Skip to content

Latest commit

 

History

History
105 lines (77 loc) · 3.96 KB

CHANGELOG.rst

File metadata and controls

105 lines (77 loc) · 3.96 KB

Release Changelog

0.8.0 (2022-05-06)

  • Replace poetry with poetry-core as a build dependency

0.8.0rc1 (2022-04-25)

  • Drop support for Python 3.7
  • Add support for Python 3.10
  • Add support for Windows
  • Update scikit-learn constraint to 1.0

0.7.0 (2021-12-08)

  • Drop support for Python 3.6
  • Fix quantile regression predictions for single record
  • Fix Tree.value for classifiers
  • Fix Tree.feature to use proper value for leaf nodes
  • Allow skranger predictors to work with shap using skranger.utils.shap.shap_patch context manager
  • Fix package includes to prevent installing extra files to site-packages

0.6.1 (2021-09-05)

  • Use oldest supported numpy for builds

0.6.0 (2021-08-23)

  • Remove numpy from dependency spec; numpy is already a requirement of scikit-learn
  • Change tree detail training code to be optional in ensembles due to expensive operations
  • Change quantile regression to use np.quantile in lieu of np.nanquantile` for faster predictions
  • Fix bug in tree classmethods when setting sample_fraction
  • Added more documentation around tree detail calculations

0.5.0 (2021-07-20)

  • Move split_select_weights, always_split_features, categorical_features params from init to fit methods.
  • Sample weight checking is now a base class method.
  • Remove sparse matrix args from bindings.
  • Fix a bug with the output of predict_quantiles not being oriented properly for multiple quantiles
  • Regression's predict_quantiles now requires a passed list of quantiles and the default is removed
  • Regression's predict now takes an optional list of quantiles
  • Remove snp_data and order_snps from bindings
  • Moves class_weights to fit in classifier, and changes the arg type to a dictionary.
  • Add RangerTreeClassifier, RangerTreeRegressor, and RangerTreeSurvival decision tree estimators which inherit between RangerMixin and BaseRangerTree. Also provide a BaseRangerForest class for ensemble estimators.
  • Add a low level Tree class which implements most of the sklearn.tree._tree.Tree interface.
  • Fix incorrect documentation for num_random_splits.

0.4.1 (2021-07-04)

  • Set an explicit lower bound on sklearn version.
  • Fix a bug with split_select_weights

0.4.0 (2021-04-23)

  • Add get_importance_pvalues method to estimators (thanks kmacdon)
  • Add feature_importances_ attribute, similar to sklearn forests
  • Ensure self.respect_categorical_features is unchanged when fitting by introducing self.respect_categorical_features_
  • Change self.n_features_ to self.n_features_in_
  • Add validation to classification targets, ensuring regression targets can't be passed to classifier
  • Add sample weight validation to ensure that passing weights of ones results in identical output when passing None. We do this because ranger does additional RNG on weighted sampling when non-null weights are passed.
  • Use self._validate_data in lieu of check_X_y when possible
  • Use self._check_n_features in lieu of manually setting n features
  • Add tags to estimators

0.3.2 (2021-01-18)

  • Fixed a bug related to incorrect sample_fraction input type
  • Fixed a bug in which sample_fraction was being passed on predict, raising a ranger error

0.3.1 (2020-12-05)

  • Fixed a bug with incorrect output for quantile regression

0.3.0 (2020-10-28)

  • Enable quantile regression on RangerForestRegressor.

0.2.0 (2020-10-23)

  • Fix bug in classifier to reorder probabilities properly using forest's class_values.

0.1.1 (2020-07-09)

  • Unpin sklearn/numpy deps, drop Cython dependency, since building wheels.
  • Use numpy pyobjects to improve compilation.

0.1.0 (2020-06-03)

  • First release.