remove-blacklisted.py

#
# This script creates a file with entries from 'file_all_entries_path' file that are
# not blacklisted by entries in 'blacklisted_entries_path' file.
#
# Usage:
#   python remove-blacklisted.py

import datetime
import time


def build_timestamp():
  t = time.time()
  return datetime.datetime.fromtimestamp(t).strftime('%Y%m%d-%H%M%S')

def blacklisted_by_arr(e, arr):
  for a in arr:
    if a.strip() in e.strip():
      return True
  return False

file_all_entries_path = 'all-lob-control.txt'
blacklisted_entries_path = 'lob-control-remove.txt'
output_file_path = 'only-wanted-entries-' + build_timestamp() + '.txt'

control_all_file = open(file_all_entries_path, 'r')
blacklisted_entries_file = open(blacklisted_entries_path, 'r')
output_file = open(output_file_path, 'w')

control_all = control_all_file.read().splitlines()
control_all_file.close()

blacklisted_entries = blacklisted_entries_file.read().splitlines()
blacklisted_entries_file.close()

size_a = len(control_all)
size_b = len(blacklisted_entries)
print str(size_a) + ' <- size'
print str(size_b) + ' <- remove size'
print str(size_a - size_b) + ' <- expected'

for line in control_all:
  if not blacklisted_by_arr(line, blacklisted_entries):
    output_file.write(line + '\n')
  else:
    print '- ' + line

output_file.close()