aplr 7.0.0__cp311-cp311-win_amd64.whl → 10.20.0__cp311-cp311-win_amd64.whl

aplr/aplr.py

@@ -1,49 +1,134 @@
+from typing import List, Callable, Optional, Dict, Union
 import numpy as np
-import numpy.typing as npt
-from typing import List, Callable, Optional
+import pandas as pd
 import aplr_cpp
+import itertools
+
+FloatVector = np.ndarray
+FloatMatrix = np.ndarray
+IntVector = np.ndarray
+IntMatrix = np.ndarray
+
+
+def _dataframe_to_cpp_dataframe(df: pd.DataFrame) -> aplr_cpp.CppDataFrame:
+    """Converts a pandas DataFrame to a CppDataFrame."""
+    cpp_df = aplr_cpp.CppDataFrame()
+    for col_name in df.columns:
+        col = df[col_name]
+        if pd.api.types.is_numeric_dtype(col.dtype):
+            # Convert numeric columns to std::vector<double>
+            # NaNs are preserved and handled in C++
+            cpp_df.add_numeric_column(
+                col_name, col.to_numpy(dtype=np.float64, na_value=np.nan)
+            )
+        elif (
+            isinstance(col.dtype, pd.CategoricalDtype)
+            or pd.api.types.is_object_dtype(col.dtype)
+            or pd.api.types.is_string_dtype(col.dtype)
+        ):
+            # Convert categorical/object/string columns to std::vector<std::string>
+            # Missing values (None, np.nan) are converted to empty strings for C++ handling
+            cpp_df.add_categorical_column(col_name, col.astype(str).fillna("").tolist())
+        else:
+            raise TypeError(
+                f"Unsupported column type for column '{col_name}': {col.dtype}"
+            )
+    return cpp_df
+
+
+def _prepare_input_data(
+    X: Union[pd.DataFrame, FloatMatrix], preprocess: bool
+) -> Union[aplr_cpp.CppDataFrame, FloatMatrix]:
+    """
+    Prepares the input data for the C++ backend.
+
+    If X is a pandas DataFrame, it's converted. If preprocess is True, it becomes
+    a CppDataFrame. If preprocess is False, it's converted to a NumPy array.
+    NumPy arrays are passed through as is.
+    """
+    if isinstance(X, pd.DataFrame):
+        if preprocess:
+            return _dataframe_to_cpp_dataframe(X)
+        else:
+            # Check if all columns are numeric before converting
+            if not all(pd.api.types.is_numeric_dtype(X[col]) for col in X.columns):
+                raise RuntimeError(
+                    "Cannot convert DataFrame to matrix if it contains non-numeric columns. "
+                    "Please ensure all columns are numeric or set preprocess=True."
+                )
+            return X.to_numpy(dtype=np.float64)
+    return X
 
 
 class APLRRegressor:
     def __init__(
         self,
-        m: int = 1000,
-        v: float = 0.1,
+        m: int = 3000,
+        v: float = 0.5,
         random_state: int = 0,
         loss_function: str = "mse",
         link_function: str = "identity",
         n_jobs: int = 0,
-        validation_ratio: float = 0.2,
+        cv_folds: int = 5,
         bins: int = 300,
         max_interaction_level: int = 1,
         max_interactions: int = 100000,
-        min_observations_in_split: int = 20,
-        ineligible_boosting_steps_added: int = 10,
-        max_eligible_terms: int = 5,
+        min_observations_in_split: int = 4,
+        ineligible_boosting_steps_added: int = 15,
+        max_eligible_terms: int = 7,
         verbosity: int = 0,
         dispersion_parameter: float = 1.5,
         validation_tuning_metric: str = "default",
         quantile: float = 0.5,
         calculate_custom_validation_error_function: Optional[
             Callable[
-                [npt.ArrayLike, npt.ArrayLike, npt.ArrayLike, npt.ArrayLike], float
+                [
+                    FloatVector,
+                    FloatVector,
+                    FloatVector,
+                    FloatVector,
+                    FloatMatrix,
+                ],
+                float,
             ]
         ] = None,
         calculate_custom_loss_function: Optional[
             Callable[
-                [npt.ArrayLike, npt.ArrayLike, npt.ArrayLike, npt.ArrayLike], float
+                [
+                    FloatVector,
+                    FloatVector,
+                    FloatVector,
+                    FloatVector,
+                    FloatMatrix,
+                ],
+                float,
             ]
         ] = None,
         calculate_custom_negative_gradient_function: Optional[
-            Callable[[npt.ArrayLike, npt.ArrayLike, npt.ArrayLike], npt.ArrayLike]
+            Callable[
+                [FloatVector, FloatVector, FloatVector, FloatMatrix],
+                FloatVector,
+            ]
         ] = None,
         calculate_custom_transform_linear_predictor_to_predictions_function: Optional[
-            Callable[[npt.ArrayLike], npt.ArrayLike]
+            Callable[[FloatVector], FloatVector]
         ] = None,
         calculate_custom_differentiate_predictions_wrt_linear_predictor_function: Optional[
-            Callable[[npt.ArrayLike], npt.ArrayLike]
+            Callable[[FloatVector], FloatVector]
         ] = None,
-        boosting_steps_before_pruning_is_done: int = 500,
+        boosting_steps_before_interactions_are_allowed: int = 0,
+        monotonic_constraints_ignore_interactions: bool = False,
+        group_mse_by_prediction_bins: int = 10,
+        group_mse_cycle_min_obs_in_bin: int = 30,
+        early_stopping_rounds: int = 200,
+        num_first_steps_with_linear_effects_only: int = 0,
+        penalty_for_non_linearity: float = 0.0,
+        penalty_for_interactions: float = 0.0,
+        max_terms: int = 0,
+        ridge_penalty: float = 0.0001,
+        mean_bias_correction: bool = False,
+        faster_convergence: bool = False,
+        preprocess: bool = True,
     ):
         self.m = m
         self.v = v
@@ -51,7 +136,7 @@ class APLRRegressor:
         self.loss_function = loss_function
         self.link_function = link_function
         self.n_jobs = n_jobs
-        self.validation_ratio = validation_ratio
+        self.cv_folds = cv_folds
         self.bins = bins
         self.max_interaction_level = max_interaction_level
         self.max_interactions = max_interactions
@@ -75,9 +160,25 @@ class APLRRegressor:
         self.calculate_custom_differentiate_predictions_wrt_linear_predictor_function = (
             calculate_custom_differentiate_predictions_wrt_linear_predictor_function
         )
-        self.boosting_steps_before_pruning_is_done = (
-            boosting_steps_before_pruning_is_done
+        self.boosting_steps_before_interactions_are_allowed = (
+            boosting_steps_before_interactions_are_allowed
+        )
+        self.monotonic_constraints_ignore_interactions = (
+            monotonic_constraints_ignore_interactions
+        )
+        self.group_mse_by_prediction_bins = group_mse_by_prediction_bins
+        self.group_mse_cycle_min_obs_in_bin = group_mse_cycle_min_obs_in_bin
+        self.early_stopping_rounds = early_stopping_rounds
+        self.num_first_steps_with_linear_effects_only = (
+            num_first_steps_with_linear_effects_only
         )
+        self.penalty_for_non_linearity = penalty_for_non_linearity
+        self.penalty_for_interactions = penalty_for_interactions
+        self.max_terms = max_terms
+        self.ridge_penalty = ridge_penalty
+        self.mean_bias_correction = mean_bias_correction
+        self.faster_convergence = faster_convergence
+        self.preprocess = preprocess
 
