scripts/data_cleansing.py

import pandas as pd
import numpy as np


def cleanse_data(df, remove_duplicates, missing_strategy):
    """
    Perform data cleansing on the dataframe.
    
    Args:
        df: pandas DataFrame
        remove_duplicates: bool, whether to remove duplicate rows
        missing_strategy: str, 'drop', 'impute_mean', 'impute_median', 'impute_mode'
    
    Returns:
        df_clean: cleaned DataFrame
        original_shape: tuple (rows, cols) before cleansing
        cleaned_shape: tuple (rows, cols) after cleansing
    """
    df = df.copy()
    original_shape = df.shape
    
    # Remove duplicates
    if remove_duplicates:
        df = df.drop_duplicates()
    
    # Handle missing values
    if missing_strategy == 'drop':
        df = df.dropna()
    elif missing_strategy in ['impute_mean', 'impute_median']:
        for col in df.select_dtypes(include=[np.number]).columns:
            if missing_strategy == 'impute_mean':
                df[col] = df[col].fillna(df[col].mean())
            elif missing_strategy == 'impute_median':
                df[col] = df[col].fillna(df[col].median())
    elif missing_strategy == 'impute_mode':
        for col in df.columns:
            mode_val = df[col].mode()
            if not mode_val.empty:
                df[col] = df[col].fillna(mode_val[0])
    
    cleaned_shape = df.shape
    return df, original_shape, cleaned_shape