add interpoint to botorch (#426)

bofire/data_models/domain/domain.py

@@ -21,6 +21,7 @@
 from bofire.data_models.constraints.api import (
     AnyConstraint,
     ConstraintNotFulfilledError,
+    InterpointEqualityConstraint,
     LinearConstraint,
     NChooseKConstraint,
     ProductConstraint,
@@ -154,6 +155,9 @@ def validate_constraints(self):
             for f in c.features:  # type: ignore
                 if f not in keys:
                     raise ValueError(f"feature {f} in constraint unknown ({keys})")
+        for c in self.constraints.get(InterpointEqualityConstraint):
+            if c.feature not in keys:
+                raise ValueError(f"feature {c.feature} not known.")
         return self
 
     @model_validator(mode="after")

bofire/data_models/strategies/predictives/botorch.py

@@ -7,13 +7,18 @@
 from bofire.data_models.base import BaseModel
 from bofire.data_models.constraints.api import (
     Constraint,
+    InterpointConstraint,
     LinearConstraint,
     NonlinearEqualityConstraint,
     NonlinearInequalityConstraint,
 )
 from bofire.data_models.domain.api import Domain, Outputs
 from bofire.data_models.enum import CategoricalEncodingEnum, CategoricalMethodEnum
-from bofire.data_models.features.api import CategoricalDescriptorInput, CategoricalInput
+from bofire.data_models.features.api import (
+    CategoricalDescriptorInput,
+    CategoricalInput,
+    ContinuousInput,
+)
 from bofire.data_models.outlier_detection.api import OutlierDetections
 from bofire.data_models.strategies.predictives.predictive import PredictiveStrategy
 from bofire.data_models.strategies.shortest_path import has_local_search_region
@@ -126,6 +131,16 @@ def is_constraint_implemented(cls, my_type: Type[Constraint]) -> bool:
             return False
         return True
 
+    @model_validator(mode="after")
+    def validate_interpoint_constraints(self):
+        if self.domain.constraints.get(InterpointConstraint) and len(
+            self.domain.inputs.get(ContinuousInput)
+        ) != len(self.domain.inputs):
+            raise ValueError(
+                "Interpoint constraints can only be used for pure continuous search spaces."
+            )
+        return self
+
     @model_validator(mode="after")
     def validate_surrogate_specs(self):
         """Ensures that a prediction model is specified for each output feature"""

bofire/data_models/strategies/predictives/qparego.py

@@ -3,6 +3,12 @@
 from pydantic import Field
 
 from bofire.data_models.acquisition_functions.api import qEI, qLogEI, qLogNEI, qNEI
+from bofire.data_models.constraints.api import (
+    Constraint,
+    InterpointConstraint,
+    NonlinearEqualityConstraint,
+    NonlinearInequalityConstraint,
+)
 from bofire.data_models.features.api import Feature
 from bofire.data_models.objectives.api import (
     CloseToTargetObjective,
@@ -40,3 +46,21 @@ def is_objective_implemented(cls, my_type: Type[Objective]) -> bool:
         ]:
             return False
         return True
+
+    @classmethod
+    def is_constraint_implemented(cls, my_type: Type[Constraint]) -> bool:
+        """Method to check if a specific constraint type is implemented for the strategy
+
+        Args:
+            my_type (Type[Constraint]): Constraint class
+
+        Returns:
+            bool: True if the constraint type is valid for the strategy chosen, False otherwise
+        """
+        if my_type in [
+            NonlinearInequalityConstraint,
+            NonlinearEqualityConstraint,
+            InterpointConstraint,
+        ]:
+            return False
+        return True

bofire/strategies/predictives/botorch.py

@@ -45,6 +45,7 @@
 from bofire.surrogates.botorch_surrogates import BotorchSurrogates
 from bofire.utils.torch_tools import (
     get_initial_conditions_generator,
+    get_interpoint_constraints,
     get_linear_constraints,
     get_nonlinear_constraints,
     tkwargs,
@@ -358,6 +359,9 @@ def _optimize_acqf_continuous(
                     options=self._get_optimizer_options(),  # type: ignore
                 )
             else:
+                interpoints = get_interpoint_constraints(
+                    domain=self.domain, n_candidates=candidate_count
+                )
                 candidates, acqf_vals = optimize_acqf(
                     acq_function=acqfs[0],
                     bounds=bounds,
@@ -367,7 +371,8 @@ def _optimize_acqf_continuous(
                     equality_constraints=get_linear_constraints(
                         domain=self.domain,
                         constraint=LinearEqualityConstraint,  # type: ignore
-                    ),
+                    )
+                    + interpoints,
                     inequality_constraints=get_linear_constraints(
                         domain=self.domain,
                         constraint=LinearInequalityConstraint,  # type: ignore

