v0.2.6

What's Changed

New Features 🎉

• Allow method="basinhopping" in find_MAP and fit_laplace by @jessegrabowski in #467
• Expose multivariate normal method argument in post-estimation tasks by @jessegrabowski in #484

Bugfixes 🪲

• Fix bug in fit_MAP when shared variables are used in graph by @jessegrabowski in #468
• Update Test CI by @jessegrabowski in #474

Documentation 📖

• Small typos corrections in notebooks SARMA and VARMAX examples by @andreacate in #440

Maintenance 🔧

• clean up project root by @theorashid in #370
• Delete buildosx from project root by @jessegrabowski in #471
• theme update and doc fixes by @OriolAbril in #473
• Pr 451 - modified and added tests to statespace by @Dekermanjian in #466
• Remove mode argument passed from Statespace.build_statespace_graph to scan by @jessegrabowski in #482

New Contributors

• @Dekermanjian made their first contribution in #466
• @andreacate made their first contribution in #440

Full Changelog: v0.2.5...v0.2.6

v0.2.5

What's Changed

Bugfixes 🪲

• Fix bug in fit_laplace when model has exactly one variable by @jessegrabowski in #437

Maintenance 🔧

• Make index check on statespace data less strict by @jessegrabowski in #434
• add python 3.13 to the list by @williambdean in #436
• Minor fix of blackjax import in fit_pathfinder function by @aphc14 in #443
• Filter warning for batched_dot until we change it by @zaxtax in #455
• Fix failed test related to PR #443 by replacing pt.batched_dot with pt.vectorize(pt.dot,...) by @aphc14 in #453
• Progress bar fix for PF by @fonnesbeck in #457
• Make fit_pathfinder more similar to fit_laplace and pm.sample by @velochy in #447
• Add regularisation term before Cholesky decomposition by @aphc14 in #459
• Fix LBFGS iteration conditions and status handling by @aphc14 in #461
• Refactor alpha_recover and inverse_hessian_factors to remove update_m… by @aphc14 in #462
• Added a PartialOrder transform by @velochy in #444

New Contributors

• @velochy made their first contribution in #447

Full Changelog: v0.2.4...v0.2.5

v0.2.4

What's Changed

Bugfixes 🪲

• Bump PyMC version requirement by @jessegrabowski in #431

Maintenance 🔧

• Fix pathfinder progress bar display not completing by @aphc14 in #420
• Fix: importance_sampling=None produces error by @aphc14 in #427
• Remove MvNormalSVD Class by @jessegrabowski in #432

Full Changelog: v0.2.3...v0.2.4

v0.2.3

What's Changed

Bugfixes 🪲

• Bugfixes and Improvements for MAP and Laplace by @jessegrabowski in #422

Maintenance 🔧

• Use a valid SPDX identifier for the license by @maresb in #419
• Fix minor typo in Making a Custom Statespace Model.ipynb. by @niskrev in #417
• Bump version by @jessegrabowski in #423

New Contributors

• @niskrev made their first contribution in #417

Full Changelog: v0.2.2...v0.2.3

v0.2.2

What's Changed

New Features 🎉

• Use rich tables for statespace build reports by @jessegrabowski in #411
• PyMC/PyTensor Implementation of Pathfinder VI by @aphc14 in #387

Bugfixes 🪲

• Check model coords for unknown shapes when building predictive models by @jessegrabowski in #413

Maintenance 🔧

• as_model wraps function arguments in Data if they support it. by @zaxtax in #414
• Release version 0.2.2 by @fonnesbeck in #418

New Contributors

• @aphc14 made their first contribution in #387

Full Changelog: v0.2.1...v0.2.2

v0.2.1

What's Changed

Major Changes 🛠

• Rename to pymc-extras by @fonnesbeck in #404
• Implement marginalization as a model transformation and deprecate MarginalModel by @ricardoV94 in #388

Maintenance 🔧

• Fix bug in TimeSeasonality when pop_state = False by @jessegrabowski in #407
• Adapt the GHA PyPI publishing workflow for trusted publishing by @maresb in #409

New Contributors

• @maresb made their first contribution in #403
• @fonnesbeck made their first contribution in #404

Full Changelog: v0.1.4...v0.2.1

v0.1.4

What's Changed

New Features 🎉

• Expand models supported by automatic marginalization by @ricardoV94 in #300
• Summarize model as rich table by @ricardoV94 in #382
• Add find_MAP with close JAX integration and fix bug with Laplace fit by @jessegrabowski in #385

Documentation 📖

• Add PyPI details to homepage by @wd60622 in #390

Maintenance 🔧

• Fix Blackjax SMC integration by @ciguaran in #374
• Add ETS model by @jessegrabowski in #369
• Allow forecasting with exogenous variables by @jessegrabowski in #372
• Remove dead code from kalman_filter.py by @jessegrabowski in #393

Full Changelog: v0.1.3...v0.1.4

v0.1.3

What's Changed

New Features 🎉

• Specialized DiscreteMarkovChain step sampler by @ricardoV94 in #359

Full Changelog: v0.1.2...v0.1.3

v0.1.2

What's Changed

Major Changes 🛠

• More bugfixes for statespace by @jessegrabowski in #346

New Features 🎉

• Implement Laplace (quadratic) approximation by @carsten-j in #345
• Add Exponential distribution to model/transforms/autoreparam.py by @ferrine in #365

Maintenance 🔧

• Move tests/ to project root by @jessegrabowski in #361
• Fix autoreparam because dims are no longer static by @ferrine in #363
• Update pre-commit by @jessegrabowski in #368

New Contributors

• @carsten-j made their first contribution in #345
• @lucianopaz made their first contribution in #377

Full Changelog: v0.1.1...v0.1.2

v0.1.1

What's Changed

New Features 🎉

• Use vectorize in finite_discrete_marginal_rv_logp by @ricardoV94 in #337
• Allow creating MarginalModel from existing Model by @ricardoV94 in #333

Maintenance 🔧

• Delete warnings for graduated functionality by @ricardoV94 in #334

Full Changelog: v0.1.0...v0.1.1