         # Creating aplr_cpp and setting parameters
         self.APLRRegressor = aplr_cpp.APLRRegressor()
@@ -91,7 +192,7 @@ class APLRRegressor:
         self.APLRRegressor.loss_function = self.loss_function
         self.APLRRegressor.link_function = self.link_function
         self.APLRRegressor.n_jobs = self.n_jobs
-        self.APLRRegressor.validation_ratio = self.validation_ratio
+        self.APLRRegressor.cv_folds = self.cv_folds
         self.APLRRegressor.bins = self.bins
         self.APLRRegressor.max_interaction_level = self.max_interaction_level
         self.APLRRegressor.max_interactions = self.max_interactions
@@ -119,38 +220,74 @@ class APLRRegressor:
         self.APLRRegressor.calculate_custom_differentiate_predictions_wrt_linear_predictor_function = (
             self.calculate_custom_differentiate_predictions_wrt_linear_predictor_function
         )
-        self.APLRRegressor.boosting_steps_before_pruning_is_done = (
-            self.boosting_steps_before_pruning_is_done
+        self.APLRRegressor.boosting_steps_before_interactions_are_allowed = (
+            self.boosting_steps_before_interactions_are_allowed
         )
+        self.APLRRegressor.monotonic_constraints_ignore_interactions = (
+            self.monotonic_constraints_ignore_interactions
+        )
+        self.APLRRegressor.group_mse_by_prediction_bins = (
+            self.group_mse_by_prediction_bins
+        )
+        self.APLRRegressor.group_mse_cycle_min_obs_in_bin = (
+            self.group_mse_cycle_min_obs_in_bin
+        )
+        self.APLRRegressor.early_stopping_rounds = self.early_stopping_rounds
+        self.APLRRegressor.num_first_steps_with_linear_effects_only = (
+            self.num_first_steps_with_linear_effects_only
+        )
+        self.APLRRegressor.penalty_for_non_linearity = self.penalty_for_non_linearity
+        self.APLRRegressor.penalty_for_interactions = self.penalty_for_interactions
+        self.APLRRegressor.max_terms = self.max_terms
+        self.APLRRegressor.ridge_penalty = self.ridge_penalty
+        self.APLRRegressor.mean_bias_correction = self.mean_bias_correction
+        self.APLRRegressor.faster_convergence = self.faster_convergence
+        self.APLRRegressor.preprocess = self.preprocess
 
     def fit(
         self,
-        X: npt.ArrayLike,
-        y: npt.ArrayLike,
-        sample_weight: npt.ArrayLike = np.empty(0),
+        X: Union[pd.DataFrame, FloatMatrix],
+        y: FloatVector,
+        sample_weight: FloatVector = np.empty(0),
         X_names: List[str] = [],
-        validation_set_indexes: List[int] = [],
+        cv_observations: IntMatrix = np.empty([0, 0]),
         prioritized_predictors_indexes: List[int] = [],
         monotonic_constraints: List[int] = [],
-        group: npt.ArrayLike = np.empty(0),
+        group: FloatVector = np.empty(0),
         interaction_constraints: List[List[int]] = [],
+        other_data: FloatMatrix = np.empty([0, 0]),
+        predictor_learning_rates: List[float] = [],
+        predictor_penalties_for_non_linearity: List[float] = [],
+        predictor_penalties_for_interactions: List[float] = [],
+        predictor_min_observations_in_split: List[int] = [],
     ):
         self.__set_params_cpp()
+        X = _prepare_input_data(X, self.preprocess)
+
         self.APLRRegressor.fit(
             X,
             y,
             sample_weight,
             X_names,
-            validation_set_indexes,
+            cv_observations,
             prioritized_predictors_indexes,
             monotonic_constraints,
             group,
             interaction_constraints,
+            other_data,
+            predictor_learning_rates,
+            predictor_penalties_for_non_linearity,
+            predictor_penalties_for_interactions,
+            predictor_min_observations_in_split,
         )
 
     def predict(
-        self, X: npt.ArrayLike, cap_predictions_to_minmax_in_training: bool = True
-    ) -> npt.ArrayLike:
+        self,
+        X: Union[pd.DataFrame, FloatMatrix],
+        cap_predictions_to_minmax_in_training: bool = True,
+    ) -> FloatVector:
+        X = _prepare_input_data(X, self.preprocess)
+
         if self.link_function == "custom_function":
             self.APLRRegressor.calculate_custom_transform_linear_predictor_to_predictions_function = (
                 self.calculate_custom_transform_linear_predictor_to_predictions_function
@@ -160,46 +297,271 @@ class APLRRegressor:
     def set_term_names(self, X_names: List[str]):
         self.APLRRegressor.set_term_names(X_names)
 
