Refactoring of DoEStrategy data model class (#448)

* add draft of restrucuted doe class * refactoring doe * add formulaic to be installed always * add formulaic to be installed always * add formulaic to be installed always * add formulaic to be installed always * check style * check style * check style * remove enumns * remove enumns * remove enumns * fix branch and bound * move delta into criterion * move delta into criterion * move delta into criterion * move delta into criterion * move default criterion * move default criterion * move default criterion * move default criterion * refactor formulas and number of experiments * pyright * fix test * fix test * fix test * fix tutorial * fix tutorial * fix tutorial * fix test * fix test * fix getting started * aarons review * rmv unneded tests * formulaic version fixed bc of breaking changes * add explanatory text to doe basic examples * typo in basic_examples.ipynb * format basic doe example * consolidate spac_filling with doe * Add categoricals for `FractionalFactorialStrategy` (#480) * integrate factorial in fractional factorial * fix tests * merge main * Multiplicative additive sobo objectives (#481) * added MultiplicativeAdditive data model * added actual multiplicative model (callable is missing) * added torch functions * added test for objective * added test for sobo stategy: multiplicative_additive * changed additive/multiplicative calculations: - Removed scaling by x**(1/(....)) to avoid numerical errors, if x<0 - included weight transformation for multiplicative objectives from (0, 1] to [1, inf) scale to avoid numerical errors at weights != 1.0 - added tests for weights != 1.0 * added notebook for comparison of merging objectives * after hooks * addet .idea/ folder (pycharm) to gitignore * after hooks * Apply pre-commit fixes * Delete .idea directory * corrected tests for multiplicative_additive_botorch_objective * after pre-commit * lint specifications * corrected weightings calc in test for multiplicative objective * after hooks * changed docstrings to google docstring format * easy fixes, spelling errors * forgot linting * easy fixes, spelling errors * removed denominator additive from multiplicative_additive_sobo strategy * after hooks * fixed typing * tensor initialization of objectives * after hooks * avoiding torch size error * avoid linting error * after hooks * reverting test-renaming * revert isinstance list comprehension to tuple.... solution * testing copilot suggestions for linting errors * reverting wrong copilot suggestions * added test for _callables_and_weights * after hooks * added test for SOBO strategy data model * added test for SOBO strategy data model * added new sobo strategy to a mysterious list * after hooks * still trying to get rid of the linting error, expecting tuple(types) * WIP * WIP * WIP * WIP * WIP * minor corrections * add pbar support for hyperopt (#494) * Make the BoFire Data Models OpenAI compatible (#495) * tuples to lists * fix tests * fix linting issues * Group split kfold (#484) * add group kfold option in cross_validate of any traainable surrogate * changed to GroupShuffleSplit, added test case * improve docstring & add some inline comments in test * refactor cross_validate & add tests * imrpve tests, remove unnecessary case while checking group split col * add push * formatting --------- Co-authored-by: Jim Boelrijk Valcon <jimboelrijk1@MacBook-Pro-van-Jim.local> * fix strict candidate enforcement (#492) * Drop support for Python 3.9 (#493) * update tests and pyproject.toml * update lint workflow * update test * bump pyright * different pyright version * change linting * Update pyproject.toml (#501) BoTorch is slowed down massively by scipy 1.15: pytorch/botorch#2668. We should fix it. * kernels working on a given set of features (#476) * kernels working on a given set of features * pre-commit * test map singletaskgp with additive kernel * test active_dims of mapped kernels * add features_to_idx_mapper to outlier detection tutorial * correctly handling categorical mol features * validating mol features transforms * verifying proper type * custom hamming kernel enabling single task gp on categorical features * removed unnecessary parameter from data model * testing equivalence of mixed gp and single gp with custom kernel * (temporary) running on all py versions * (temporary) debug github actions by printing * more printing * Revert "testing equivalence of mixed gp and single gp with custom kernel" This reverts commit 4a2a547. * Revert "removed unnecessary parameter from data model" This reverts commit 6ad1dfd. * Revert "custom hamming kernel enabling single task gp on categorical features" This reverts commit 17d8350. * Revert "Revert "custom hamming kernel enabling single task gp on categorical features"" This reverts commit 2e29852. * Revert "Revert "testing equivalence of mixed gp and single gp with custom kernel"" This reverts commit 1cd2776. * removed test debug and restored to latest implemented features * pinning compatible version of formulaic * pinning compatible version of formulaic * removed old code * lint * removed scratch file * removed old code again * silencing pyright false positive * compatibility with py39 * pin compatible version of formulaic * restored old code * pinning sklearn * pinning sklearn * pinning scikit everywhere * not testing for prediction quality * matching lengthscale constraints in hamming kernel * removed equivalence test * testing hamming kernel * added test for mol features in single task gp * categorical onehot kernel uses the right lengthscale for multiple features * removed redundant check * more descriptive name for base kernel * updated docstring * improved tests and comments --------- Co-authored-by: Robert Lee <84771576+R-M-Lee@users.noreply.github.com> * Fix mapper tests (#502) * fix kernel mapper tests * bump botorch dependency * rmv unused import in staregies.api * rmv unused import in space filling * rmv unused import in space filling * fix data models tets * fix data models tets * fix data models tets * fix data models tets * fix data models tets * no more bnb in test * add fixtures for criteria * add fixtures for criteria --------- Co-authored-by: LinzneDD_basf <dominik.linzner@basf.com> Co-authored-by: Dominik Linzner <44400328+dlinzner-bcs@users.noreply.github.com> Co-authored-by: linznedd <linznedd@basfad.basf.net> Co-authored-by: Robert Lee <84771576+R-M-Lee@users.noreply.github.com> Co-authored-by: Lukas Hebing <92095234+LukasHebing@users.noreply.github.com> Co-authored-by: Julian Keupp <28810427+jkeupp@users.noreply.github.com> Co-authored-by: Jim Boelrijk Valcon <jimboelrijk1@MacBook-Pro-van-Jim.local> Co-authored-by: Emilio Dorigatti <emilio.dorigatti@gmail.com>
experimental-design · Jan 28, 2025 · ac38bcc · ac38bcc
1 parent 4d6d6c6
commit ac38bcc
Show file tree

