Rework signatures

baybe/recommenders/naive.py

@@ -6,6 +6,7 @@
 import pandas as pd
 from attrs import define, evolve, field, fields
 
+from baybe.exceptions import UnusedObjectWarning
 from baybe.objectives.base import Objective
 from baybe.recommenders.pure.base import PureRecommender
 from baybe.recommenders.pure.bayesian.base import BayesianRecommender
@@ -113,6 +114,13 @@ def recommend(  # noqa: D102
             )
 
         # We are in a hybrid setting now
+        if pending_measurements is not None:
+            raise UnusedObjectWarning(
+                f"A non-empty set of pending measurements was provided, but the "
+                f"selected recommender {self.__class__.__name__} with discrete "
+                f"sub-recommender {self.disc_recommender.__class__.__name__} only "
+                f"utilizes this information for purely discrete spaces."
+            )
 
         # We will attach continuous parts to discrete parts and the other way round.
         # To make things simple, we sample a single point in the continuous space which
@@ -123,21 +131,11 @@ def recommend(  # noqa: D102
 
         # Get discrete candidates. The metadata flags are ignored since the search space
         # is hybrid
-        # TODO Slight BOILERPLATE CODE, see recommender.py, ll. 47+
         _, candidates_comp = searchspace.discrete.get_candidates(
             allow_repeated_recommendations=True,
             allow_recommending_already_measured=True,
         )
 
-        # Transform pending measurements
-        if pending_measurements is None:
-            pending_comp_discrete = None
-            pending_comp_continuous = None
-        else:
-            pending_comp = searchspace.transform(pending_measurements)
-            pending_comp_discrete = pending_comp[candidates_comp.columns]
-            pending_comp_continuous = pending_comp[searchspace.continuous.param_names]
-
         # We now check whether the discrete recommender is bayesian.
         if isinstance(self.disc_recommender, BayesianRecommender):
             # Get access to the recommenders acquisition function
@@ -160,7 +158,6 @@ def recommend(  # noqa: D102
             subspace_discrete=searchspace.discrete,
             candidates_comp=candidates_comp,
             batch_size=batch_size,
-            pending_comp=pending_comp_discrete,
         )
 
         # Get one random discrete point that will be attached when evaluating the
@@ -183,7 +180,7 @@ def recommend(  # noqa: D102
 
         # Call the private function of the continuous recommender
         rec_cont = self.cont_recommender._recommend_continuous(
-            searchspace.continuous, batch_size, pending_comp_continuous
+            searchspace.continuous, batch_size
         )
 
         # Glue the solutions together and return them

baybe/recommenders/pure/base.py

@@ -42,17 +42,8 @@ def recommend(  # noqa: D102
     ) -> pd.DataFrame:
         # See base class
         if searchspace.type is SearchSpaceType.CONTINUOUS:
-            # Transform pending measurements
-            pending_comp = (
-                None
-                if pending_measurements is None
-                else searchspace.transform(pending_measurements)
-            )
-
             return self._recommend_continuous(
-                subspace_continuous=searchspace.continuous,
-                batch_size=batch_size,
-                pending_comp=pending_comp,
+                subspace_continuous=searchspace.continuous, batch_size=batch_size
             )
         else:
             return self._recommend_with_discrete_parts(
@@ -66,7 +57,6 @@ def _recommend_discrete(
         subspace_discrete: SubspaceDiscrete,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None,
     ) -> pd.Index:
         """Generate recommendations from a discrete search space.
 
@@ -76,7 +66,6 @@ def _recommend_discrete(
             candidates_comp: The computational representation of all discrete candidate
                 points to be considered.
             batch_size: The size of the recommendation batch.
-            pending_comp: Computational representation of pending measurements.
 
         Raises:
             NotImplementedError: If the function is not implemented by the child class.
@@ -92,7 +81,6 @@ def _recommend_discrete(
                 searchspace=SearchSpace(discrete=subspace_discrete),
                 candidates_comp=candidates_comp,
                 batch_size=batch_size,
-                pending_comp=pending_comp,
             ).index
         except NotImplementedError as exc:
             raise NotImplementedError(
@@ -108,15 +96,13 @@ def _recommend_continuous(
         self,
         subspace_continuous: SubspaceContinuous,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.DataFrame:
         """Generate recommendations from a continuous search space.
 
         Args:
             subspace_continuous: The continuous subspace from which to generate
                 recommendations.
             batch_size: The size of the recommendation batch.
-            pending_comp: Pending points in computational representation.
 
