Poniard

Latest version: v0.5.0

Safety actively analyzes 623106 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 1 of 2

0.4.0

0.3.0

What's Changed
* Migrate to nbdev by rxavier in https://github.com/rxavier/poniard/pull/29 & https://github.com/rxavier/poniard/pull/30
* New documentation at https://rxavier.github.io/poniard/
* Better handling of removed and added estimators (no longer readded if pipelines are rebuilt)
* Fix `reassign_types` for non-modified types.
* Fix for wandb config logging
* `TargetEncoder` now handles non-numeric and non-boolean classification targets.

**Full Changelog**: https://github.com/rxavier/poniard/compare/0.2.4...0.3.0

0.2.4

What's Changed
* Reuse predictions for plots in https://github.com/rxavier/poniard/pull/20
* Light refactor in https://github.com/rxavier/poniard/pull/23
* Add more regression plots in https://github.com/rxavier/poniard/pull/24
* Assorted fixes in https://github.com/rxavier/poniard/pull/25


**Full Changelog**: https://github.com/rxavier/poniard/compare/0.2.3...0.2.4

0.2.3

What's Changed
* Add artifact instead of direct logging by rxavier in https://github.com/rxavier/poniard/pull/19


**Full Changelog**: https://github.com/rxavier/poniard/compare/0.2.2...0.2.3

0.2.2

What's Changed
* Properly address multilabel and multioutput targets by rxavier in https://github.com/rxavier/poniard/pull/15
* Plugins refactor by rxavier in https://github.com/rxavier/poniard/pull/18

**Full Changelog**: https://github.com/rxavier/poniard/compare/0.2.1...0.2.2

0.2.1

What's Changed
- Allow adding preprocessing steps to existing preprocessor and add VarianceTreshold by rxavier in https://github.com/rxavier/poniard/pull/9
- Implements `add_preprocessing_step`
- Removes transformers with no assigned features from the preprocessor
- Include a VarianceThreshold transformer at the end of the default preprocessing pipeline
- Don't use TargetEncoder if the target is multi output, switch to OrdinalEncoder instead
- Set TargetEncoder to ignore unseen categories - will use the mean of the target
- Improve TargetEncoder docs
- Clear the `_fitted_estimator_ids` after adding a preprocessing step or reassigning types (which triggers a rebuild of the preprocessor)


**Full Changelog**: https://github.com/rxavier/poniard/compare/0.2.0...0.2.1

Page 1 of 2

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.