gctools/misc/csv_trim.py

import os
import re
from datetime import datetime
from pathlib import Path

MAX_AGE = datetime.now().timestamp() - 12 * 60 * 60


def csv_trim(dirname: str = "misc/data"):
    if Path(dirname).is_file():
        csv_trim_file(Path(dirname), True)
    else:
        for csv_file in Path(dirname).glob("*.csv"):
            csv_trim_file(csv_file)


def csv_trim_file(csv_file: Path, ignore_age: bool = False):
    temp_file = Path(str(csv_file) + ".tmp")
    file_mtime = csv_file.stat().st_mtime
    if not ignore_age and file_mtime < MAX_AGE:
        return
    print(csv_file.name)
    with open(csv_file, "r", encoding="latin-1", errors="ignore") as frh:
        with open(temp_file, "w", encoding="latin-1") as fwh:
            for line in frh.readlines():
                fwh.write(re.sub(r"[ ]+\t", "\t", line))

    os.utime(temp_file, (file_mtime, file_mtime))

    csv_file.unlink()
    temp_file.rename(csv_file)