generate logs w/ pii

gen_fake_data.py

import time
import json

from faker import Factory

fake = Factory.create()
start_time = time.time() + 900

while time.time() < start_time:
    p = fake.profile()
    p.pop('website')
    p.pop('current_location')

    s = time.strftime("%Y-%m-%dT%H:%M:%S", time.gmtime(time.time())) + ' - INFO - '
    new_dict = {}
    for k, v in p.items():
        new_k = str(k).replace('\n', ' ')
        new_v = str(v).replace('\n', ' ')
        new_dict[new_k] = new_v

    j = json.dumps(new_dict)
    s += j
    s += '\n'

    w = open('/home/vagrant/logs/demo/demo.log',"a+")
    w.write(s)
    w.close()
    time.sleep(0.5)

#   Global processing rules that are applied to all the logs. The available rules are
#   "exclude_at_match", "include_at_match" and "mask_sequences".
#  processing_rules:
#    - type: exclude_at_match
#      name: exclude_sensitive_info
#      pattern: \w*sensitive\-info*\w
#    - type: mask_sequences
#      name: social_security_number
#      pattern: (\d{3}-?\d{2}-?\d{4})
#      replace_placeholder: "XXX-XX-XXXX"
#    - type: mask_sequences
#      name: mask_credit_card
#      pattern: (?:4[0-9]{12}(?:[0-9]{3})?|[25][1-7][0-9]{14}|6(?:011|5[0-9][0-9])[0-9]{12}|3[47][0-9]{13}|3(?:0[0-5]|[68][0-9])[0-9]{11}|(?:2131|1800|35\d{3})\d{11})
#      replace_placeholder: "[Credit Card Number]"