0.12.2

What's Changed

Changed

• More robust settings for the GP fitting

Merged Pull Requests

• Improve GP Fit by @Scienfitz in #472

Full Changelog: 0.12.1...0.12.2

0.12.1

Hotfix Release

• Reset flag default value by @AdrianSosic in #471

Full Changelog: 0.12.0...0.12.1

0.12.0

What's Changed

Breaking Changes

• Lookup callables for simulation are now expected to accept/return dataframes with
  the corresponding parameter/target column labels

Added

• SHAP explanations via the new SHAPInsight class
• Optional insights dependency group
• Insights user guide
• Example for a traditional mixture
• allow_missing and allow_extra keyword arguments to Objective.transform
• add_noise_to_perturb_degenerate_rows utility
• benchmarks subpackage for defining and running performance tests
• Campaign.toggle_discrete_candidates to dynamically in-/exclude discrete candidates
• filter_df utility for filtering dataframe content
• arrays_to_dataframes decorator to create lookups from array-based callables
• DiscreteConstraint.get_valid to conveniently access valid candidates
• Functionality for persisting benchmarking results on S3 from a manual pipeline run
• remain_switched option to TwoPhaseMetaRecommender
• is_stateful class variable to MetaRecommender
• get_non_meta_recommender method to MetaRecommender

Changed

• SubstanceParameter encodings are now computed exclusively with the
  scikit-fingerprints package, granting access to all fingerprints available therein
• Example for slot-based mixtures has been revised and grouped together with the new
  traditional mixture example
• Memory caching is now non-verbose
• CustomDiscreteParameter does not allow duplicated rows in data anymore
• De-/activating Polars via BAYBE_DEACTIVATE_POLARS now requires passing values
  compatible with strtobool
• All arguments to MetaRecommender.select_recommender are now optional
• MetaRecommenders can now be composed of other MetaRecommenders
• For performance reasons, search space manipulation using polars is no longer
  guaranteed to produce the same row order as the corresponding pandas operations
• allow_repeated_recommendations has been renamed to
  allow_recommending_already_recommended and is now True by default

Fixed

• Rare bug arising from degenerate SubstanceParameter.comp_df rows that caused
  wrong number of recommendations being returned
• ContinuousConstraints can now be used in single point precision
• Search spaces are now stateless, preventing unintended side effects that could lead to
  incorrect candidate sets when reused in different optimization contexts
• qNIPV not working with single MIN targets
• Passing a TargetTransformation without passing bounds when creating a
  NumericalTarget now raises an error
• Crash when using ContinuousCardinalityConstraint caused by an unintended interplay
  between constraints and dropped parameters yielding empty parameter sets
• Minimizing a single NumericalTarget with specified bounds/transformation via
  SingleTargetObjective no longer erroneously maximizes it
• allow_* flags are now context-aware, i.e. setting them in a context where they are
  irrelevant now raises an error instead of passing silently

Removed

• botorch_function_wrapper utility for creating lookup callables

Deprecations

• Passing a dataframe via the data argument to Objective.transform is no longer
  possible. The dataframe must now be passed as positional argument.
• The new allow_extra flag is automatically set to True in Objective.transform
  when left unspecified
• get_transform_parameters has been replaced with get_transform_objects
• Passing a dataframe via the data argument to Target.transform is no longer
  possible. The data must now be passed as a series as first positional argument.
• SubstanceEncoding value MORGAN_FP. As a replacement, ECFP with 1024 bits and
  radius of 4 can be used.
• SubstanceEncoding value RDKIT. As a replacement, RDKIT2DDESCRIPTORS can be used.
• The metadata attribute of SubspaceDiscrete no longer exists. Metadata is now
  exclusively handled by the Campaign class.
• get_current_recommender and get_next_recommender of MetaRecommender have become
  obsolete and calling them is no longer possible
• Passing allow_* flags to recommenders is no longer supported since the necessary
  metadata required for the flags is no longer available at that level. The
  functionality has been taken over by Campaign.

