v0.1.0

What's Changed

• Added support for sklearn pipelines in powershap as a model and appropriate testing, thanks to @eduardokapp: #33
• Fixed kwargs not propagating properly to the fit of the model: #53
• Added support for Python 3.13 and appropriate testing: #45
• Solved bug that resulted in an error when the solve_power could not converge, which resulted in returning a list instead of a float, which is not supported anymore by newer numpy versions: #46
• Solved issue with Sklearn 1.6, resulting in breaking code due to an internal change to their tag system and validate_data. The minimum sklearn version is now 1.6 as a result: #54
• Removed support for tensorflow due to stopped support by them for Python 3.10 on Windows (the resulting DeepLearningExplainer does not work as a result, Pytorch is on the roadmap)
• Migrated to uv from poetry
• Added testing for Windows
• Updated testing

New Contributors

• @eduardokapp made their first contribution in #33
• @JarneVerhaeghe

Full Changelog: v0.0.11...v0.1.0

v0.0.11

What's Changed

• enable validation set for xgboost's fit by @mrzdev in #34
• Correct use of random_state for xgboost by @lorenjan in #37
• Reduce memory usage by removing data copies by @lorenjan in #38
• relaxed pandas version constraints, so it works with 2.0 by @almostintuitive in #40
• 🙏 update testing by @jonasvdd in #48
• 💨 updating dependencies - python 3.12 compatible by @jonasvdd in #49
• Fix wrong aggregation of shap values by @jvdd in #52

New Contributors

• @mrzdev made their first contribution in #34
• @lorenjan made their first contribution in #37
• @almostintuitive made their first contribution in #40
• @jonasvdd made their first contribution in #48

Full Changelog: v0.0.9...v0.0.11

v0.0.9

What's Changed

• 🐛 lgbm test bugfix for categorical data by @jvdd in #16
• ✨ cv support by @jvdd in #17
• 🤖 update dependencies + new release by @jvdd in #18
• 🧹 avoid missing package errors by @jvdd in #20
• ⬆️ update dependencies by @jvdd in #22
• ♻️ update poetry install in ci by @jvdd in #27
• ✨ support xgboost + 📦 general maintenance by @jvdd in #30

Full Changelog: powershap-v0.0.5.1...v0.0.9

powershap v0.0.5.1

• Bug fixes
• Added stratify and group support for train-test split
• Added more model support
• Added Python 3.7 support
• Optimized the algorithm

powershap v0.0.2

The first public release of PowerShap.