find_best_caesar_break_parameters.py

   1 import random
   2 import collections
   3 from cipher import *
   4 from cipherbreak import *
   5
   6 print('Loading...')
   7
   8 corpus = sanitise(''.join([open('shakespeare.txt', 'r').read(),
   9     open('sherlock-holmes.txt', 'r').read(),
  10     open('war-and-peace.txt', 'r').read()]))
  11 corpus_length = len(corpus)
  12
  13 euclidean_scaled_english_counts = norms.euclidean_scale(english_counts)
  14
  15 metrics = [{'func': norms.l1, 'name': 'l1'},
  16     {'func': norms.l2, 'name': 'l2'},
  17     {'func': norms.l3, 'name': 'l2'},
  18     {'func': norms.cosine_distance, 'name': 'cosine_distance'},
  19     {'func': norms.harmonic_mean, 'name': 'harminic_mean'},
  20     {'func': norms.geometric_mean, 'name': 'geometric_mean'},
  21     {'func': norms.inverse_log_pl, 'name': 'inverse_log_pl'}]
  22 scalings = [{'corpus_frequency': normalised_english_counts,
  23          'scaling': norms.normalise,
  24          'name': 'normalised'},
  25         {'corpus_frequency': euclidean_scaled_english_counts,
  26          'scaling': norms.euclidean_scale,
  27          'name': 'euclidean_scaled'},
  28          {'corpus_frequency': normalised_english_counts,
  29          'scaling': norms.identity_scale,
  30          'name': 'normalised_with_identity'}]
  31 message_lengths = [300, 100, 50, 30, 20, 10, 5]
  32
  33 trials = 5000
  34
  35 # rebuild with itertools.product and itertools.starmap
  36 # e.g. results = starmap(one_trial, product(metrics, scalings, message_lengths))
  37 # ... which would then be easy parallelise.
  38
  39 print('Starting:', end='', flush=True)
  40 with open('caesar_break_parameter_trials.csv', 'w') as f:
  41     print('metric,scaling,message_length,score', file = f)
  42     scores = collections.defaultdict(int)
  43     for metric in metrics:
  44         scores[metric['name']] = collections.defaultdict(int)
  45         for scaling in scalings:
  46             scores[metric['name']][scaling['name']] = collections.defaultdict(int)
  47             for message_length in message_lengths:
  48                 for i in range(trials):
  49                     sample_start = random.randint(0, corpus_length - message_length)
  50                     sample = corpus[sample_start:(sample_start + message_length)]
  51                     key = random.randint(1, 25)
  52                     sample_ciphertext = caesar_encipher(sample, key)
  53                     (found_key, score) = caesar_break(sample_ciphertext,
  54                                                       metric=metric['func'],
  55                                                       target_counts=scaling['corpus_frequency'],
  56                                                       message_frequency_scaling=scaling['scaling'])
  57                     if found_key == key:
  58                         scores[metric['name']][scaling['name']][message_length] += 1
  59                 print('.', end='', flush=True)
  60                 print(', '.join([metric['name'],
  61                                  scaling['name'],
  62                                  str(message_length),
  63                                  str(scores[metric['name']][scaling['name']][message_length] / trials) ]),
  64                     file = f)
  65 print()