Questions about the server

[dat-lequoc]

Hi, I have two questions.

1. The "DEL" row will be modified by the server before it public anonymized data for the attack phrase? Or does the server shuffle just rows without changing these lines?

2. Could you upload the script which checks the exception below, please?
   

[ceichler]

After verification with Pr. Nguyen regarding the first point, this can be configured on the server (keeping the DEL line or removing them in the DB provided to attackers).

It is therefore advised to fill a del line with garbage (that respect formatting)

[beng-git]

Il s'agit du fichier Footprint.py qui calcule le fichier F :

L'erreur correspondante est l'erreur -5

import json
import csv
from threading import Thread
from statistics import mean
from statistics import median
from Utility import *
from Utils import *
from datetime import date

#################################
#        Footprint Thread       #
#################################

# Generating the submitted file footprint for assessment and computes teh utility
class Footprint(Thread):
    def __init__(self, input, origin, footprint, dbconn):
        Thread.__init__(self)
        self.input = input
        self.origin = origin
        self.footprint = footprint
        self.exception = 0
        self.dbconn = dbconn


    def run(self):
        fd_nona_file = open(self.origin, "r")
        fd_anon_file = open(self.input, "r")
        nona_reader = csv.reader(fd_nona_file, delimiter=separator)
        anon_reader = csv.reader(fd_anon_file, delimiter=separator)
        found_ids_weeks = {}  # Dictionary {key = user id : value = list of months in which the user appears}
        linktable = {}  # Dictionary {key = user id : value = dictionary {key = month : value = list of anonymized ids}}
        index = 0
        for row1, row2 in zip(nona_reader, anon_reader):
            # Reads simultaneously both origin and edited file
            index += 1
            if row2[0]:
                if row2[0] != "DEL":
                    try: #Computing both week numbers
                        y2, m2, d2 = row2[1][0:10].split("-")
                        calendar2 = date(int(y2), int(m2), int(d2)).isocalendar()
                        weeknum2 = f"{calendar2[0]}-{calendar2[1]}"
                    except:
                        self.exception = (-10, index)
                        return
                    y1, m1, d1 = row1[1][0:10].split("-")
                    calendar1 = date(int(y1), int(m1), int(d1)).isocalendar()
                    weeknum1 = f"{calendar1[0]}-{calendar1[1]}"
                    if weeknum1 == weeknum2: #Checking if the week number is the same in both files
                        if row1[0] not in found_ids_weeks.keys() and weeknum1 == weeknum2: #Filling the correction table
                            found_ids_weeks[row1[0]] = [weeknum1]
                            linktable[row1[0]] = {}
                            linktable[row1[0]][weeknum1] = [row2[0]]
                        elif weeknum1 not in found_ids_weeks[row1[0]] and weeknum1 == weeknum2:
                            found_ids_weeks[row1[0]] += [weeknum1]
                            linktable[row1[0]][weeknum1] = [row2[0]]
                        else:
                            if linktable[row1[0]][weeknum1][0] != row2[0]:
                                # 2 pseudo IDs are standing for the same user, the same week
                                self.exception = (-5, index)
                                return
                    else:
                        self.exception = (-5, index)
                        return
            else :
                self.exception = (-6, index)
                return

        # Turns the python table into a json table
        with open(self.footprint, 'w') as result:
            json.dump(linktable, result)

        #Update the database
        self.dbconn.cursor().execute(f"UPDATE anonymisation \
                                       set status='Génération de la correction terminée' \
                                       where fileLink='{self.input.split('/')[1]}'")
        self.dbconn.commit()

    def result(self):
        return self.exception