words/scripts/gen_wordlist.py

#!/usr/bin/env python3

import sys
import json


def get_args():
    """ Get command line arguments """

    import argparse
    parser = argparse.ArgumentParser()
    parser.add_argument('wordlist')
    parser.add_argument('word_length', type=int)
    parser.add_argument('frequency_min', type=int)
    return parser.parse_args()


def main(args):
    """ Entry point for script """
    with open(args.wordlist) as fp:
        words = [ (word[1], int(word[4])) for word in [ word.lower().split('\t') for word in fp.read().strip().split('\n') ] ]

    words = [ word[0] for word in words if word[1] > args.frequency_min and word[0].isalpha() and len(word[0]) == args.word_length ]

    words.sort(key=lambda word: word[1])
    print(f"wordlist = {json.dumps(words)}")
    print(f"{args}", file=sys.stderr)
    print(f"{len(words)=}", file=sys.stderr)

    return 0


if __name__ == '__main__':
    try:
        sys.exit(main(get_args()))
    except KeyboardInterrupt:
        sys.exit(0)