-    def calculate_local_feature_importance(self, X: npt.ArrayLike) -> npt.ArrayLike:
-        return self.APLRRegressor.calculate_local_feature_importance(X)
+    def calculate_feature_importance(
+        self,
+        X: Union[pd.DataFrame, FloatMatrix],
+        sample_weight: FloatVector = np.empty(0),
+    ) -> FloatVector:
+        X = _prepare_input_data(X, self.preprocess)
+        return self.APLRRegressor.calculate_feature_importance(X, sample_weight)
 
-    def calculate_local_feature_importance_for_terms(
-        self, X: npt.ArrayLike
-    ) -> npt.ArrayLike:
-        return self.APLRRegressor.calculate_local_feature_importance_for_terms(X)
+    def calculate_term_importance(
+        self,
+        X: Union[pd.DataFrame, FloatMatrix],
+        sample_weight: FloatVector = np.empty(0),
+    ) -> FloatVector:
+        X = _prepare_input_data(X, self.preprocess)
+        return self.APLRRegressor.calculate_term_importance(X, sample_weight)
+
+    def calculate_local_feature_contribution(
+        self, X: Union[pd.DataFrame, FloatMatrix]
+    ) -> FloatMatrix:
+        X = _prepare_input_data(X, self.preprocess)
+        return self.APLRRegressor.calculate_local_feature_contribution(X)
+
+    def calculate_local_term_contribution(
+        self, X: Union[pd.DataFrame, FloatMatrix]
+    ) -> FloatMatrix:
+        X = _prepare_input_data(X, self.preprocess)
+        return self.APLRRegressor.calculate_local_term_contribution(X)
+
+    def calculate_local_contribution_from_selected_terms(
+        self, X: Union[pd.DataFrame, FloatMatrix], predictor_indexes: List[int]
+    ) -> FloatVector:
+        X = _prepare_input_data(X, self.preprocess)
+        return self.APLRRegressor.calculate_local_contribution_from_selected_terms(
+            X, predictor_indexes
+        )
 
-    def calculate_terms(self, X: npt.ArrayLike) -> npt.ArrayLike:
+    def calculate_terms(self, X: Union[pd.DataFrame, FloatMatrix]) -> FloatMatrix:
+        X = _prepare_input_data(X, self.preprocess)
         return self.APLRRegressor.calculate_terms(X)
 
     def get_term_names(self) -> List[str]:
         return self.APLRRegressor.get_term_names()
 
-    def get_term_coefficients(self) -> npt.ArrayLike:
-        return self.APLRRegressor.get_term_coefficients()
+    def get_term_affiliations(self) -> List[str]:
+        return self.APLRRegressor.get_term_affiliations()
+
+    def get_unique_term_affiliations(self) -> List[str]:
+        return self.APLRRegressor.get_unique_term_affiliations()
 
-    def get_term_coefficient_steps(self, term_index: int) -> npt.ArrayLike:
-        return self.APLRRegressor.get_term_coefficient_steps(term_index)
+    def get_base_predictors_in_each_unique_term_affiliation(self) -> List[List[int]]:
+        return self.APLRRegressor.get_base_predictors_in_each_unique_term_affiliation()
 
-    def get_validation_error_steps(self) -> npt.ArrayLike:
+    def get_term_coefficients(self) -> FloatVector:
+        return self.APLRRegressor.get_term_coefficients()
+
+    def get_validation_error_steps(self) -> FloatMatrix:
         return self.APLRRegressor.get_validation_error_steps()
 
-    def get_feature_importance(self) -> npt.ArrayLike:
+    def get_feature_importance(self) -> FloatVector:
         return self.APLRRegressor.get_feature_importance()
 
+    def get_term_importance(self) -> FloatVector:
+        return self.APLRRegressor.get_term_importance()
+
+    def get_term_main_predictor_indexes(self) -> IntVector:
+        return self.APLRRegressor.get_term_main_predictor_indexes()
+
+    def get_term_interaction_levels(self) -> IntVector:
+        return self.APLRRegressor.get_term_interaction_levels()
+
     def get_intercept(self) -> float:
         return self.APLRRegressor.get_intercept()
 
-    def get_intercept_steps(self) -> npt.ArrayLike:
-        return self.APLRRegressor.get_intercept_steps()
-
     def get_optimal_m(self) -> int:
         return self.APLRRegressor.get_optimal_m()
 
     def get_validation_tuning_metric(self) -> str:
         return self.APLRRegressor.get_validation_tuning_metric()
 