         Raises:
             NotImplementedError: If the function is not implemented by the child class.
@@ -131,7 +117,6 @@ def _recommend_continuous(
                 searchspace=SearchSpace(continuous=subspace_continuous),
                 candidates_comp=pd.DataFrame(),
                 batch_size=batch_size,
-                pending_comp=pending_comp,
             )
         except NotImplementedError as exc:
             raise NotImplementedError(
@@ -148,7 +133,6 @@ def _recommend_hybrid(
         searchspace: SearchSpace,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None,
     ) -> pd.DataFrame:
         """Generate recommendations from a hybrid search space.
 
@@ -162,8 +146,6 @@ def _recommend_hybrid(
             candidates_comp: The computational representation of all discrete candidate
                 points to be considered.
             batch_size: The size of the recommendation batch.
-            pending_comp: Pending points in computational representation.
-
 
         Raises:
             NotImplementedError: If the function is not implemented by the child class.
@@ -177,7 +159,7 @@ def _recommend_with_discrete_parts(
         self,
         searchspace: SearchSpace,
         batch_size: int,
-        pending_measurements: pd.DataFrame | None = None,
+        pending_measurements: pd.DataFrame | None,
     ) -> pd.DataFrame:
         """Obtain recommendations in search spaces with a discrete part.
 
@@ -200,18 +182,13 @@ def _recommend_with_discrete_parts(
 
         # Get discrete candidates
         # Repeated recommendations are always allowed for hybrid spaces
+        # Pending points are excluded for discrete spaces
         _, candidates_comp = searchspace.discrete.get_candidates(
             allow_repeated_recommendations=is_hybrid_space
             or self.allow_repeated_recommendations,
             allow_recommending_already_measured=is_hybrid_space
             or self.allow_recommending_already_measured,
-        )
-
-        # Transform pending measurements
-        pending_comp = (
-            None
-            if pending_measurements is None
-            else searchspace.transform(pending_measurements)
+            exclude=None if is_hybrid_space else pending_measurements,
         )
 
         # Check if enough candidates are left
@@ -228,13 +205,11 @@ def _recommend_with_discrete_parts(
 
         # Get recommendations
         if is_hybrid_space:
-            rec = self._recommend_hybrid(
-                searchspace, candidates_comp, batch_size, pending_comp
-            )
+            rec = self._recommend_hybrid(searchspace, candidates_comp, batch_size)
             idxs = rec.index
         else:
             idxs = self._recommend_discrete(
-                searchspace.discrete, candidates_comp, batch_size, pending_comp
+                searchspace.discrete, candidates_comp, batch_size
             )
             rec = searchspace.discrete.exp_rep.loc[idxs, :]
 

baybe/recommenders/pure/bayesian/botorch.py

@@ -77,7 +77,6 @@ def _recommend_discrete(
         subspace_discrete: SubspaceDiscrete,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.Index:
         """Generate recommendations from a discrete search space.
 
@@ -87,7 +86,6 @@ def _recommend_discrete(
             candidates_comp: The computational representation of all discrete candidate
                 points to be considered.
             batch_size: The size of the recommendation batch.
-            pending_comp: Computational representation of pending measurements.
 
         Raises:
             IncompatibleAcquisitionFunctionError: If a non-Monte Carlo acquisition
@@ -131,15 +129,13 @@ def _recommend_continuous(
         self,
         subspace_continuous: SubspaceContinuous,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.DataFrame:
         """Generate recommendations from a continuous search space.
 
         Args:
             subspace_continuous: The continuous subspace from which to generate
                 recommendations.
             batch_size: The size of the recommendation batch.
-            pending_comp: Computational representation of pending measurements.
 
         Raises:
             IncompatibleAcquisitionFunctionError: If a non-Monte Carlo acquisition
@@ -186,7 +182,6 @@ def _recommend_hybrid(
         searchspace: SearchSpace,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.DataFrame:
         """Recommend points using the ``optimize_acqf_mixed`` function of BoTorch.
 
@@ -202,7 +197,6 @@ def _recommend_hybrid(
             candidates_comp: The computational representation of the candidates
                 of the discrete subspace.
             batch_size: The size of the calculated batch.
-            pending_comp: Computational representation of pending measurements.
 
