Added keybase signature
[cipher-training.git] / cipherbreak.py
1 """A set of functions to break the ciphers give in ciphers.py.
2 """
3
4 import string
5 import collections
6 import norms
7 import logging
8 import random
9 import math
10 from itertools import starmap
11 from segment import segment
12 from multiprocessing import Pool
13
14 import matplotlib.pyplot as plt
15
16 logger = logging.getLogger(__name__)
17 logger.addHandler(logging.FileHandler('cipher.log'))
18 logger.setLevel(logging.WARNING)
19 #logger.setLevel(logging.INFO)
20 #logger.setLevel(logging.DEBUG)
21
22 from cipher import *
23 from language_models import *
24
25 # To time a run:
26 #
27 # import timeit
28 # c5a = open('2012/5a.ciphertext', 'r').read()
29 # timeit.timeit('keyword_break(c5a)', setup='gc.enable() ; from __main__ import c5a ; from cipher import keyword_break', number=1)
30 # timeit.repeat('keyword_break_mp(c5a, chunksize=500)', setup='gc.enable() ; from __main__ import c5a ; from cipher import keyword_break_mp', repeat=5, number=1)
31
32 transpositions = collections.defaultdict(list)
33 for word in keywords:
34 transpositions[transpositions_of(word)] += [word]
35
36 def frequencies(text):
37 """Count the number of occurrences of each character in text
38
39 >>> sorted(frequencies('abcdefabc').items())
40 [('a', 2), ('b', 2), ('c', 2), ('d', 1), ('e', 1), ('f', 1)]
41 >>> sorted(frequencies('the quick brown fox jumped over the lazy ' \
42 'dog').items()) # doctest: +NORMALIZE_WHITESPACE
43 [(' ', 8), ('a', 1), ('b', 1), ('c', 1), ('d', 2), ('e', 4), ('f', 1),
44 ('g', 1), ('h', 2), ('i', 1), ('j', 1), ('k', 1), ('l', 1), ('m', 1),
45 ('n', 1), ('o', 4), ('p', 1), ('q', 1), ('r', 2), ('t', 2), ('u', 2),
46 ('v', 1), ('w', 1), ('x', 1), ('y', 1), ('z', 1)]
47 >>> sorted(frequencies('The Quick BROWN fox jumped! over... the ' \
48 '(9lazy) DOG').items()) # doctest: +NORMALIZE_WHITESPACE
49 [(' ', 8), ('!', 1), ('(', 1), (')', 1), ('.', 3), ('9', 1), ('B', 1),
50 ('D', 1), ('G', 1), ('N', 1), ('O', 2), ('Q', 1), ('R', 1), ('T', 1),
51 ('W', 1), ('a', 1), ('c', 1), ('d', 1), ('e', 4), ('f', 1), ('h', 2),
52 ('i', 1), ('j', 1), ('k', 1), ('l', 1), ('m', 1), ('o', 2), ('p', 1),
53 ('r', 1), ('t', 1), ('u', 2), ('v', 1), ('x', 1), ('y', 1), ('z', 1)]
54 >>> sorted(frequencies(sanitise('The Quick BROWN fox jumped! over... '\
55 'the (9lazy) DOG')).items()) # doctest: +NORMALIZE_WHITESPACE
56 [('a', 1), ('b', 1), ('c', 1), ('d', 2), ('e', 4), ('f', 1), ('g', 1),
57 ('h', 2), ('i', 1), ('j', 1), ('k', 1), ('l', 1), ('m', 1), ('n', 1),
58 ('o', 4), ('p', 1), ('q', 1), ('r', 2), ('t', 2), ('u', 2), ('v', 1),
59 ('w', 1), ('x', 1), ('y', 1), ('z', 1)]
60 >>> frequencies('abcdefabcdef')['x']
61 0
62 """
63 return collections.Counter(c for c in text)
64
65
66 def caesar_break(message, fitness=Pletters):
67 """Breaks a Caesar cipher using frequency analysis
68
69 >>> caesar_break('ibxcsyorsaqcheyklxivoexlevmrimwxsfiqevvmihrsasrxliwyrh' \
70 'ecjsppsamrkwleppfmergefifvmhixscsymjcsyqeoixlm') # doctest: +ELLIPSIS
71 (4, -130.849989015...)