-    def get_validation_indexes(self) -> List[int]:
-        return self.APLRRegressor.get_validation_indexes()
+    def get_main_effect_shape(self, predictor_index: int) -> Dict[float, float]:
+        return self.APLRRegressor.get_main_effect_shape(predictor_index)
+
+    def get_unique_term_affiliation_shape(
+        self,
+        unique_term_affiliation: str,
+        max_rows_before_sampling: int = 500000,
+        additional_points: int = 250,
+    ) -> FloatMatrix:
+        return self.APLRRegressor.get_unique_term_affiliation_shape(
+            unique_term_affiliation, max_rows_before_sampling, additional_points
+        )
+
+    def get_cv_error(self) -> float:
+        return self.APLRRegressor.get_cv_error()
+
+    def get_num_cv_folds(self) -> int:
+        """
+        Gets the number of cross-validation folds used during training.
+
+        :return: The number of folds.
+        """
+        return self.APLRRegressor.get_num_cv_folds()
+
+    def get_cv_validation_predictions(self, fold_index: int) -> FloatVector:
+        """
+        Gets the validation predictions for a specific cross-validation fold.
+
+        Note that these predictions may be conservative, as the final model is an ensemble of the models
+        from all cross-validation folds, which has a variance-reducing effect similar to bagging.
+
+        :param fold_index: The index of the fold.
+        :return: A numpy array containing the validation predictions.
+        """
+        return self.APLRRegressor.get_cv_validation_predictions(fold_index)
+
+    def get_cv_y(self, fold_index: int) -> FloatVector:
+        """
+        Gets the validation response values (y) for a specific cross-validation fold.
+
+        :param fold_index: The index of the fold.
+        :return: A numpy array containing the validation response values.
+        """
+        return self.APLRRegressor.get_cv_y(fold_index)
+
+    def get_cv_sample_weight(self, fold_index: int) -> FloatVector:
+        """
+        Gets the validation sample weights for a specific cross-validation fold.
+
+        :param fold_index: The index of the fold.
+        :return: A numpy array containing the validation sample weights.
+        """
+        return self.APLRRegressor.get_cv_sample_weight(fold_index)
+
+    def get_cv_validation_indexes(self, fold_index: int) -> IntVector:
+        """
+        Gets the original indexes of the validation observations for a specific cross-validation fold.
+
+        :param fold_index: The index of the fold.
+        :return: A numpy array containing the original indexes.
+        """
+        return self.APLRRegressor.get_cv_validation_indexes(fold_index)
+
+    def set_intercept(self, value: float):
+        self.APLRRegressor.set_intercept(value)
+
+    def plot_affiliation_shape(
+        self,
+        affiliation: str,
+        plot: bool = True,
+        save: bool = False,
+        path: str = "",
+    ):
+        """
+        Plots or saves the shape of a given unique term affiliation.
+
+        For main effects, it produces a line plot. For two-way interactions, it produces a heatmap.
+        Plotting for higher-order interactions is not supported.
+
+        :param affiliation: A string specifying which unique_term_affiliation to use.
+        :param plot: If True, displays the plot.
+        :param save: If True, saves the plot to a file.
+        :param path: The file path to save the plot. If empty and save is True, a default path will be used.
+        """
+        try:
+            import matplotlib.pyplot as plt
+        except ImportError:
+            raise ImportError("matplotlib is required for plotting. Please install it.")
+
+        all_affiliations = self.get_unique_term_affiliations()
+        if affiliation not in all_affiliations:
+            raise ValueError(
+                f"Affiliation '{affiliation}' not found in model. "
+                f"Available affiliations are: {all_affiliations}"
+            )
+
+        affiliation_index = all_affiliations.index(affiliation)
+
+        predictors_in_each_affiliation = (
+            self.get_base_predictors_in_each_unique_term_affiliation()
+        )
+        predictor_indexes_used = predictors_in_each_affiliation[affiliation_index]
+
+        shape = self.get_unique_term_affiliation_shape(affiliation)
+        if shape.shape[0] == 0:
+            print(f"No shape data available for affiliation '{affiliation}'.")
+            return
+
+        predictor_names = affiliation.split(" & ")
+
+        is_main_effect: bool = len(predictor_indexes_used) == 1
+        is_two_way_interaction: bool = len(predictor_indexes_used) == 2
+
+        if is_main_effect:
+            fig = plt.figure()
+            # Sort by predictor value for a clean line plot
+            sorted_indices = np.argsort(shape[:, 0])
+            plt.plot(shape[sorted_indices, 0], shape[sorted_indices, 1])
+            plt.xlabel(predictor_names[0])
+            plt.ylabel("Contribution to linear predictor")
+            plt.title(f"Main effect of {predictor_names[0]}")
+            plt.grid(True)
+        elif is_two_way_interaction:
+            fig = plt.figure(figsize=(8, 6))
+
+            # Get unique coordinates and their inverse mapping
+            y_unique, y_inv = np.unique(shape[:, 0], return_inverse=True)
+            x_unique, x_inv = np.unique(shape[:, 1], return_inverse=True)
+
+            # Create grid for sums and counts
+            grid_sums = np.zeros((len(y_unique), len(x_unique)))
+            grid_counts = np.zeros((len(y_unique), len(x_unique)))
+
+            # Populate sums and counts to later calculate the mean
+            np.add.at(grid_sums, (y_inv, x_inv), shape[:, 2])
+            np.add.at(grid_counts, (y_inv, x_inv), 1)
+
+            # Calculate mean, avoiding division by zero
+            with np.errstate(divide="ignore", invalid="ignore"):
+                pivot_table_values = np.true_divide(grid_sums, grid_counts)
+                # Where there's no data, pivot_table_values will be nan, which is fine for imshow.
+
+            plt.imshow(
+                pivot_table_values,
+                aspect="auto",
+                origin="lower",
+                extent=[
+                    x_unique.min(),
+                    x_unique.max(),
+                    y_unique.min(),
+                    y_unique.max(),
+                ],
+                cmap="Blues_r",
+            )
+            plt.colorbar(label="Contribution to the linear predictor")
+            plt.xlabel(predictor_names[1])
+            plt.ylabel(predictor_names[0])
+            plt.title(
+                f"Interaction between {predictor_names[0]} and {predictor_names[1]}"
+            )
+        else:
+            print(
+                f"Plotting for interaction level > 2 is not supported. Affiliation: {affiliation}"
+            )
+            return
+
+        if save:
+            save_path = path or f"shape_of_{affiliation.replace(' & ', '_')}.png"
+            plt.savefig(save_path)
+
+        if plot:
+            plt.show()
+
+        plt.close(fig)
+
+    def remove_provided_custom_functions(self):
+        self.APLRRegressor.remove_provided_custom_functions()
+        self.calculate_custom_validation_error_function = None
+        self.calculate_custom_loss_function = None
+        self.calculate_custom_negative_gradient_function = None
+
+    def clear_cv_results(self):
+        """
+        Clears the stored cross-validation results (predictions, y, etc.) to free up memory.
+        """
+        self.APLRRegressor.clear_cv_results()
 