Hide file tree

Showing 25 changed files with 1,685 additions and 962 deletions.
diff --git a/bofire/data_models/strategies/actual_strategy_type.py b/bofire/data_models/strategies/actual_strategy_type.py
@@ -25,7 +25,6 @@
 )
 from bofire.data_models.strategies.random import RandomStrategy
 from bofire.data_models.strategies.shortest_path import ShortestPathStrategy
-from bofire.data_models.strategies.space_filling import SpaceFillingStrategy
 
 
 ActualStrategy = Union[
@@ -40,7 +39,6 @@
     QnehviStrategy,
     QparegoStrategy,
     EntingStrategy,
-    SpaceFillingStrategy,
     RandomStrategy,
     DoEStrategy,
     FactorialStrategy,

diff --git a/bofire/data_models/strategies/api.py b/bofire/data_models/strategies/api.py
@@ -1,7 +1,17 @@
 from typing import Union
 
 from bofire.data_models.strategies.actual_strategy_type import ActualStrategy
-from bofire.data_models.strategies.doe import DoEStrategy
+from bofire.data_models.strategies.doe import (
+    AnyDoEOptimalityCriterion,
+    AnyOptimalityCriterion,
+    AOptimalityCriterion,
+    DoEStrategy,
+    DOptimalityCriterion,
+    EOptimalityCriterion,
+    GOptimalityCriterion,
+    KOptimalityCriterion,
+    SpaceFillingCriterion,
+)
 from bofire.data_models.strategies.factorial import FactorialStrategy
 from bofire.data_models.strategies.fractional_factorial import (
     FractionalFactorialStrategy,
@@ -32,7 +42,6 @@
 )
 from bofire.data_models.strategies.random import RandomStrategy
 from bofire.data_models.strategies.shortest_path import ShortestPathStrategy
-from bofire.data_models.strategies.space_filling import SpaceFillingStrategy
 from bofire.data_models.strategies.stepwise.conditions import (
     AlwaysTrueCondition,
     AnyCondition,

diff --git a/bofire/data_models/strategies/doe.py b/bofire/data_models/strategies/doe.py
@@ -1,24 +1,99 @@
-from typing import Literal, Optional, Type, Union
+from typing import Annotated, Dict, Literal, Optional, Type, Union
 
+from formulaic import Formula
+from formulaic.errors import FormulaSyntaxError
+from pydantic import Field, field_validator
+
+from bofire.data_models.base import BaseModel
 from bofire.data_models.constraints.api import Constraint
 from bofire.data_models.features.api import Feature, MolecularInput
 from bofire.data_models.objectives.api import Objective
 from bofire.data_models.strategies.strategy import Strategy
 from bofire.data_models.types import Bounds
-from bofire.strategies.enum import OptimalityCriterionEnum
 