Merged Pull Requests

• Fix math in CardinalityConstraint docstring by @AVHopp in #405
• Typing overrides by @AdrianSosic in #400
• Bump Streamlit Requirement by @Scienfitz in #407
• Refactor objective transform interface by @AdrianSosic in #398
• Add Polars installation issues to known_issues by @AVHopp in #404
• Refactor target transform interface by @AdrianSosic in #406
• Drop batch_quantity keyword by @AdrianSosic in #411
• Add downloads badge by @AdrianSosic in #418
• Unpin protobuf by @Scienfitz in #421
• Fix some typos by @AdrianSosic in #426
• Fix Asserts in Dependency Constraint Example by @Scienfitz in #422
• Rework Mixture Examples by @Scienfitz in #394
• Use correct precision in botorch wrapper by @AVHopp in #417
• Enable scikit-fingerprints by @Hrovatin in #364
• Fix rare degenerate comp_df bug by @Scienfitz in #399
• Basic benchmarking structure by @fabianliebig in #413
• Require truthy/falsy values for BAYBE_DEACTIVATE_POLARS by @AdrianSosic in #427
• Fix dtype in continuous comp rep by @AVHopp in #415
• Drop CI audit dependence by @AdrianSosic in #434
• Immutable searchspace by @AdrianSosic in #412
• Fix: qNIPV not working with single MIN target by @Scienfitz in #435
• Add manual benchmark workflow and S3 result persistence by @fabianliebig in #429
• Raise error when specifying transformation without bounds by @AVHopp in #451
• Fix empty constraints after dropping parameters by @AVHopp in #448
• Include robust check for extracting column names for fingerprints by @AVHopp in #447
• Correct typo in docs on acquisition functions by @Hrovatin in #442
• Refactor callable lookup mechanism by @AdrianSosic in #441
• Pin version of scipy to <1.15 by @AVHopp in #458
• Hotfix: Minimization with bounds properly minimize targets by @AVHopp in #462
• Insights module and SHAPInsight by @Alex6022 in #391
• Simplify and extend meta recommender logic by @AdrianSosic in #457
• Fix minimization of improvement-based MC acquisition functions by @AdrianSosic in #465
• Maintain row order after cross join by @AdrianSosic in #463
• Move allow_* flags to Campaign by @AdrianSosic in #423
• Insights User Guide by @Scienfitz in #466
• Include 0.11.4 Hotfix Changes in main by @Scienfitz in #469
• Release 0.12.0 by @Scienfitz in #470

New Contributors

• @Hrovatin made their first contribution in #364
• @fabianliebig made their first contribution in #413
• @Alex6022 made their first contribution in #391

Full Changelog: 0.11.2...0.12.0

0.11.4

Hotfix Release

Changed

• Polars lazy streaming has been deactivated due to instabilities

Fixed

• Improvement-based Monte Carlo acquisition functions now use the correct
  reference value for single-target minimization

Full Changelog: 0.11.3...0.11.4

0.11.3

Hotfix Release

Fixed

• protobuf dependency issue, version pin was removed

Full Changelog: 0.11.2...0.11.3

0.11.2

What's Changed

Added

• n_restarts and n_raw_samples keywords to configure continuous optimization
  behavior for BotorchRecommender
• User guide for utilities

Changed

• Utility add_fake_results renamed to add_fake_measurements
• Utilities add_fake_measurements and add_parameter_noise now also return the
  dataframe they modified in-place

Fixed

• Leftover attrs-decorated classes are garbage collected before the subclass tree is
  traversed, avoiding sporadic serialization problems

Merged Pull Requests

• Subclass garbage collection by @AdrianSosic in #397
• Update known_issues.md by @Scienfitz in #390
• Update CONTRIBUTING.md by @Scienfitz in #395
• Add Continuous Optimization Controls to BotorchRecommender by @Scienfitz in #389
• Add Explanation of some Utilities to User Guide by @Scienfitz in #392
• Release 0.11.2 by @Scienfitz in #401

Full Changelog: 0.11.1...0.11.2

0.11.1

What's Changed

Added

• Continuous linear constraints have been consolidated in the new
  ContinuousLinearConstraint class

Changed

• get_surrogate now also returns the model for transformed single targets or
  desirability objectives

Fixed

• Unsafe name-based matching of columns in get_comp_rep_parameter_indices

Deprecated

• ContinuousLinearEqualityConstraint and ContinuousLinearInequalityConstraint
  replaced by ContinuousLinearConstraint with the corresponding operator keyword

Merged Pull Requests

• Speedup Bernoulli Multi-Armed Bandit Example by @julianStreibel in #380
• Fix intersphinx link for torch by @AVHopp in #382
• Fix Name Matching for Retrieving Comp Rep Indices by @Scienfitz in #379
• Remove get_surrogate Restrictions by @Scienfitz in #386
• Consolidate Continuous Linear Constraints by @Scienfitz in #384
• Release 0.11.1 by @Scienfitz in #388

