prepare_data.py

import pandas as pd
import os
from typing import Union

import datasets
from datasets import load_dataset


def save_and_compress(dataset: Union[datasets.Dataset, pd.DataFrame], name: str, idx=None):
    if idx:
        path = f"{name}_{idx}.jsonl"
    else:
        path = f"{name}.jsonl"

    print("Saving to", path)
    dataset.to_json(path, force_ascii=False, orient='records', lines=True)

    print("Compressing...")
    os.system(f'xz -zkf -T0 --memlimit-compress=60% {path}')  # -TO to use multithreading


for split in ["train", "valid", "test"]:
    dataset = load_dataset("parquet", data_files=f"original_data/{split}_en.parquet", split="train")
    dataset = dataset.remove_columns(['case_marked_as_closed'])  # this column brings problems
    # these are also potentially an issue: overflows
    dataset = dataset.remove_columns(['filing_date', 'date_first_instance_ruling', 'date_appeal_panel_session'])
    print(dataset[0])
    save_and_compress(dataset, f"data/{split}")