72 >>> caesar_break('wxwmaxdgheetgwuxztgptedbgznitgwwhpguxyhkxbmhvvtlbhgtee' \
73 'raxlmhiixweblmxgxwmhmaxybkbgztgwztsxwbgmxgmert') # doctest: +ELLIPSIS
74 (19, -128.82410410...)
75 >>> caesar_break('yltbbqnqnzvguvaxurorgenafsbezqvagbnornfgsbevpnaabjurer' \
76 'svaquvzyvxrnznazlybequrvfohgriraabjtbaruraprur') # doctest: +ELLIPSIS
77 (13, -126.25403935...)
78 """
79 sanitised_message = sanitise(message)
80 best_shift = 0
81 best_fit = float('-inf')
82 for shift in range(26):
83 plaintext = caesar_decipher(sanitised_message, shift)
84 fit = fitness(plaintext)
85 logger.debug('Caesar break attempt using key {0} gives fit of {1} '
86 'and decrypt starting: {2}'.format(shift, fit,
87 plaintext[:50]))
88 if fit > best_fit:
89 best_fit = fit
90 best_shift = shift
91 logger.info('Caesar break best fit: key {0} gives fit of {1} and '
92 'decrypt starting: {2}'.format(best_shift, best_fit,
93 caesar_decipher(sanitised_message, best_shift)[:50]))
94 return best_shift, best_fit
95
96 def affine_break(message, fitness=Pletters):
97 """Breaks an affine cipher using frequency analysis
98
99 >>> affine_break('lmyfu bkuusd dyfaxw claol psfaom jfasd snsfg jfaoe ls ' \
100 'omytd jlaxe mh jm bfmibj umis hfsul axubafkjamx. ls kffkxwsd jls ' \
101 'ofgbjmwfkiu olfmxmtmwaokttg jlsx ls kffkxwsd jlsi zg tsxwjl. jlsx ' \
102 'ls umfjsd jlsi zg hfsqysxog. ls dmmdtsd mx jls bats mh bkbsf. ls ' \
103 'bfmctsd kfmyxd jls lyj, mztanamyu xmc jm clm cku tmmeaxw kj lai ' \
104 'kxd clm ckuxj.') # doctest: +ELLIPSIS
105 ((15, 22, True), -340.601181913...)
106 """
107 sanitised_message = sanitise(message)
108 best_multiplier = 0
109 best_adder = 0
110 best_one_based = True
111 best_fit = float("-inf")
112 for one_based in [True, False]:
113 for multiplier in [x for x in range(1, 26, 2) if x != 13]:
114 for adder in range(26):
115 plaintext = affine_decipher(sanitised_message,
116 multiplier, adder, one_based)
117 fit = fitness(plaintext)
118 logger.debug('Affine break attempt using key {0}x+{1} ({2}) '
119 'gives fit of {3} and decrypt starting: {4}'.
120 format(multiplier, adder, one_based, fit,
121 plaintext[:50]))
122 if fit > best_fit:
123 best_fit = fit
124 best_multiplier = multiplier
125 best_adder = adder
126 best_one_based = one_based
127 logger.info('Affine break best fit with key {0}x+{1} ({2}) gives fit of '
128 '{3} and decrypt starting: {4}'.format(
129 best_multiplier, best_adder, best_one_based, best_fit,
130 affine_decipher(sanitised_message, best_multiplier,
131 best_adder, best_one_based)[:50]))
132 return (best_multiplier, best_adder, best_one_based), best_fit
133
134 def keyword_break(message, wordlist=keywords, fitness=Pletters):
135 """Breaks a keyword substitution cipher using a dictionary and
136 frequency analysis.
