max_col_widths.py

#!/usr/bin/env python
import functools
import argparse
import json

from pathlib import Path
import typing

import csv


def optimal_widths(in_file: typing.IO, delimiter: str):
    reader = csv.reader(in_file, delimiter=delimiter)
    headers = next(reader)
    widths = functools.reduce(
        lambda w, r: list(map(max, zip(w, map(len, r)))),
        reader,
        [0] * len(headers),
    )
    return dict(zip(headers, widths))


DESCRIPTION = """\
Determine max column widths for CSV.
"""

if __name__ == "__main__":
    parser = argparse.ArgumentParser("clean_roll", description=DESCRIPTION)

    parser.add_argument("in_path", type=Path)
    parser.add_argument("-d", "--delim", default="|")

    args = parser.parse_args()

    with open(args.in_path, "r") as f:
        res = optimal_widths(f, args.delim)
        print(json.dumps(res, indent=2))