#!/usr/bin/env python3 import io import re from pathlib import Path import requests import pyarrow.parquet as pq BASE = "https://huggingface.co/datasets/Estwld/empathetic_dialogues_llm/resolve/refs%2Fconvert%2Fparquet/default" SPLITS = {"train": "train", "val": "valid"} OUT_DIR = Path("data/sft") BOS = "" EOS = "" ROLES = ("<|user|>", "<|assistant|>") _MULTI = re.compile(r"\s+") _APOS = re.compile(r"\s+'\s*") _PUNCT = re.compile(r"\s+([.,!?;:])") _CODE = re.compile(r"`||def |import |class |printf|console\.log|#include| {out}") if __name__ == "__main__": main()