Full Changelog: 0.11.0...0.11.1

0.11.0

What's Changed

Breaking Changes

• The public methods of Surrogate models now operate on dataframes in experimental
  representation instead of tensors in computational representation
• Surrogate.posterior models now returns a Posterior object
• param_bounds_comp of SearchSpace, SubspaceDiscrete and SubspaceContinuous has
  been replaced with comp_rep_bounds, which returns a dataframe

Added

• py.typed file to enable the use of type checkers on the user side
• IndependentGaussianSurrogate base class for surrogate models providing independent
  Gaussian posteriors for all candidates (cannot be used for batch prediction)
• comp_rep_columns property for Parameter, SearchSpace, SubspaceDiscrete
  and SubspaceContinuous classes
• New mechanisms for surrogate input/output scaling configurable per class
• SurrogateProtocol as an interface for user-defined surrogate architectures
• Support for binary targets via BinaryTarget class
• Support for bandit optimization via BetaBernoulliMultiArmedBanditSurrogate class
• Bandit optimization example
• qThompsonSampling acquisition function
• BetaPrior class
• recommend now accepts the pending_experiments argument, informing the algorithm
  about points that were already selected for evaluation
• Pure recommenders now have the allow_recommending_pending_experiments flag,
  controlling whether pending experiments are excluded from candidates in purely
  discrete search spaces
• get_surrogate and posterior methods to Campaign
• tenacity test dependency
• Multi-version documentation

Changed

• The transition from experimental to computational representation no longer happens
  in the recommender but in the surrogate
• Fallback models created by catch_constant_targets are stored outside the surrogate
• to_tensor now also handles numpy arrays
• MIN mode of NumericalTarget is now implemented via the acquisition function
  instead of negating the computational representation
• Search spaces now store their parameters in alphabetical order by name
• Improvement-based acquisition functions now consider the maximum posterior mean
  instead of the maximum noisy measurement as reference value
• Iteration tests now attempt up to 5 repeated executions if they fail due to numerical
  reasons

Fixed

• CategoricalParameter and TaskParameter no longer incorrectly coerce a single
  string input to categories/tasks
• farthest_point_sampling no longer depends on the provided point order
• Batch predictions for RandomForestSurrogate
• Surrogates providing only marginal posterior information can no longer be used for
  batch recommendation
• SearchSpace.from_dataframe now creates a proper empty discrete subspace without
  index when called with continuous parameters only
• Metadata updates are now only triggered when a discrete subspace is present
• Unintended reordering of discrete search space parts for recommendations obtained
  with BotorchRecommender

Removed

• register_custom_architecture decorator
• Scalar and DefaultScaler classes

Deprecations

• The role of register_custom_architecture has been taken over by
  baybe.surrogates.base.SurrogateProtocol
• BayesianRecommender.surrogate_model has been replaced with get_surrogate

Merged Pull Requests

• Comp Rep Transition Point by @AdrianSosic in #278
• Surrogate Posteriors by @AdrianSosic in #309
• Surrogate scaling by @AdrianSosic in #315
• Surrogate interface by @AdrianSosic in #325
• Mypy for surrogates by @AdrianSosic in #337
• Add py.typed by @AdrianSosic in #342
• Lockfile for doc generation by @AdrianSosic in #323
• Fix parameter value conversion by @AdrianSosic in #259
• Improve FPS by @AdrianSosic in #336
• Refactor Surrogates by @AdrianSosic in #338
• Silence mypy type-abstract errors by @AdrianSosic in #354
• Hotfix: Remove python version in tox docs calls by @AVHopp in #351
• Fix local doc building by @AVHopp in #339
• MIN mode via acquisition function by @AdrianSosic in #340
• Fix batch recommendation by @AdrianSosic in #348
• Remove token from publishing pipelines by @Scienfitz in #361
• Parameter permutation by @AdrianSosic in #341
• Add workflow permissions by @Scienfitz in #362
• Best_f from posterior means by @AdrianSosic in #350
• Fix SubspaceDiscrete.from_dataframe edge case by @AdrianSosic in #360
• Multi-Armed Bandit by @julianStreibel in #343
• Fix visuals in examples by @AVHopp in #347
• Enable Pending Points by @Scienfitz in #319
• Fix Recommendation Order for BotorchRecommender by @Scienfitz in #370
• Expose surrogate by @AdrianSosic in #355
• Add iteration test retries by @Scienfitz in #374
• Multi-version documentation by @AVHopp in #317
• Fix broken links to stable by @AVHopp in #377
• Release 0.11.0 by @Scienfitz in #376

