find_best_caesar_break_parameters.py

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