     # For sklearn
     def get_params(self, deep=True):
@@ -210,7 +572,7 @@ class APLRRegressor:
             "loss_function": self.loss_function,
             "link_function": self.link_function,
             "n_jobs": self.n_jobs,
-            "validation_ratio": self.validation_ratio,
+            "cv_folds": self.cv_folds,
             "bins": self.bins,
             "max_interaction_level": self.max_interaction_level,
             "max_interactions": self.max_interactions,
@@ -226,7 +588,19 @@ class APLRRegressor:
             "calculate_custom_negative_gradient_function": self.calculate_custom_negative_gradient_function,
             "calculate_custom_transform_linear_predictor_to_predictions_function": self.calculate_custom_transform_linear_predictor_to_predictions_function,
             "calculate_custom_differentiate_predictions_wrt_linear_predictor_function": self.calculate_custom_differentiate_predictions_wrt_linear_predictor_function,
-            "boosting_steps_before_pruning_is_done": self.boosting_steps_before_pruning_is_done,
+            "boosting_steps_before_interactions_are_allowed": self.boosting_steps_before_interactions_are_allowed,
+            "monotonic_constraints_ignore_interactions": self.monotonic_constraints_ignore_interactions,
+            "group_mse_by_prediction_bins": self.group_mse_by_prediction_bins,
+            "group_mse_cycle_min_obs_in_bin": self.group_mse_cycle_min_obs_in_bin,
+            "early_stopping_rounds": self.early_stopping_rounds,
+            "num_first_steps_with_linear_effects_only": self.num_first_steps_with_linear_effects_only,
+            "penalty_for_non_linearity": self.penalty_for_non_linearity,
+            "penalty_for_interactions": self.penalty_for_interactions,
+            "max_terms": self.max_terms,
+            "ridge_penalty": self.ridge_penalty,
+            "mean_bias_correction": self.mean_bias_correction,
+            "faster_convergence": self.faster_convergence,
+            "preprocess": self.preprocess,
         }
 
     # For sklearn
@@ -236,29 +610,44 @@ class APLRRegressor:
         self.__set_params_cpp()
         return self
 
+    def __setstate__(self, state):
+        # For backwards compatibility with older pickled models
+        if "preprocess" not in state:
+            state["preprocess"] = False
+        self.__dict__.update(state)
+        self.__set_params_cpp()
+
 
 class APLRClassifier:
     def __init__(
         self,
-        m: int = 9000,
-        v: float = 0.1,
+        m: int = 3000,
+        v: float = 0.5,
         random_state: int = 0,
         n_jobs: int = 0,
-        validation_ratio: float = 0.2,
+        cv_folds: int = 5,
         bins: int = 300,
         verbosity: int = 0,
         max_interaction_level: int = 1,
         max_interactions: int = 100000,
-        min_observations_in_split: int = 20,
-        ineligible_boosting_steps_added: int = 10,
-        max_eligible_terms: int = 5,
-        boosting_steps_before_pruning_is_done: int = 500,
+        min_observations_in_split: int = 4,
+        ineligible_boosting_steps_added: int = 15,
+        max_eligible_terms: int = 7,
+        boosting_steps_before_interactions_are_allowed: int = 0,
+        monotonic_constraints_ignore_interactions: bool = False,
+        early_stopping_rounds: int = 200,
+        num_first_steps_with_linear_effects_only: int = 0,
+        penalty_for_non_linearity: float = 0.0,
+        penalty_for_interactions: float = 0.0,
+        max_terms: int = 0,
+        ridge_penalty: float = 0.0001,
+        preprocess: bool = True,
     ):
         self.m = m
         self.v = v
         self.random_state = random_state
         self.n_jobs = n_jobs
-        self.validation_ratio = validation_ratio
+        self.cv_folds = cv_folds
         self.bins = bins
         self.verbosity = verbosity
         self.max_interaction_level = max_interaction_level
@@ -266,9 +655,21 @@ class APLRClassifier:
         self.min_observations_in_split = min_observations_in_split
         self.ineligible_boosting_steps_added = ineligible_boosting_steps_added
         self.max_eligible_terms = max_eligible_terms
-        self.boosting_steps_before_pruning_is_done = (
-            boosting_steps_before_pruning_is_done
+        self.boosting_steps_before_interactions_are_allowed = (
+            boosting_steps_before_interactions_are_allowed
+        )
+        self.monotonic_constraints_ignore_interactions = (
+            monotonic_constraints_ignore_interactions
+        )
+        self.early_stopping_rounds = early_stopping_rounds
+        self.num_first_steps_with_linear_effects_only = (
+            num_first_steps_with_linear_effects_only
         )
+        self.penalty_for_non_linearity = penalty_for_non_linearity
+        self.penalty_for_interactions = penalty_for_interactions
+        self.max_terms = max_terms
+        self.ridge_penalty = ridge_penalty
+        self.preprocess = preprocess
 