137
138 >>> keyword_break(keyword_encipher('this is a test message for the ' \
139 'keyword decipherment', 'elephant', KeywordWrapAlphabet.from_last), \
140 wordlist=['cat', 'elephant', 'kangaroo']) # doctest: +ELLIPSIS
141 (('elephant', <KeywordWrapAlphabet.from_last: 2>), -52.834575011...)
142 """
143 best_keyword = ''
144 best_wrap_alphabet = True
145 best_fit = float("-inf")
146 for wrap_alphabet in KeywordWrapAlphabet:
147 for keyword in wordlist:
148 plaintext = keyword_decipher(message, keyword, wrap_alphabet)
149 fit = fitness(plaintext)
150 logger.debug('Keyword break attempt using key {0} (wrap={1}) '
151 'gives fit of {2} and decrypt starting: {3}'.format(
152 keyword, wrap_alphabet, fit,
153 sanitise(plaintext)[:50]))
154 if fit > best_fit:
155 best_fit = fit
156 best_keyword = keyword
157 best_wrap_alphabet = wrap_alphabet
158 logger.info('Keyword break best fit with key {0} (wrap={1}) gives fit of '
159 '{2} and decrypt starting: {3}'.format(best_keyword,
160 best_wrap_alphabet, best_fit, sanitise(
161 keyword_decipher(message, best_keyword,
162 best_wrap_alphabet))[:50]))
163 return (best_keyword, best_wrap_alphabet), best_fit
164
165 def keyword_break_mp(message, wordlist=keywords, fitness=Pletters,
166 number_of_solutions=1, chunksize=500):
167 """Breaks a keyword substitution cipher using a dictionary and
168 frequency analysis
169
170 >>> keyword_break_mp(keyword_encipher('this is a test message for the ' \
171 'keyword decipherment', 'elephant', KeywordWrapAlphabet.from_last), \
172 wordlist=['cat', 'elephant', 'kangaroo']) # doctest: +ELLIPSIS
173 (('elephant', <KeywordWrapAlphabet.from_last: 2>), -52.834575011...)
174 >>> keyword_break_mp(keyword_encipher('this is a test message for the ' \
175 'keyword decipherment', 'elephant', KeywordWrapAlphabet.from_last), \
176 wordlist=['cat', 'elephant', 'kangaroo'], \
177 number_of_solutions=2) # doctest: +ELLIPSIS, +NORMALIZE_WHITESPACE
178 [(('elephant', <KeywordWrapAlphabet.from_last: 2>), -52.834575011...),
179 (('elephant', <KeywordWrapAlphabet.from_largest: 3>), -52.834575011...)]
180 """
181 with Pool() as pool:
182 helper_args = [(message, word, wrap, fitness)
183 for word in wordlist
184 for wrap in KeywordWrapAlphabet]
185 # Gotcha: the helper function here needs to be defined at the top level
186 # (limitation of Pool.starmap)
187 breaks = pool.starmap(keyword_break_worker, helper_args, chunksize)
188 if number_of_solutions == 1:
189 return max(breaks, key=lambda k: k[1])
190 else:
191 return sorted(breaks, key=lambda k: k[1], reverse=True)[:number_of_solutions]
192
193 def keyword_break_worker(message, keyword, wrap_alphabet, fitness):
194 plaintext = keyword_decipher(message, keyword, wrap_alphabet)
195 fit = fitness(plaintext)
196 logger.debug('Keyword break attempt using key {0} (wrap={1}) gives fit of '
197 '{2} and decrypt starting: {3}'.format(keyword,
198 wrap_alphabet, fit, sanitise(plaintext)[:50]))
199 return (keyword, wrap_alphabet), fit
200
201 def monoalphabetic_break_hillclimbing(message, max_iterations=10000000,
202 fitness=Pletters):
203 ciphertext = unaccent(message).lower()
204 alphabet = list(string.ascii_lowercase)
205 random.shuffle(alphabet)
206 alphabet = ''.join(alphabet)
