filter.py

import sys
import os
import collections
import itertools
import pprint

filenames = list(os.listdir(sys.argv[1]))
# filenames = list(filter(lambda x: any(map(lambda y: x.startswith('%s@' % y), andrew_ids)), filenames))
grouped_filenames = collections.defaultdict(list)
for filename in filenames:
    grouped_filenames[filename.split('@')[0]].append(filename)
# pprint.pprint(grouped_files)
final_submission_nums = dict(map(lambda x: (x[0], max(map(lambda y: int(y.split('_')[1]), x[1]))), grouped_filenames.items()))
final_submissions = dict(map(lambda x: (x[0], list(filter(lambda y: int(y.split('_')[1]) == final_submission_nums[x[0]], x[1]))), grouped_filenames.items()))
final_filenames = list(itertools.chain(*final_submissions.values()))
for filename in filenames:
	if filename not in final_filenames:
		os.remove(os.path.join(sys.argv[1], filename))