         # Creating aplr_cpp and setting parameters
         self.APLRClassifier = aplr_cpp.APLRClassifier()
@@ -280,7 +681,7 @@ class APLRClassifier:
         self.APLRClassifier.v = self.v
         self.APLRClassifier.random_state = self.random_state
         self.APLRClassifier.n_jobs = self.n_jobs
-        self.APLRClassifier.validation_ratio = self.validation_ratio
+        self.APLRClassifier.cv_folds = self.cv_folds
         self.APLRClassifier.bins = self.bins
         self.APLRClassifier.verbosity = self.verbosity
         self.APLRClassifier.max_interaction_level = self.max_interaction_level
@@ -290,66 +691,146 @@ class APLRClassifier:
             self.ineligible_boosting_steps_added
         )
         self.APLRClassifier.max_eligible_terms = self.max_eligible_terms
-        self.APLRClassifier.boosting_steps_before_pruning_is_done = (
-            self.boosting_steps_before_pruning_is_done
+        self.APLRClassifier.boosting_steps_before_interactions_are_allowed = (
+            self.boosting_steps_before_interactions_are_allowed
         )
+        self.APLRClassifier.monotonic_constraints_ignore_interactions = (
+            self.monotonic_constraints_ignore_interactions
+        )
+        self.APLRClassifier.early_stopping_rounds = self.early_stopping_rounds
+        self.APLRClassifier.num_first_steps_with_linear_effects_only = (
+            self.num_first_steps_with_linear_effects_only
+        )
+        self.APLRClassifier.penalty_for_non_linearity = self.penalty_for_non_linearity
+        self.APLRClassifier.penalty_for_interactions = self.penalty_for_interactions
+        self.APLRClassifier.max_terms = self.max_terms
+        self.APLRClassifier.ridge_penalty = self.ridge_penalty
+        self.APLRClassifier.preprocess = self.preprocess
 
     def fit(
         self,
-        X: npt.ArrayLike,
-        y: List[str],
-        sample_weight: npt.ArrayLike = np.empty(0),
+        X: Union[pd.DataFrame, FloatMatrix],
+        y: Union[FloatVector, List[str]],
+        sample_weight: FloatVector = np.empty(0),
         X_names: List[str] = [],
-        validation_set_indexes: List[int] = [],
+        cv_observations: IntMatrix = np.empty([0, 0]),
         prioritized_predictors_indexes: List[int] = [],
         monotonic_constraints: List[int] = [],
         interaction_constraints: List[List[int]] = [],
+        predictor_learning_rates: List[float] = [],
+        predictor_penalties_for_non_linearity: List[float] = [],
+        predictor_penalties_for_interactions: List[float] = [],
+        predictor_min_observations_in_split: List[int] = [],
     ):
         self.__set_params_cpp()
+
+        X = _prepare_input_data(X, self.preprocess)
+
+        if isinstance(y, np.ndarray):
+            y = y.astype(str).tolist()
+        elif isinstance(y, list) and y and not isinstance(y[0], str):
+            y = [str(val) for val in y]
+
         self.APLRClassifier.fit(
             X,
             y,
             sample_weight,
             X_names,
-            validation_set_indexes,
+            cv_observations,
             prioritized_predictors_indexes,
             monotonic_constraints,
             interaction_constraints,
+            predictor_learning_rates,
+            predictor_penalties_for_non_linearity,
+            predictor_penalties_for_interactions,
+            predictor_min_observations_in_split,
         )
+        # For sklearn
+        self.classes_ = np.arange(len(self.APLRClassifier.get_categories()))
 
     def predict_class_probabilities(
-        self, X: npt.ArrayLike, cap_predictions_to_minmax_in_training: bool = False
-    ) -> npt.ArrayLike:
+        self,
+        X: Union[pd.DataFrame, FloatMatrix],
+        cap_predictions_to_minmax_in_training: bool = False,
+    ) -> FloatMatrix:
+        X = _prepare_input_data(X, self.preprocess)
+
         return self.APLRClassifier.predict_class_probabilities(
             X, cap_predictions_to_minmax_in_training
         )
 
     def predict(
-        self, X: npt.ArrayLike, cap_predictions_to_minmax_in_training: bool = False
+        self,
+        X: Union[pd.DataFrame, FloatMatrix],
+        cap_predictions_to_minmax_in_training: bool = False,
     ) -> List[str]:
+        X = _prepare_input_data(X, self.preprocess)
+
         return self.APLRClassifier.predict(X, cap_predictions_to_minmax_in_training)
 
-    def calculate_local_feature_importance(self, X: npt.ArrayLike) -> npt.ArrayLike:
-        return self.APLRClassifier.calculate_local_feature_importance(X)
+    def calculate_local_feature_contribution(
+        self, X: Union[pd.DataFrame, FloatMatrix]
+    ) -> FloatMatrix:
+        X = _prepare_input_data(X, self.preprocess)
+        return self.APLRClassifier.calculate_local_feature_contribution(X)
 
     def get_categories(self) -> List[str]:
         return self.APLRClassifier.get_categories()
 