207 return monoalphabetic_break_hillclimbing_worker(ciphertext, alphabet,
208 max_iterations, fitness)
209
210 def monoalphabetic_break_hillclimbing_mp(message, workers=10,
211 max_iterations = 10000000, fitness=Pletters, chunksize=1):
212 worker_args = []
213 ciphertext = unaccent(message).lower()
214 for i in range(workers):
215 alphabet = list(string.ascii_lowercase)
216 random.shuffle(alphabet)
217 alphabet = ''.join(alphabet)
218 worker_args.append((ciphertext, alphabet, max_iterations, fitness))
219 with Pool() as pool:
220 breaks = pool.starmap(monoalphabetic_break_hillclimbing_worker,
221 worker_args, chunksize)
222 return max(breaks, key=lambda k: k[1])
223
224 def monoalphabetic_break_hillclimbing_worker(message, alphabet,
225 max_iterations, fitness):
226 def swap(letters, i, j):
227 if i > j:
228 i, j = j, i
229 if i == j:
230 return letters
231 else:
232 return (letters[:i] + letters[j] + letters[i+1:j] + letters[i] +
233 letters[j+1:])
234 best_alphabet = alphabet
235 best_fitness = float('-inf')
236 for i in range(max_iterations):
237 alphabet = swap(alphabet, random.randrange(26), random.randrange(26))
238 cipher_translation = ''.maketrans(string.ascii_lowercase, alphabet)
239 plaintext = message.translate(cipher_translation)
240 if fitness(plaintext) > best_fitness:
241 best_fitness = fitness(plaintext)
242 best_alphabet = alphabet
243 print(i, best_alphabet, best_fitness, plaintext)
244 return best_alphabet, best_fitness
245
246
247 def vigenere_keyword_break_mp(message, wordlist=keywords, fitness=Pletters,
248 chunksize=500):
249 """Breaks a vigenere cipher using a dictionary and frequency analysis.
250
251 >>> vigenere_keyword_break_mp(vigenere_encipher(sanitise('this is a test ' \
252 'message for the vigenere decipherment'), 'cat'), \
253 wordlist=['cat', 'elephant', 'kangaroo']) # doctest: +ELLIPSIS
254 ('cat', -52.947271216...)
255 """
256 with Pool() as pool:
257 helper_args = [(message, word, fitness)
258 for word in wordlist]
259 # Gotcha: the helper function here needs to be defined at the top level
260 # (limitation of Pool.starmap)
261 breaks = pool.starmap(vigenere_keyword_break_worker, helper_args,
262 chunksize)
263 return max(breaks, key=lambda k: k[1])
264 vigenere_keyword_break = vigenere_keyword_break_mp
265
266 def vigenere_keyword_break_worker(message, keyword, fitness):
267 plaintext = vigenere_decipher(message, keyword)
268 fit = fitness(plaintext)
269 logger.debug('Vigenere keyword break attempt using key {0} gives fit of '
270 '{1} and decrypt starting: {2}'.format(keyword,
271 fit, sanitise(plaintext)[:50]))
272 return keyword, fit
273
274
275 def vigenere_frequency_break(message, max_key_length=20, fitness=Pletters):
276 """Breaks a Vigenere cipher with frequency analysis
277
278 >>> vigenere_frequency_break(vigenere_encipher(sanitise("It is time to " \
279 "run. She is ready and so am I. I stole Daniel's pocketbook this " \
280 "afternoon when he left his jacket hanging on the easel in the " \
281 "attic. I jump every time I hear a footstep on the stairs, " \
282 "certain that the theft has been discovered and that I will " \
283 "be caught. The SS officer visits less often now that he is " \
284 "sure"), 'florence')) # doctest: +ELLIPSIS
285 ('florence', -307.5473096791...)
