Source code for evalml.pipelines.components.transformers.imputers.simple_imputer

"""Component that imputes missing data according to a specified imputation strategy."""
import pandas as pd
import woodwork
from sklearn.impute import SimpleImputer as SkImputer
from woodwork.logical_types import Double

from evalml.pipelines.components.transformers import Transformer
from evalml.pipelines.components.utils import (
    drop_natural_language_columns,
    set_boolean_columns_to_categorical,
)
from evalml.utils import infer_feature_types
from evalml.utils.gen_utils import is_categorical_actually_boolean


class SimpleImputer(Transformer):
    """Imputes missing data according to a specified imputation strategy.  Natural language columns are ignored.

    Args:
        impute_strategy (string): Impute strategy to use. Valid values include "mean", "median", "most_frequent", "constant" for
           numerical data, and "most_frequent", "constant" for object data types.
        fill_value (string): When impute_strategy == "constant", fill_value is used to replace missing data.
           Defaults to 0 when imputing numerical data and "missing_value" for strings or object data types.
        random_seed (int): Seed for the random number generator. Defaults to 0.

    """

    name = "Simple Imputer"
    hyperparameter_ranges = {"impute_strategy": ["mean", "median", "most_frequent"]}
    """{
        "impute_strategy": ["mean", "median", "most_frequent"]
    }"""

    def __init__(
        self, impute_strategy="most_frequent", fill_value=None, random_seed=0, **kwargs
    ):
        parameters = {"impute_strategy": impute_strategy, "fill_value": fill_value}
        parameters.update(kwargs)
        self.impute_strategy = impute_strategy
        imputer = SkImputer(
            strategy=impute_strategy,
            fill_value=fill_value,
            missing_values=pd.NA,
            **kwargs,
        )
        self._all_null_cols = None
        super().__init__(
            parameters=parameters,
            component_obj=imputer,
            random_seed=random_seed,
        )

    def fit(self, X, y=None):
        """Fits imputer to data. 'None' values are converted to np.nan before imputation and are treated as the same.

        Args:
            X (pd.DataFrame or np.ndarray): the input training data of shape [n_samples, n_features]
            y (pd.Series, optional): the target training data of length [n_samples]

        Returns:
            self

        Raises:
            ValueError: if the SimpleImputer receives a dataframe with both Boolean and Categorical data.

        """
        X = infer_feature_types(X)

        if set([lt.type_string for lt in X.ww.logical_types.values()]) == {
            "boolean",
            "categorical",
        } and not all(
            [
                is_categorical_actually_boolean(X, col)
                for col in X.ww.select("Categorical")
            ],
        ):
            raise ValueError(
                "SimpleImputer cannot handle dataframes with both boolean and categorical features.  Use Imputer instead.",
            )
        nan_ratio = X.isna().sum() / X.shape[0]
        self._all_null_cols = nan_ratio[nan_ratio == 1].index.tolist()

        X, _ = drop_natural_language_columns(X)
        X = set_boolean_columns_to_categorical(X)

        # If the Dataframe only had natural language columns, do nothing.
        if X.shape[1] == 0:
            return self
        self._component_obj.fit(X, y)
        return self

    def transform(self, X, y=None):
        """Transforms input by imputing missing values. 'None' and np.nan values are treated as the same.

        Args:
            X (pd.DataFrame): Data to transform.
            y (pd.Series, optional): Ignored.

        Returns:
            pd.DataFrame: Transformed X
        """
        X = infer_feature_types(X)
        original_schema = X.ww.schema
        X = set_boolean_columns_to_categorical(X)

        # Return early since bool dtype doesn't support nans and sklearn errors if all cols are bool
        if (X.dtypes == bool).all():
            return X

        not_all_null_cols = [col for col in X.columns if col not in self._all_null_cols]
        original_index = X.index

        # Drop natural language columns and transform the other columns
        X_t, natural_language_cols = drop_natural_language_columns(X)
        if X_t.shape[1] == 0:
            return X
        not_all_null_or_natural_language_cols = [
            col for col in not_all_null_cols if col not in natural_language_cols
        ]

        X_t = self._component_obj.transform(X_t)
        X_t = pd.DataFrame(X_t, columns=not_all_null_or_natural_language_cols)

        new_schema = original_schema.get_subset_schema(X_t.columns)

        # TODO: Fix this after WW adds inference of object type booleans to BooleanNullable
        # Iterate through categorical columns that might have been boolean and convert them back to boolean
        for col in X.ww.select(["Categorical"], return_schema=True).columns:
            if is_categorical_actually_boolean(X, col):
                X_t[col] = X_t[col].astype(bool)

        # Convert Nullable Integers to Doubles for the "mean" and "median" strategies
        if self.impute_strategy in ["mean", "median"]:
            nullable_int_cols = X.ww.select(["IntegerNullable"], return_schema=True)
            nullable_int_cols = [x for x in nullable_int_cols.columns.keys()]
            for col in nullable_int_cols:
                new_schema.set_types({col: Double})
        X_t.ww.init(schema=new_schema)

        # Add back in natural language columns, unchanged
        if len(natural_language_cols) > 0:
            X_t = woodwork.concat_columns([X_t, X[natural_language_cols]])

        if not_all_null_or_natural_language_cols:
            X_t.index = original_index
        return X_t

    def fit_transform(self, X, y=None):
        """Fits on X and transforms X.

        Args:
            X (pd.DataFrame): Data to fit and transform
            y (pd.Series, optional): Target data.

        Returns:
            pd.DataFrame: Transformed X
        """
        return self.fit(X, y).transform(X, y)