import logging

import pandas as pd
from zenml import step
from typing_extensions import Annotated
from typing import Tuple

from src.data_preprocessing import DataPreprocessing, DataCombiningBaseline

@step
def preprocess_data(
    df_movie: pd.DataFrame,
    df_user: pd.DataFrame,
    df_rating: pd.DataFrame,
) -> Annotated[pd.DataFrame, "df_total"]:  
    """
    Preprocess

    Args:
        df: Raw data
    Returns:
        X_train: Training data
        X_test: Testing data
        y_train: Training labels
        y_test: Testing labels
    """
    try:
        process_strategy = DataCombiningBaseline()
        data_preprocess_class = DataPreprocessing(df_movie, df_user, df_rating, process_strategy)
        processed_data = data_preprocess_class.handle_data()
        
        logging.info(f"df_test: \n{processed_data}")

        return processed_data
    except Exception as e:
        logging.error(f"Error in handling data: {e}")
        raise e