New Contributors

• @julianStreibel made their first contribution in #343

Full Changelog: 0.10.0...0.11.0

0.10.0

What's Changed

Breaking Changes

• Providing an explicit batch_size is now mandatory when asking for recommendations
• RecommenderProtocol.recommend now accepts an optional Objective
• RecommenderProtocol.recommend now expects training data to be provided as a single
  dataframe in experimental representation instead of two separate dataframes in
  computational representation
• Parameter.is_numeric has been replaced with Parameter.is_numerical
• DiscreteParameter.transform_rep_exp2comp has been replaced with
  DiscreteParameter.transform
• filter_attributes has been replaced with match_attributes

Added

• Surrogate base class now exposes a to_botorch method
• SubspaceDiscrete.to_searchspace and SubspaceContinuous.to_searchspace
  convenience constructor
• Validators for Campaign attributes
• _optional subpackage for managing optional dependencies
• New acquisition functions for active learning: qNIPV (negative integrated posterior
  variance) and PSTD (posterior standard deviation)
• Acquisition function: qKG (knowledge gradient)
• Abstract ContinuousNonlinearConstraint class
• Abstract CardinalityConstraint class and
  DiscreteCardinalityConstraint/ContinuousCardinalityConstraint subclasses
• Uniform sampling mechanism for continuous spaces with cardinality constraints
• register_hooks utility enabling user-defined augmentation of arbitrary callables
• transform methods of SearchSpace, SubspaceDiscrete and SubspaceContinuous
  now take additional allow_missing and allow_extra keyword arguments
• More details to the transfer learning user guide
• Activated doctests
• SubspaceDiscrete.from_parameter, SubspaceContinuous.from_parameter,
  SubspaceContinuous.from_product and SearchSpace.from_parameter
  convenience constructors
• DiscreteParameter.to_subspace, ContinuousParameter.to_subspace and
  Parameter.to_searchspace convenience constructors
• Utilities for permutation and dependency data augmentation
• Validation and translation tests for kernels
• BasicKernel and CompositeKernel base classes
• Activated pre-commit.ci with auto-update
• User guide for active learning
• Polars expressions for DiscreteSumConstraint, DiscreteProductConstraint,
  DiscreteExcludeConstraint, DiscreteLinkedParametersConstraint and
  DiscreteNoLabelDuplicatesConstraint
• Discrete search space Cartesian product can be created lazily via Polars
• Examples demonstrating the register_hooks utility: basic registration mechanism,
  monitoring the probability of improvement, and automatic campaign stopping

Changed

• Passing an Objective to Campaign is now optional
• GaussianProcessSurrogate models are no longer wrapped when cast to BoTorch
• Restrict upper versions of main dependencies, motivated by major numpy release
• Sampling methods in qNIPV and BotorchRecommender are now specified via
  DiscreteSamplingMethod enum
• Interval class now supports degenerate intervals containing only one element
• add_fake_results now directly processes Target objects instead of a Campaign
• path argument in plotting utility is now optional and defaults to Path(".")
• UnusedObjectWarning by non-predictive recommenders is now ignored during simulations
• The default kernel factory now avoids strong jumps by linearly interpolating between
  two fixed low and high dimensional prior regimes
• The previous default kernel factory has been renamed to EDBOKernelFactory and now
  fully reflects the original logic
• The default acquisition function has been changed from qEI to qLogEI for improved
  numerical stability

Removed

• Support for Python 3.9 removed due to new BoTorch requirements
  and guidelines from Scientific Python
• Linter typos for spellchecking

Fixed

• sequential flag of SequentialGreedyRecommender is now set to True
• Serialization bug related to class layout of SKLearnClusteringRecommender
• MetaRecommenders no longer trigger warnings about non-empty objectives or
  measurements when calling a NonPredictiveRecommender
• Bug introduced in 0.9.0 (PR #221, commit 3078f3), where arguments to to_gpytorch
  are not passed on to the GPyTorch kernels
• Positive-valued kernel attributes are now correctly handled by validators
  and hypothesis strategies
• As a temporary workaround to compensate for missing IndexKernel priors,
  fit_gpytorch_mll_torch is used instead of fit_gpytorch_mll when a TaskParameter
  is present, which acts as regularization via early stopping during model fitting