bofire/utils/torch_tools.py

@@ -106,6 +106,8 @@ def get_interpoint_constraints(
             of a tensor with the feature indices, coefficients and a float for the rhs.
     """
     constraints = []
+    if n_candidates == 1:
+        return constraints
     for constraint in domain.constraints.get(InterpointEqualityConstraint):
         assert isinstance(constraint, InterpointEqualityConstraint)
         coefficients = torch.tensor([1.0, -1.0]).to(**tkwargs)

tests/bofire/data_models/specs/domain.py

@@ -1,3 +1,4 @@
+from bofire.data_models.constraints.api import InterpointEqualityConstraint
 from bofire.data_models.domain.api import Constraints, Domain, Inputs, Outputs
 from bofire.data_models.features.api import ContinuousInput, ContinuousOutput
 from tests.bofire.data_models.specs.features import specs as features
@@ -71,3 +72,25 @@
     error=ValueError,
     message="Feature keys are not unique",
 )
+
+specs.add_invalid(
+    Domain,
+    lambda: {
+        "inputs": Inputs(
+            features=[
+                features.valid(ContinuousInput).obj(key="i1"),
+                features.valid(ContinuousInput).obj(key="i2"),
+            ]
+        ),
+        "outputs": Outputs(
+            features=[
+                features.valid(ContinuousOutput).obj(key="o1"),
+            ]
+        ),
+        "constraints": Constraints(
+            constraints=[InterpointEqualityConstraint(feature="i3")]
+        ),
+    },
+    error=ValueError,
+    message="feature i3 not known.",
+)

tests/bofire/data_models/specs/strategies.py

@@ -1,6 +1,7 @@
 import bofire.data_models.strategies.api as strategies
 from bofire.data_models.acquisition_functions.api import qEI, qLogNEHVI, qPI
 from bofire.data_models.constraints.api import (
+    InterpointEqualityConstraint,
     LinearEqualityConstraint,
     LinearInequalityConstraint,
     NChooseKConstraint,
@@ -416,6 +417,29 @@
     message="LSR-BO only supported for linear constraints.",
 )
 
+specs.add_invalid(
+    strategies.SoboStrategy,
+    lambda: {
+        "domain": Domain(
+            inputs=Inputs(
+                features=[
+                    ContinuousInput(
+                        key=k, bounds=(0, 1), local_relative_bounds=(0.1, 0.1)
+                    )
+                    for k in ["a", "b", "c"]
+                ]
+                + [CategoricalInput(key="d", categories=["a", "b", "c"])]
+            ),
+            outputs=Outputs(features=[ContinuousOutput(key="alpha")]),
+            constraints=Constraints(
+                constraints=[InterpointEqualityConstraint(feature="a")]
+            ),
+        ).model_dump(),
+    },
+    error=ValueError,
+    message="Interpoint constraints can only be used for pure continuous search spaces.",
+)
+
 specs.add_valid(
     strategies.FractionalFactorialStrategy,
     lambda: {

tests/bofire/strategies/test_sobo.py

@@ -31,7 +31,10 @@
     qSR,
     qUCB,
 )
-from bofire.data_models.constraints.api import NChooseKConstraint
+from bofire.data_models.constraints.api import (
+    InterpointEqualityConstraint,
+    NChooseKConstraint,
+)
 from bofire.data_models.domain.api import Domain, Inputs, Outputs
 from bofire.data_models.features.api import ContinuousInput, ContinuousOutput
 from bofire.data_models.objectives.api import (
@@ -478,3 +481,14 @@ def test_sobo_get_optimizer_options():
     strategy_data = data_models.SoboStrategy(domain=domain, maxiter=500, batch_limit=4)
     strategy = SoboStrategy(data_model=strategy_data)
     assert strategy._get_optimizer_options() == {"maxiter": 500, "batch_limit": 1}
+
+
+def test_sobo_interpoint():
+    bench = Himmelblau()
+    experiments = bench.f(bench.domain.inputs.sample(4), return_complete=True)
+    domain = bench._domain
+    domain.constraints.constraints.append(InterpointEqualityConstraint(feature="x_1"))
+    strategy_data = data_models.SoboStrategy(domain=domain)
+    strategy = SoboStrategy(data_model=strategy_data)
+    strategy.tell(experiments)
+    strategy.ask(2)

tests/bofire/utils/test_torch_tools.py

@@ -352,6 +352,8 @@ def test_get_interpoint_equality_constraints():
             dtype=torch.int64,
         ),
     )
+    constraints = get_interpoint_constraints(domain=domain, n_candidates=1)
+    assert len(constraints) == 0
 
 
 def test_get_linear_constraints():