benchmark.py

from contextlib import contextmanager
from operator import itemgetter
from sys import argv, stdout
from time import time

@contextmanager
def duration(outfile=stdout):
    start = time()
    yield
    end = time()
    outfile.write(str(end - start) + '\n')

def groupby(seq, func):
    d = dict()
    for item in seq:
        d.setdefault(func(item), []).append(item)
    return d

with duration():
    with open(argv[1], 'r') as file:
        word_pairs = [line.strip().split(',') for line in file]
        result = groupby(word_pairs, itemgetter(0))