truncate

#!/usr/bin/env python3

"""
Truncates a sentence to a set length based on the number of tokens.
If "-l zh" is passed, Han characters count as tokens.
"""

import sys
import regex as re

def main(args):
    # TODO: should preserve the particular whitespaces
    for line in sys.stdin:
        if args.lang == "zh":
            count = 0
            for char in line.strip():
                if re.match(r'\p{Han}', char) or char == " ":
                    count += 1
                if count <= args.length:
                    print(char, end="")

            print()
        else:
            tokens = line.strip().split()
            print(" ".join(tokens[0:args.length]))

if __name__ == "__main__":
    import argparse

    parser = argparse.ArgumentParser()
    parser.add_argument("length", type=int)
    parser.add_argument("--lang", "-l", default="en")
    args = parser.parse_args()

    main(args)