Move add_parameter_noise to simulation package

baybe/simulation/core.py

@@ -15,8 +15,8 @@
 from baybe.campaign import Campaign
 from baybe.exceptions import NotEnoughPointsLeftError, NothingToSimulateError
 from baybe.simulation.lookup import look_up_targets
+from baybe.simulation.utils import add_parameter_noise
 from baybe.targets.enum import TargetMode
-from baybe.utils.dataframe import add_parameter_noise
 from baybe.utils.numerical import DTypeFloatNumpy, closer_element, closest_element
 from baybe.utils.random import temporary_seed
 

baybe/simulation/utils.py

@@ -0,0 +1,61 @@
+"""Simulation utilities."""
+
+from __future__ import annotations
+
+from collections.abc import Iterable
+from typing import TYPE_CHECKING, Literal
+
+import numpy as np
+import pandas as pd
+
+if TYPE_CHECKING:
+    from baybe.parameters import Parameter
+
+
+def add_parameter_noise(
+    data: pd.DataFrame,
+    parameters: Iterable[Parameter],
+    noise_type: Literal["absolute", "relative_percent"] = "absolute",
+    noise_level: float = 1.0,
+) -> None:
+    """Apply uniform noise to the parameter values of a recommendation frame.
+
+    The noise can be additive or multiplicative.
+    This can be used to simulate experimental noise or imperfect user input containing
+    numerical parameter values that differ from the recommendations. Note that the
+    dataframe is modified in-place, and that no new dataframe is returned.
+
+    Args:
+        data: Output of the ``recommend`` function of a ``Campaign`` object, see
+            :func:`baybe.campaign.Campaign.recommend`.
+        parameters: The parameters for which the values shall be corrupted.
+        noise_type: Defines whether the noise should be additive or multiplicative.
+        noise_level: Level/magnitude of the noise. Must be provided as numerical value
+            for noise type ``absolute`` and as percentage for noise type
+            ``relative_percent``.
+
+    Raises:
+        ValueError: If ``noise_type`` is neither ``absolute`` nor
+            ``relative_percent``.
+    """
+    # Validate input
+    if noise_type not in ("relative_percent", "absolute"):
+        raise ValueError(
+            f"Parameter 'noise_type' was {noise_type} but must be either "
+            "'absolute' or 'relative_percent'."
+        )
+
+    for param in (p for p in parameters if p.is_numeric):
+        # Add selected noise type
+        if noise_type == "relative_percent":
+            data[param.name] *= np.random.uniform(
+                1.0 - noise_level / 100.0, 1.0 + noise_level / 100.0, len(data)
+            )
+        elif noise_type == "absolute":
+            data[param.name] += np.random.uniform(-noise_level, noise_level, len(data))
+
+        # Respect continuous intervals
+        if param.is_continuous:
+            data[param.name] = data[param.name].clip(
+                param.bounds.lower, param.bounds.upper
+            )

baybe/utils/dataframe.py

@@ -3,14 +3,9 @@
 from __future__ import annotations
 
 import logging
-from collections.abc import Iterable, Iterator, Sequence
-from typing import (
-    TYPE_CHECKING,
-    Literal,
-    overload,
-)
-
-import numpy as np
+from collections.abc import Iterator, Sequence
+from typing import TYPE_CHECKING, overload
+
 import pandas as pd
 
 from baybe.utils.numerical import DTypeFloatNumpy
@@ -63,55 +58,6 @@ def to_tensor(*dfs: pd.DataFrame) -> Tensor | Iterator[Tensor]:
     return out
 
 
-def add_parameter_noise(
-    data: pd.DataFrame,
-    parameters: Iterable[Parameter],
-    noise_type: Literal["absolute", "relative_percent"] = "absolute",
-    noise_level: float = 1.0,
-) -> None:
-    """Apply uniform noise to the parameter values of a recommendation frame.
-
-    The noise can be additive or multiplicative.
-    This can be used to simulate experimental noise or imperfect user input containing
-    numerical parameter values that differ from the recommendations. Note that the
-    dataframe is modified in-place, and that no new dataframe is returned.
-
-    Args:
-        data: Output of the ``recommend`` function of a ``Campaign`` object, see
-            :func:`baybe.campaign.Campaign.recommend`.
-        parameters: The parameters for which the values shall be corrupted.
-        noise_type: Defines whether the noise should be additive or multiplicative.
-        noise_level: Level/magnitude of the noise. Must be provided as numerical value
-            for noise type ``absolute`` and as percentage for noise type
-            ``relative_percent``.
-
-    Raises:
-        ValueError: If ``noise_type`` is neither ``absolute`` nor
-            ``relative_percent``.
-    """
-    # Validate input
-    if noise_type not in ("relative_percent", "absolute"):
-        raise ValueError(
-            f"Parameter 'noise_type' was {noise_type} but must be either "
-            "'absolute' or 'relative_percent'."
-        )
-
-    for param in (p for p in parameters if p.is_numeric):
-        # Add selected noise type
-        if noise_type == "relative_percent":
-            data[param.name] *= np.random.uniform(
-                1.0 - noise_level / 100.0, 1.0 + noise_level / 100.0, len(data)
-            )
-        elif noise_type == "absolute":
-            data[param.name] += np.random.uniform(-noise_level, noise_level, len(data))
-
-        # Respect continuous intervals
-        if param.is_continuous:
-            data[param.name] = data[param.name].clip(
-                param.bounds.lower, param.bounds.upper
-            )
-
-
 def df_drop_single_value_columns(
     df: pd.DataFrame, lst_exclude: list = None
 ) -> pd.DataFrame:

tests/conftest.py

@@ -52,6 +52,7 @@
 )
 from baybe.recommenders.pure.nonpredictive.sampling import RandomRecommender
 from baybe.searchspace import SearchSpace
+from baybe.simulation.utils import add_parameter_noise
 from baybe.surrogates import _ONNX_INSTALLED, GaussianProcessSurrogate
 from baybe.targets import NumericalTarget
 from baybe.telemetry import (
@@ -61,7 +62,7 @@
 )
 from baybe.utils.basic import hilberts_factory
 from baybe.utils.boolean import strtobool
-from baybe.utils.dataframe import add_fake_results, add_parameter_noise
+from baybe.utils.dataframe import add_fake_results
 
 try:
     import baybe.utils.chemistry  # noqa: F401  # Tests if chem deps are available