DoWhy Files

• [New] Doubly robust estimator [Example Notebook]
• [New] DecisionTreeClassifier as an option in the set of 'GOOD' gcm classifiers
• Improved calibration of auto-selected classifier models in gcm
• Performance improvements in causal prediction algorithm
• Python 3.13 support, can be installed with uv

A big thank you to @amit-sharma @bloebp @emmanuel-ferdman @JPZ4-5 @nparent1 for contributing to this release. More details below.

What's Changed

• Add Python 3.13 support and update dependencies by @bloebp in https://github.com/py-why/dowhy/pull/1328
• Correct vertical stacking in linearity check by @emmanuel-ferdman in https://github.com/py-why/dowhy/pull/1336
• Fix CUDA errors in conditional regularizer by @JPZ4-5 in https://github.com/py-why/dowhy/pull/1341
• fixed pandas futurewarning errors by @amit-sharma in https://github.com/py-why/dowhy/pull/1339
• feat: use torch.cdist to compute Gaussian kernel for performance improvement by @JPZ4-5 in https://github.com/py-why/dowhy/pull/1358
• Doubly Robust Estimator by @nparent1 in https://github.com/py-why/dowhy/pull/1346
• Revise GCM classifier model selection logic by @bloebp in https://github.com/py-why/dowhy/pull/1349
• Enabling install using uv by @amit-sharma in https://github.com/py-why/dowhy/pull/1362

New Contributors

• @emmanuel-ferdman made their first contribution in https://github.com/py-why/dowhy/pull/1336
• @JPZ4-5 made their first contribution in https://github.com/py-why/dowhy/pull/1341

Full Changelog: https://github.com/py-why/dowhy/compare/v0.13...v0.14