     def get_logit_model(self, category: str) -> APLRRegressor:
-        return self.APLRClassifier.get_logit_model(category)
+        logit_model_cpp = self.APLRClassifier.get_logit_model(category)
+
+        logit_model_py = APLRRegressor(
+            m=self.m,
+            v=self.v,
+            random_state=self.random_state,
+            loss_function="binomial",
+            link_function="logit",
+            n_jobs=self.n_jobs,
+            cv_folds=self.cv_folds,
+            bins=self.bins,
+            max_interaction_level=self.max_interaction_level,
+            max_interactions=self.max_interactions,
+            min_observations_in_split=self.min_observations_in_split,
+            ineligible_boosting_steps_added=self.ineligible_boosting_steps_added,
+            max_eligible_terms=self.max_eligible_terms,
+            verbosity=self.verbosity,
+            boosting_steps_before_interactions_are_allowed=self.boosting_steps_before_interactions_are_allowed,
+            monotonic_constraints_ignore_interactions=self.monotonic_constraints_ignore_interactions,
+            early_stopping_rounds=self.early_stopping_rounds,
+            num_first_steps_with_linear_effects_only=self.num_first_steps_with_linear_effects_only,
+            penalty_for_non_linearity=self.penalty_for_non_linearity,
+            penalty_for_interactions=self.penalty_for_interactions,
+            max_terms=self.max_terms,
+            ridge_penalty=self.ridge_penalty,
+            preprocess=self.preprocess,
+        )
+
+        logit_model_py.APLRRegressor = logit_model_cpp
 
-    def get_validation_indexes(self) -> List[int]:
-        return self.APLRClassifier.get_validation_indexes()
+        return logit_model_py
 
-    def get_validation_error_steps(self) -> npt.ArrayLike:
+    def get_validation_error_steps(self) -> FloatMatrix:
         return self.APLRClassifier.get_validation_error_steps()
 
-    def get_validation_error(self) -> float:
-        return self.APLRClassifier.get_validation_error()
+    def get_cv_error(self) -> float:
+        return self.APLRClassifier.get_cv_error()
 
-    def get_feature_importance(self) -> npt.ArrayLike:
+    def get_feature_importance(self) -> FloatVector:
         return self.APLRClassifier.get_feature_importance()
 
+    def get_unique_term_affiliations(self) -> List[str]:
+        return self.APLRClassifier.get_unique_term_affiliations()
+
+    def get_base_predictors_in_each_unique_term_affiliation(self) -> List[List[int]]:
+        return self.APLRClassifier.get_base_predictors_in_each_unique_term_affiliation()
+
+    def clear_cv_results(self):
+        """
+        Clears the stored cross-validation results from all underlying logit models to free up memory.
+        """
+        self.APLRClassifier.clear_cv_results()
+
     # For sklearn
     def get_params(self, deep=True):
         return {
@@ -357,7 +838,7 @@ class APLRClassifier:
             "v": self.v,
             "random_state": self.random_state,
             "n_jobs": self.n_jobs,
-            "validation_ratio": self.validation_ratio,
+            "cv_folds": self.cv_folds,
             "bins": self.bins,
             "verbosity": self.verbosity,
             "max_interaction_level": self.max_interaction_level,
@@ -365,7 +846,15 @@ class APLRClassifier:
             "min_observations_in_split": self.min_observations_in_split,
             "ineligible_boosting_steps_added": self.ineligible_boosting_steps_added,
             "max_eligible_terms": self.max_eligible_terms,
-            "boosting_steps_before_pruning_is_done": self.boosting_steps_before_pruning_is_done,
+            "boosting_steps_before_interactions_are_allowed": self.boosting_steps_before_interactions_are_allowed,
+            "monotonic_constraints_ignore_interactions": self.monotonic_constraints_ignore_interactions,
+            "early_stopping_rounds": self.early_stopping_rounds,
+            "num_first_steps_with_linear_effects_only": self.num_first_steps_with_linear_effects_only,
+            "penalty_for_non_linearity": self.penalty_for_non_linearity,
+            "penalty_for_interactions": self.penalty_for_interactions,
+            "max_terms": self.max_terms,
+            "ridge_penalty": self.ridge_penalty,
+            "preprocess": self.preprocess,
         }
 
     # For sklearn
@@ -374,3 +863,79 @@ class APLRClassifier:
             setattr(self, parameter, value)
         self.__set_params_cpp()
         return self
+
+    # For sklearn
+    def predict_proba(self, X: FloatMatrix) -> FloatMatrix:
+        return self.predict_class_probabilities(X)
+
+    def __setstate__(self, state):
+        # For backwards compatibility with older pickled models
+        if "preprocess" not in state:
+            state["preprocess"] = False
+        self.__dict__.update(state)
+        self.__set_params_cpp()
+
+
+class APLRTuner:
+    def __init__(
+        self,
+        parameters: Union[Dict[str, List[float]], List[Dict[str, List[float]]]] = {
+            "max_interaction_level": [0, 1],
+            "min_observations_in_split": [4, 10, 20, 100, 500, 1000],
+        },
+        is_regressor: bool = True,
+    ):
+        self.parameters = parameters
+        self.is_regressor = is_regressor
+        self.parameter_grid = self._create_parameter_grid()
+
+    def _create_parameter_grid(self) -> List[Dict[str, float]]:
+        items = sorted(self.parameters.items())
+        keys, values = zip(*items)
+        combinations = list(itertools.product(*values))
+        grid = [dict(zip(keys, combination)) for combination in combinations]
+        return grid
+
+    def fit(self, X: Union[pd.DataFrame, FloatMatrix], y: FloatVector, **kwargs):
+        self.cv_results: List[Dict[str, float]] = []
+        best_validation_result = np.inf
+        for params in self.parameter_grid:
+            if self.is_regressor:
+                model = APLRRegressor(**params)
+            else:
+                model = APLRClassifier(**params)
+            model.fit(X, y, **kwargs)
+            cv_error_for_this_model = model.get_cv_error()
+            cv_results_for_this_model = model.get_params()
+            cv_results_for_this_model["cv_error"] = cv_error_for_this_model
+            self.cv_results.append(cv_results_for_this_model)
+            if cv_error_for_this_model < best_validation_result:
+                best_validation_result = cv_error_for_this_model
+                self.best_model = model
+        self.cv_results = sorted(self.cv_results, key=lambda x: x["cv_error"])
+
+    def predict(
+        self, X: Union[pd.DataFrame, FloatMatrix], **kwargs
+    ) -> Union[FloatVector, List[str]]:
+        return self.best_model.predict(X, **kwargs)
+
+    def predict_class_probabilities(
+        self, X: Union[pd.DataFrame, FloatMatrix], **kwargs
+    ) -> FloatMatrix:
+        if self.is_regressor == False:
+            return self.best_model.predict_class_probabilities(X, **kwargs)
+        else:
+            raise TypeError(
+                "predict_class_probabilities is only possible when is_regressor is False"
+            )
+
+    def predict_proba(
+        self, X: Union[pd.DataFrame, FloatMatrix], **kwargs
+    ) -> FloatMatrix:
+        return self.predict_class_probabilities(X, **kwargs)
+
+    def get_best_estimator(self) -> Union[APLRClassifier, APLRRegressor]:
+        return self.best_model
+
+    def get_cv_results(self) -> List[Dict[str, float]]:
+        return self.cv_results

