Randomly sample a large ass file

sample.py

import argparse
import mmap
import random


def get_line_count(filename):
    linecount = 0

    with open(args.filename, "r+b") as f:
        mm = mmap.mmap(f.fileno(), 0, prot=mmap.PROT_READ)
        try:
            for line in iter(mm.readline, ""):
                linecount += 1
        finally:
            mm.close()

    return linecount


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument("filename", type=str, help="name of file to sample")
    parser.add_argument("size", type=int, help="sample size")
    args = parser.parse_args()

    linecount = get_line_count(args.filename)
    lines_to_grab = set(random.sample(xrange(linecount), args.size))

    with open(args.filename, 'r', 1) as f, open('outfile.csv', 'w') as of:
        for l, line in enumerate(f):
            if l in lines_to_grab:
                of.write(line)