Deprecations

• SequentialGreedyRecommender class replaced with BotorchRecommender
• SubspaceContinuous.samples_random has been replaced with
  SubspaceContinuous.sample_uniform
• SubspaceContinuous.samples_full_factorial has been replaced with
  SubspaceContinuous.sample_from_full_factorial
• Passing a dataframe via the data argument to the transform methods of
  SearchSpace, SubspaceDiscrete and SubspaceContinuous is no longer possible.
  The dataframe must now be passed as positional argument.
• The new allow_extra flag is automatically set to True in transform methods
  of search space classes when left unspecified

Expired Deprecations (from 0.7.*)

• Interval.is_finite property
• Specifying target configs without type information
• Specifying parameters/constraints at the top level of a campaign configs
• Passing numerical_measurements_must_be_within_tolerance to Campaign
• batch_quantity argument
• Passing allow_repeated_recommendations or allow_recommending_already_measured
  to MetaRecommender (or former Strategy)
• *Strategy classes and baybe.strategies subpackage
• Specifying MetaRecommender (or former Strategy) configs without type information

Merged Pull Requests

• Drop Python 3.9 by @AdrianSosic in #253
• Refactor recommender signature by @AdrianSosic in #220
• Python 3.10 syntax upgrade by @AdrianSosic in #255
• Fix changelog by @AdrianSosic in #261
• Refactor adapter model by @AdrianSosic in #260
• Deprecate SequentialGreedyRecommender by @AdrianSosic in #256
• Campaign validators by @AdrianSosic in #262
• Refactor optional imports by @AdrianSosic in #266
• Add upper version limits by @AdrianSosic in #271
• Add qNIPV acqf by @Scienfitz in #264
• Fix optional dependency flags by @AdrianSosic in #272
• Cardinality constraints for random recommender by @Waschenbacher in #243
• Register hook with the same signature by @RimRihana in #275
• register_hooks utility by @AdrianSosic in #276
• Remove core test caching by @Scienfitz in #277
• Add CI dev branch triggers by @Scienfitz in #279
• Fix attrs incompatibility of pydoclint by @AdrianSosic in #282
• Replace is_numeric class attribute with is_numerical by @AdrianSosic in #281
• Add discrete cardinality constraint by @Waschenbacher in #270
• Remove tox caching in Regular by @Scienfitz in #293
• Fix serialization and caching by @AdrianSosic in #298
• Refactor search space transform methods by @AdrianSosic in #289
• Fix MetaRecommender Warnings by @Scienfitz in #292
• Fix Deprecated Calls by @Scienfitz in #302
• Polars pipeline for discrete sum, product and exclusion constraints by @rjavadi in #252
• Update TL User Guide by @Scienfitz in #299
• Reorganize simulation lookup by @AdrianSosic in #265
• Change path from sys.path[0] to . by @AVHopp in #288
• Polars: Translate linked parameter and no label duplicate to polars by @rjavadi in #305
• Searchspace convenience constructors by @AdrianSosic in #303
• Add KnowledgeGradient ACQF by @Scienfitz in #308
• Augmentation Utilities by @Scienfitz in #290
• Fix kernel translation by @AdrianSosic in #311
• Upgrade dev tools by @AdrianSosic in #316
• Polars: from_product() implementation by @rjavadi in #306
• Add Active Learning Userguide by @Scienfitz in #273
• Ignore UnusedObjectWarning caused by nonpred recommenders in simulation by @Scienfitz in #313
• Temp Fix: Avoid Prior Jumps by @Scienfitz in #310
• Restore old MLL fit by @Scienfitz in #320
• Fix broken docstring links by @AVHopp in #324
• Custom Hooks: PI Example by @RimRihana in #287
• S...

0.9.1

What's Changed

Changed

• Discrete searchspace memory estimate is now natively represented in bytes

Fixed

• Non-GP surrogates not working with deepcopy and the simulation package due to
  slotted base class
• Datatype inconsistencies for various parameters' values and comp_df and
  SubSelectionCondition's selection related to floating point precision

Merged Pull Requests

• Fix wrong picture in README by @AVHopp in #248
• Fix surrogate slots by @Scienfitz in #251
• Refine MemorySize class by @Scienfitz in #249
• Separate tox commands for doc building by @AVHopp in #234
• Fix dtypes related to floating point precision by @Scienfitz in #254
• Release 0.9.1 by @AdrianSosic in #258

Full Changelog: 0.9.0...0.9.1