data_preparation.py

import json
from sklearn.model_selection import train_test_split


def split_data(file_name: str, data_type: str):
    if data_type == "json":
        with open(file_name, 'r') as json_file:
            data = json.load(json_file)["data"]
            json_file.close()

    train, test = train_test_split(data, train_size=0.9, random_state=42)
    return(train, test)


def save_json(data: dict, file_name: str):
    """
    Method to save the json file.
    Parameters:
    ----------
    data: dict,
        data to be saved in file.
    file_name: str,
        name of the file.
    Returns:
    --------
    None
    """

    # save the split
    with open(file_name, "w") as data_file:
        json.dump(data, data_file, indent=2)
        data_file.close()


if __name__ == "__main__":
    # split the train data
    train, test = split_data("policies-qa-training.json", "json")

    # save the train split
    save_json({"data": train}, "train.json")

    # save the test split
    save_json({"data": test}, "test.json")