aplr-10.20.0.dist-info/METADATA

@@ -0,0 +1,34 @@
+Metadata-Version: 2.4
+Name: aplr
+Version: 10.20.0
+Summary: Automatic Piecewise Linear Regression
+Home-page: https://github.com/ottenbreit-data-science/aplr
+Author: Mathias von Ottenbreit
+Author-email: ottenbreitdatascience@gmail.com
+License: MIT
+Platform: Windows
+Platform: Linux
+Platform: MacOS
+Classifier: License :: OSI Approved :: MIT License
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy>=1.11
+Requires-Dist: pandas>=1.0.0
+Provides-Extra: plots
+Requires-Dist: matplotlib>=3.0; extra == "plots"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: license
+Dynamic: license-file
+Dynamic: platform
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+The documentation for Automatic Piecewise Linear Regression is available at [https://github.com/ottenbreit-data-science/aplr](https://github.com/ottenbreit-data-science/aplr).

aplr-10.20.0.dist-info/RECORD

@@ -0,0 +1,8 @@
+aplr_cpp.cp311-win_amd64.pyd,sha256=ecU1pYcAJ3K_B_pcYJS06DiNSMpCCovciVQWGpbWiwA,825856
+aplr/__init__.py,sha256=oDFSgVytP_qQ8ilun6oHxKr-DYEeqjEQp5FciX45lls,21
+aplr/aplr.py,sha256=I_LyS_uH9wmW7wE8uP6nvPhfCeeA1cQXIuLPqwT21OE,40125
+aplr-10.20.0.dist-info/licenses/LICENSE,sha256=YOMo-RaL4P7edMZGD96-NskKpxyMZdP3-WiiMMmihNk,1134
+aplr-10.20.0.dist-info/METADATA,sha256=sRYG3XKunuA5qB-cm-6XBqUd4XRIgK1tVriXV3PX_5Y,1048
+aplr-10.20.0.dist-info/WHEEL,sha256=JLOMsP7F5qtkAkINx5UnzbFguf8CqZeraV8o04b0I8I,101
+aplr-10.20.0.dist-info/top_level.txt,sha256=DXVC0RIFGpzVnPeKWAZTXQdJheOEZL51Wip6Fx7zbR4,14
+aplr-10.20.0.dist-info/RECORD,,

{aplr-7.0.0.dist-info → aplr-10.20.0.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.40.0)
+Generator: setuptools (80.9.0)
 Root-Is-Purelib: false
 Tag: cp311-cp311-win_amd64
 

aplr-7.0.0.dist-info/METADATA

@@ -1,17 +0,0 @@
-Metadata-Version: 2.1
-Name: aplr
-Version: 7.0.0
-Summary: Automatic Piecewise Linear Regression
-Home-page: https://github.com/ottenbreit-data-science/aplr
-Author: Mathias von Ottenbreit
-Author-email: ottenbreitdatascience@gmail.com
-License: MIT
-Platform: Windows
-Platform: Linux
-Classifier: License :: OSI Approved :: MIT License
-Requires-Python: >=3.8
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: numpy (>=1.20)
-
-Build predictive and interpretable parametric regression or classification machine learning models in Python based on the Automatic Piecewise Linear Regression methodology developed by Mathias von Ottenbreit.

aplr-7.0.0.dist-info/RECORD

@@ -1,8 +0,0 @@
-aplr_cpp.cp311-win_amd64.pyd,sha256=pOvJghR3IOWSQcVzvvtAgQnlKIG9HMCerOMrwhuuACE,471040
-aplr/__init__.py,sha256=oDFSgVytP_qQ8ilun6oHxKr-DYEeqjEQp5FciX45lls,21
-aplr/aplr.py,sha256=0gVeqsL1WyXHZpIDqn6rXoT0tuv76nHZVs6LCzOxNEM,16012
-aplr-7.0.0.dist-info/LICENSE,sha256=YOMo-RaL4P7edMZGD96-NskKpxyMZdP3-WiiMMmihNk,1134
-aplr-7.0.0.dist-info/METADATA,sha256=KwO6mV-2SWzBGp3lXXT-UQBhgpv99T1NVwI6aBeoSoc,671
-aplr-7.0.0.dist-info/WHEEL,sha256=9wvhO-5NhjjD8YmmxAvXTPQXMDOZ50W5vklzeoqFtkM,102
-aplr-7.0.0.dist-info/top_level.txt,sha256=DXVC0RIFGpzVnPeKWAZTXQdJheOEZL51Wip6Fx7zbR4,14
-aplr-7.0.0.dist-info/RECORD,,