Refactor adapter model (#260)

This PR moves the `AdapterModel` (i.e., the connecting layer between
baybe surrogates and botorch models) to the surrogate package, where it
actually belongs. Additionally, the surrogates are equipped with a
`to_botorch` method that simplifies the model translation and can be
customized per subclass. In particular, GP surrogates are no longer
wrapped using the adapter but now expose their internal botorch model
instance directly.

CHANGELOG.md

@@ -5,8 +5,12 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+### Added
+- `Surrogate` base class now exposes a `to_botorch` method
+
 ### Changed
 - Passing an `Objective` to `Campaign` is now optional
+- `GaussianProcessSurrogate` models are no longer wrapped when cast to BoTorch
 
 ### Removed
 - Support for Python 3.9 removed due to new [BoTorch requirements](https://github.com/pytorch/botorch/pull/2293) 

baybe/acquisition/base.py

@@ -43,15 +43,13 @@ def to_botorch(
         """Create the botorch-ready representation of the function."""
         import botorch.acquisition as botorch_analytical_acqf
 
-        from baybe.acquisition._adapter import AdapterModel
-
         acqf_cls = getattr(botorch_analytical_acqf, self.__class__.__name__)
         params_dict = filter_attributes(object=self, callable_=acqf_cls.__init__)
 
         additional_params = {
             p: v
             for p, v in {
-                "model": AdapterModel(surrogate),
+                "model": surrogate.to_botorch(),
                 "best_f": train_y.max().item(),
                 "X_baseline": to_tensor(train_x),
             }.items()

baybe/acquisition/_adapter.py → baybe/surrogates/_adapter.py

@@ -1,4 +1,4 @@
-"""Adapter for making BoTorch's acquisition functions work with BayBE models."""
+"""Adapter functionality for making BayBE surrogates BoTorch-ready."""
 
 from collections.abc import Callable
 from typing import Any
@@ -19,7 +19,7 @@ class AdapterModel(Model):
     surrogate model usable in conjunction with BoTorch acquisition functions.
 
     Args:
-        surrogate: The internal surrogate model
+        surrogate: The internal surrogate model.
     """
 
     def __init__(self, surrogate: Surrogate):

baybe/surrogates/base.py

@@ -25,6 +25,7 @@
 from baybe.surrogates.utils import _prepare_inputs, _prepare_targets
 
 if TYPE_CHECKING:
+    from botorch.models.model import Model
     from torch import Tensor
 
 # Define constants
@@ -55,6 +56,12 @@ class Surrogate(ABC, SerialMixin):
     """Class variable encoding whether or not the surrogate supports transfer
     learning."""
 
+    def to_botorch(self) -> Model:
+        """Create the botorch-ready representation of the model."""
+        from baybe.surrogates._adapter import AdapterModel
+
+        return AdapterModel(self)
+
     def posterior(self, candidates: Tensor) -> tuple[Tensor, Tensor]:
         """Evaluate the surrogate model at the given candidate points.
 

baybe/surrogates/gaussian_process/core.py

@@ -22,6 +22,7 @@
 )
 
 if TYPE_CHECKING:
+    from botorch.models.model import Model
     from torch import Tensor
 
 
@@ -59,6 +60,11 @@ def from_preset(preset: GaussianProcessPreset) -> GaussianProcessSurrogate:
         """Create a Gaussian process surrogate from one of the defined presets."""
         return make_gp_from_preset(preset)
 
+    def to_botorch(self) -> Model:  # noqa: D102
+        # See base class.
+
+        return self._model
+
     def _posterior(self, candidates: Tensor) -> tuple[Tensor, Tensor]:
         # See base class.
         posterior = self._model.posterior(candidates)

tests/test_imports.py

@@ -58,8 +58,8 @@ def test_imports(module: str):
 
 WHITELISTS = {
     "torch": [
-        "baybe.acquisition._adapter",
         "baybe.acquisition.partial",
+        "baybe.surrogates._adapter",
         "baybe.utils.botorch_wrapper",
         "baybe.utils.torch",
     ],