Page MenuHomec4science
Files F102782754

run_anonymization.py
No OneTemporary
Actions

File Metadata

Created
Mon, Feb 24, 03:36

run_anonymization.py
View Options

#!/usr/bin/env python3
import random, datetime, sys, csv
import string
from random import randrange, randint
import os
import pickle
import hashlib
date_start = datetime.date(2000, 1, 1)
date_period = 365 * 17
# Reads in emails.txt and movies.txt and creates 'nbr_movies' entries for each
# email.
# Returns the database, the emails and the movies in the following format:
# [ [ user, movie, date, grade ], ... ]
def create_db(nbr_movies):
with open("emails.txt") as f:
emails = f.read().split("\n");
while "" in emails:
emails.remove("")
with open("movies.txt") as f:
movies = f.read().split("\n");
while "" in movies:
movies.remove("")
db = []
for email in emails:
movies_index = list(range(0, len(movies)))
random.shuffle(movies_index)
for i, f in enumerate(movies_index[0:nbr_movies]):
dat = date_start + datetime.timedelta(randint(1, date_period))
db.append(
[email, movies[f], dat.strftime("%Y/%m/%d"), randint(1, 5)])
return db, emails, movies
# Anonymize the given database, but still let the get_movies_with_rating
# function give the right answers.
def anonymize_1(db):
#just overwrite email and dates to anonymize them completely, do not need them for movie ratings
for i,_ in enumerate(db):
#name
db[i][0] = '*'
#date
db[i][2] = '*'
return db
# For a given anonymized-database and a rating, this function should return
# the films with the given rating.
def get_movies_with_rating(anon, rating):
movies = []
for movie in anon:
if rating == movie[3] and not movie[1] in movies:
movies.append(movie[1])
return movies
# A bit lesser anonymization than anonymize_1, but still no date. The returned
# database should have enough information to be used by get_top_rated. If you
# use a too simple hashing-function like sha-256, the result will be rejected.
def anonymize_2(db):
for i,_ in enumerate(db):
db[i][0] = str(i)
db[i][2] = '*'
return db
# get_top_rated searches for all users having rated a movie and searches their
# top-rated movie(s). It returns a list of all found movies, also doubles!
def get_top_rated(anon, movie):
users = {}
#each user will have a rating dictionary with its movies
users_who_rated = []
for line in anon:
email = line[0]
movie_an = line[1]
rating = line[3]
if email in users:
users[email][rating].append(movie_an)
else:
#create rating dictionary for that movie, add all keys to prevent key error
rating_dict = {1: [], 2: [], 3: [], 4: [], 5: []}
rating_dict[rating] = [movie_an]
users[email] = rating_dict
#save the users that have rated the given movie to check their ratings later
if movie == movie_an:
users_who_rated += [email]
movies = []
for user in users_who_rated:
#pick all movies from best rating
ratings = users[user]
if ratings[5]:
movies += ratings[5]
elif ratings[4]:
movies += ratings[4]
elif ratings[3]:
movies += ratings[3]
elif ratings[2]:
movies += ratings[2]
else:
movies += ratings[1]
return movies
# This is called when you start the script on localhost, and when the
# checker wants to run your functions.
if __name__ == "__main__":
# This part can be modified at your convenience.
"""if len(sys.argv) == 1:
print("Testing mode")
#db, emails, movies = create_db(10000000)
anon_db1 = anonymize_1(db)
anon_db2 = anonymize_2(db)"""
# If you modify this part, don't complain if it doesn't work anymore!
# This part is used to communicate with the verification-script. So you
# should not touch it (unless you're looking for a bug to exploit the
# verification script - but we didn't plan to put one in there).
if len(sys.argv) >= 3:
db_file, ex = sys.argv[1:3]
with open(db_file) as f:
db = list(csv.reader(f, skipinitialspace=True))
# Get nice ints for comparisons
for i, line in enumerate(db):
db[i][3] = int(line[3])
result = []
if ex == "ex1aa":
result = anonymize_1(db)
elif ex == "ex1ag":
rating = int(sys.argv[3])
result = [get_movies_with_rating(db, rating)]
elif ex == "ex1ba":
result = anonymize_2(db)
elif ex == "ex1bg":
movie = sys.argv[3]
result = [get_top_rated(db, movie)]
with open("/tmp/student.csv", "w") as f:
writer = csv.writer(f)
writer.writerows(iter(result))

Event Timeline

c4science · Help