """Hyperparameters for NelderMead optimizer"""
-
-    xatol: float = 0.0
-    fatol: float = 0.0
-    adaptive: bool = False
+NelderMead_DEFAULTS = {'xatol': 0.0, 'fatol': 0.0,
+                       'adaptive': False}
 
 
 class NelderMead(_SciPyOptimizer):
     """Nelder-Mead"""
     require_gradients: bool = False
     method: str = "Nelder-Mead"
-    hyperparameters: NelderMead_Parameters = NelderMead_Parameters()
+    default_hyperparameters = NelderMead_DEFAULTS
 
     def get_info(self) -> List[str]:
         return ['Fast', 'Global', 'First-Order', 'Single-Solution']

src/f3dasm/optimization.py

@@ -5,7 +5,7 @@
 # =============================================================================
 
 # Local
-from ._src.optimization.optimizer import Optimizer, OptimizerParameters
+from ._src.optimization.optimizer import Optimizer
 from ._src.optimization.optimizer_factory import OPTIMIZERS
 
 #                                                          Authorship & Credits
@@ -19,6 +19,5 @@
 
 __all__ = [
     'Optimizer',
-    'OptimizerParameters',
     'OPTIMIZERS',
 ]

tests/optimization/test_all_optimizers.py

@@ -115,7 +115,7 @@ def test_optimizer_iterations(iterations: int, data_generator: str,
 
     _optimizer = _optimizer_factory(optimizer, domain=domain)
 
-    if x0_selection == "new" and iterations < _optimizer.hyperparameters.population:
+    if x0_selection == "new" and iterations < _optimizer.hyperparameters['population']:
         with pytest.raises(ValueError):
             data.optimize(optimizer=optimizer, data_generator=data_generator,
                           iterations=iterations, kwargs={'seed': seed, 'noise': None,