 
-class DoEStrategy(Strategy):
-    type: Literal["DoEStrategy"] = "DoEStrategy"
+PREDEFINED_MODEL_TYPES = Literal[
+    "linear",
+    "linear-and-quadratic",
+    "linear-and-interactions",
+    "fully-quadratic",
+]
+
+
+class OptimalityCriterion(BaseModel):
+    type: str
+    delta: float = 1e-6
+    transform_range: Optional[Bounds] = None
+
+
+class SpaceFillingCriterion(OptimalityCriterion):
+    type: Literal["SpaceFillingCriterion"] = "SpaceFillingCriterion"  # type: ignore
+    sampling_fraction: Annotated[float, Field(gt=0, lt=1)] = 0.3
+
+
+class DoEOptimalityCriterion(OptimalityCriterion):
+    type: str
     formula: Union[
-        Literal[
-            "linear",
-            "linear-and-quadratic",
-            "linear-and-interactions",
-            "fully-quadratic",
-        ],
+        PREDEFINED_MODEL_TYPES,
         str,
     ]
+    """
+    model_type (str, Formula): keyword or formulaic Formula describing the model. Known keywords
+    are "linear", "linear-and-interactions", "linear-and-quadratic", "fully-quadratic".
+    """
+
+    @field_validator("formula")
+    @classmethod
+    def validate_formula(cls, formula: str) -> str:
+        if formula not in PREDEFINED_MODEL_TYPES.__args__:  # type: ignore
+            # check that it is a valid formula
+            try:
+                Formula(formula)
+            except FormulaSyntaxError:
+                raise ValueError(f"Invalid formula: {formula}")
+        return formula
+
+
+class DOptimalityCriterion(DoEOptimalityCriterion):
+    type: Literal["DOptimalityCriterion"] = "DOptimalityCriterion"  # type: ignore
+
+
+class EOptimalityCriterion(DoEOptimalityCriterion):
+    type: Literal["EOptimalityCriterion"] = "EOptimalityCriterion"  # type: ignore
+
+
+class AOptimalityCriterion(DoEOptimalityCriterion):
+    type: Literal["AOptimalityCriterion"] = "AOptimalityCriterion"  # type: ignore
+
+
+class GOptimalityCriterion(DoEOptimalityCriterion):
+    type: Literal["GOptimalityCriterion"] = "GOptimalityCriterion"  # type: ignore
+
+
+class KOptimalityCriterion(DoEOptimalityCriterion):
+    type: Literal["KOptimalityCriterion"] = "KOptimalityCriterion"  # type: ignore
+
+
+AnyDoEOptimalityCriterion = Union[
+    KOptimalityCriterion,
+    GOptimalityCriterion,
+    AOptimalityCriterion,
+    EOptimalityCriterion,
+    DOptimalityCriterion,
+]
+
+AnyOptimalityCriterion = Union[
+    AnyDoEOptimalityCriterion,
+    SpaceFillingCriterion,
+]
+
+
+class DoEStrategy(Strategy):
+    type: Literal["DoEStrategy"] = "DoEStrategy"  # type: ignore
+
+    criterion: AnyOptimalityCriterion = Field(
+        default_factory=lambda: DOptimalityCriterion(formula="fully-quadratic")
+    )
     optimization_strategy: Literal[
         "default",
         "exhaustive",
@@ -28,11 +103,8 @@ class DoEStrategy(Strategy):
         "iterative",
     ] = "default"
 
-    verbose: bool = False
-
-    objective: OptimalityCriterionEnum = OptimalityCriterionEnum.D_OPTIMALITY
-
-    transform_range: Optional[Bounds] = None
+    verbose: bool = False  # get rid of this at a later stage
+    ipopt_options: Optional[Dict] = None
 
     @classmethod
     def is_constraint_implemented(cls, my_type: Type[Constraint]) -> bool:

diff --git a/bofire/data_models/strategies/space_filling.py b/bofire/data_models/strategies/space_filling.py
diff --git a/bofire/strategies/api.py b/bofire/strategies/api.py
@@ -17,6 +17,5 @@
 )
 from bofire.strategies.random import RandomStrategy
 from bofire.strategies.shortest_path import ShortestPathStrategy
-from bofire.strategies.space_filling import SpaceFillingStrategy
 from bofire.strategies.stepwise.stepwise import StepwiseStrategy
 from bofire.strategies.strategy import Strategy