286 """
287 def worker(message, key_length, fitness):
288 splits = every_nth(sanitised_message, key_length)
289 key = ''.join([chr(caesar_break(s)[0] + ord('a')) for s in splits])
290 plaintext = vigenere_decipher(message, key)
291 fit = fitness(plaintext)
292 return key, fit
293 sanitised_message = sanitise(message)
294 results = starmap(worker, [(sanitised_message, i, fitness)
295 for i in range(1, max_key_length+1)])
296 return max(results, key=lambda k: k[1])
297
298
299 def beaufort_frequency_break(message, max_key_length=20, fitness=Pletters):
300 """Breaks a Beaufort cipher with frequency analysis
301
302 >>> beaufort_frequency_break(beaufort_encipher(sanitise("It is time to " \
303 "run. She is ready and so am I. I stole Daniel's pocketbook this " \
304 "afternoon when he left his jacket hanging on the easel in the " \
305 "attic. I jump every time I hear a footstep on the stairs, " \
306 "certain that the theft has been discovered and that I will " \
307 "be caught. The SS officer visits less often now " \
308 "that he is sure"), 'florence')) # doctest: +ELLIPSIS
309 ('florence', -307.5473096791...)
310 """
311 def worker(message, key_length, fitness):
312 splits = every_nth(sanitised_message, key_length)
313 key = ''.join([chr(-caesar_break(s)[0] % 26 + ord('a'))
314 for s in splits])
315 plaintext = beaufort_decipher(message, key)
316 fit = fitness(plaintext)
317 return key, fit
318 sanitised_message = sanitise(message)
319 results = starmap(worker, [(sanitised_message, i, fitness)
320 for i in range(1, max_key_length+1)])
321 return max(results, key=lambda k: k[1])
322
323
324 def column_transposition_break_mp(message, translist=transpositions,
325 fitness=Pbigrams, chunksize=500):
326 """Breaks a column transposition cipher using a dictionary and
327 n-gram frequency analysis
328
329 >>> column_transposition_break_mp(column_transposition_encipher(sanitise( \
330 "It is a truth universally acknowledged, that a single man in \
331 possession of a good fortune, must be in want of a wife. However \
332 little known the feelings or views of such a man may be on his \
333 first entering a neighbourhood, this truth is so well fixed in \
334 the minds of the surrounding families, that he is considered the \
335 rightful property of some one or other of their daughters."), \
336 'encipher'), \
337 translist={(2, 0, 5, 3, 1, 4, 6): ['encipher'], \
338 (5, 0, 6, 1, 3, 4, 2): ['fourteen'], \
339 (6, 1, 0, 4, 5, 3, 2): ['keyword']}) # doctest: +ELLIPSIS
340 (((2, 0, 5, 3, 1, 4, 6), False, False), -709.4646722...)
341 >>> column_transposition_break_mp(column_transposition_encipher(sanitise( \
342 "It is a truth universally acknowledged, that a single man in \
343 possession of a good fortune, must be in want of a wife. However \
344 little known the feelings or views of such a man may be on his \
345 first entering a neighbourhood, this truth is so well fixed in \
346 the minds of the surrounding families, that he is considered the \
347 rightful property of some one or other of their daughters."), \
348 'encipher'), \
349 translist={(2, 0, 5, 3, 1, 4, 6): ['encipher'], \
350 (5, 0, 6, 1, 3, 4, 2): ['fourteen'], \
351 (6, 1, 0, 4, 5, 3, 2): ['keyword']}, \
352 fitness=Ptrigrams) # doctest: +ELLIPSIS
353 (((2, 0, 5, 3, 1, 4, 6), False, False), -997.0129085...)
