Feature Engineering & Selection Basics Questions

import pandas as pd
from typing import List, Dict

def impute_missing(df: pd.DataFrame,
                   numeric_cols: List[str],
                   categorical_cols: List[str],
                   inplace: bool = False) -> pd.DataFrame:
    """
    Impute numeric_cols with median and categorical_cols with mode.
    Returns the imputed DataFrame (a copy unless inplace=True).
    """
    if not inplace:
        df = df.copy()

    # Compute medians for numeric columns
    medians: Dict[str, float] = df[numeric_cols].median(skipna=True).to_dict()
    # Compute modes for categorical columns (take first mode; if no mode, leave NaN)
    modes: Dict[str, object] = {}
    for col in categorical_cols:
        if col in df.columns:
            mode_series = df[col].mode(dropna=True)
            modes[col] = mode_series.iloc[0] if not mode_series.empty else pd.NA

    # Fill numeric and categorical
    df = df.fillna({**medians, **modes})
    return df

# Example usage:
# Assume df has columns: ['user_id','age','country','last_login_days_ago','purchase_amount']
numeric_cols = ['age', 'last_login_days_ago', 'purchase_amount']
categorical_cols = ['country']

imputed_df = impute_missing(df, numeric_cols, categorical_cols)