         Raises:
             IncompatibleAcquisitionFunctionError: If a non-Monte Carlo acquisition

baybe/recommenders/pure/nonpredictive/base.py

@@ -9,7 +9,7 @@
 from baybe.exceptions import UnusedObjectWarning
 from baybe.objectives.base import Objective
 from baybe.recommenders.pure.base import PureRecommender
-from baybe.searchspace.core import SearchSpace
+from baybe.searchspace.core import SearchSpace, SearchSpaceType
 
 
 @define
@@ -40,9 +40,19 @@ def recommend(  # noqa: D102
                 f"consider any objectives, meaning that the argument is ignored.",
                 UnusedObjectWarning,
             )
+        if (
+            pending_measurements is not None
+            and searchspace.type is not SearchSpaceType.DISCRETE
+        ):
+            raise UnusedObjectWarning(
+                f"A non-empty set of pending measurements was provided, but the "
+                f"selected recommender {self.__class__.__name__} only utilizes this "
+                f"information for purely discrete spaces."
+            )
         return super().recommend(
             batch_size=batch_size,
             searchspace=searchspace,
             objective=objective,
             measurements=measurements,
+            pending_measurements=pending_measurements,
         )

baybe/recommenders/pure/nonpredictive/clustering.py

@@ -98,7 +98,6 @@ def _recommend_discrete(
         subspace_discrete: SubspaceDiscrete,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.Index:
         # See base class.
 
@@ -107,14 +106,6 @@ def _recommend_discrete(
         scaler = StandardScaler()
         scaler.fit(subspace_discrete.comp_rep)
 
-        # Ignore exact pending point matches in the candidates
-        if pending_comp is not None:
-            candidates_comp = (
-                candidates_comp.merge(pending_comp, indicator=True, how="outer")
-                .query('_merge == "left_only"')
-                .drop(columns=["_merge"])
-            )
-
         # Scale candidates
         candidates_scaled = np.ascontiguousarray(scaler.transform(candidates_comp))
 

baybe/recommenders/pure/nonpredictive/sampling.py

@@ -1,13 +1,11 @@
 """Recommenders based on sampling."""
 
-import warnings
 from typing import ClassVar
 
 import numpy as np
 import pandas as pd
 from sklearn.preprocessing import StandardScaler
 
-from baybe.exceptions import UnusedObjectWarning
 from baybe.recommenders.pure.nonpredictive.base import NonPredictiveRecommender
 from baybe.searchspace import SearchSpace, SearchSpaceType, SubspaceDiscrete
 from baybe.utils.sampling_algorithms import farthest_point_sampling
@@ -25,18 +23,8 @@ def _recommend_hybrid(
         searchspace: SearchSpace,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.DataFrame:
         # See base class.
-
-        if (pending_comp is not None) and (len(pending_comp) != 0):
-            warnings.warn(
-                f"'{self.recommend.__name__}' was called with a non-empty "
-                f"set of pending measurements but '{self.__class__.__name__}' does not "
-                f"utilize this information, meaning that the argument is ignored.",
-                UnusedObjectWarning,
-            )
-
         if searchspace.type == SearchSpaceType.DISCRETE:
             return candidates_comp.sample(batch_size)
 
@@ -69,7 +57,6 @@ def _recommend_discrete(
         subspace_discrete: SubspaceDiscrete,
         candidates_comp: pd.DataFrame,
         batch_size: int,
-        pending_comp: pd.DataFrame | None = None,
     ) -> pd.Index:
         # See base class.
 
@@ -78,14 +65,6 @@ def _recommend_discrete(
         scaler = StandardScaler()
         scaler.fit(subspace_discrete.comp_rep)
 
-        # Ignore exact pending point matches in the candidates
-        if pending_comp is not None:
-            candidates_comp = (
-                candidates_comp.merge(pending_comp, indicator=True, how="outer")
-                .query('_merge == "left_only"')
-                .drop(columns=["_merge"])
-            )
-
         # Scale and sample
         candidates_scaled = np.ascontiguousarray(scaler.transform(candidates_comp))
         ilocs = farthest_point_sampling(candidates_scaled, batch_size)

baybe/searchspace/discrete.py

@@ -640,6 +640,7 @@ def get_candidates(
         self,
         allow_repeated_recommendations: bool = False,
         allow_recommending_already_measured: bool = False,
+        exclude: pd.DataFrame | None = None,
     ) -> tuple[pd.DataFrame, pd.DataFrame]:
         """Return the set of candidate parameter settings that can be tested.
 
@@ -652,6 +653,8 @@ def get_candidates(
             allow_recommending_already_measured: If ``True``, parameters settings for
                 which there are already target values available are still considered as
                 valid candidates.
+            exclude: Points in experimental representation that should be excluded as
+                candidates.
 
         Returns:
             The candidate parameter settings both in experimental and computational
@@ -664,6 +667,12 @@ def get_candidates(
         if not allow_recommending_already_measured:
             mask_todrop |= self.metadata["was_measured"]
 
+        # Remove additional excludes
+        if exclude is not None:
+            mask_todrop |= pd.merge(self.exp_rep, exclude, indicator=True, how="left")[
+                "_merge"
+            ].eq("both")
+
         return self.exp_rep.loc[~mask_todrop], self.comp_rep.loc[~mask_todrop]
 
     def transform(