354 """
355 with Pool() as pool:
356 helper_args = [(message, trans, fillcolumnwise, emptycolumnwise,
357 fitness)
358 for trans in translist.keys()
359 for fillcolumnwise in [True, False]
360 for emptycolumnwise in [True, False]]
361 # Gotcha: the helper function here needs to be defined at the top level
362 # (limitation of Pool.starmap)
363 breaks = pool.starmap(column_transposition_break_worker,
364 helper_args, chunksize)
365 return max(breaks, key=lambda k: k[1])
366 column_transposition_break = column_transposition_break_mp
367
368 def column_transposition_break_worker(message, transposition,
369 fillcolumnwise, emptycolumnwise, fitness):
370 plaintext = column_transposition_decipher(message, transposition,
371 fillcolumnwise=fillcolumnwise, emptycolumnwise=emptycolumnwise)
372 fit = fitness(sanitise(plaintext))
373 logger.debug('Column transposition break attempt using key {0} '
374 'gives fit of {1} and decrypt starting: {2}'.format(
375 transposition, fit,
376 sanitise(plaintext)[:50]))
377 return (transposition, fillcolumnwise, emptycolumnwise), fit
378
379
380 def scytale_break_mp(message, max_key_length=20,
381 fitness=Pbigrams, chunksize=500):
382 """Breaks a scytale cipher using a range of lengths and
383 n-gram frequency analysis
384
385 >>> scytale_break_mp(scytale_encipher(sanitise( \
386 "It is a truth universally acknowledged, that a single man in \
387 possession of a good fortune, must be in want of a wife. However \
388 little known the feelings or views of such a man may be on his \
389 first entering a neighbourhood, this truth is so well fixed in \
390 the minds of the surrounding families, that he is considered the \
391 rightful property of some one or other of their daughters."), \
392 5)) # doctest: +ELLIPSIS
393 (5, -709.4646722...)
394 >>> scytale_break_mp(scytale_encipher(sanitise( \
395 "It is a truth universally acknowledged, that a single man in \
396 possession of a good fortune, must be in want of a wife. However \
397 little known the feelings or views of such a man may be on his \
398 first entering a neighbourhood, this truth is so well fixed in \
399 the minds of the surrounding families, that he is considered the \
400 rightful property of some one or other of their daughters."), \
401 5), \
402 fitness=Ptrigrams) # doctest: +ELLIPSIS
403 (5, -997.0129085...)
404 """
405 with Pool() as pool:
406 helper_args = [(message, trans, False, True, fitness)
407 for trans in
408 [[col for col in range(math.ceil(len(message)/rows))]
409 for rows in range(1,max_key_length+1)]]
410 # Gotcha: the helper function here needs to be defined at the top level
411 # (limitation of Pool.starmap)
412 breaks = pool.starmap(column_transposition_break_worker,
413 helper_args, chunksize)
414 best = max(breaks, key=lambda k: k[1])
415 return math.trunc(len(message) / len(best[0][0])), best[1]
416 scytale_break = scytale_break_mp
417
418
419 def railfence_break(message, max_key_length=20,
420 fitness=Pletters, chunksize=500):
421 """Breaks a hill cipher using a matrix of given rank and letter frequencies
422
423
424 """
425
426 sanitised_message = sanitise(message)
427 results = starmap(worker, [(sanitised_message, i, fitness)
428 for i in range(2, max_key_length+1)])
429 return max(results, key=lambda k: k[1])
430
431
432 def railfence_break(message, max_key_length=20,
433 fitness=Pbigrams, chunksize=500):
434 """Breaks a railfence cipher using a range of lengths and
435 n-gram frequency analysis
436
437 >>> railfence_break(railfence_encipher(sanitise( \
438 "It is a truth universally acknowledged, that a single man in \
439 possession of a good fortune, must be in want of a wife. However \
440 little known the feelings or views of such a man may be on his \
441 first entering a neighbourhood, this truth is so well fixed in \
442 the minds of the surrounding families, that he is considered the \
443 rightful property of some one or other of their daughters."), \
444 7)) # doctest: +ELLIPSIS
445 (7, -709.46467226...)
446 >>> railfence_break(railfence_encipher(sanitise( \
447 "It is a truth universally acknowledged, that a single man in \
448 possession of a good fortune, must be in want of a wife. However \
449 little known the feelings or views of such a man may be on his \
450 first entering a neighbourhood, this truth is so well fixed in \
451 the minds of the surrounding families, that he is considered the \
452 rightful property of some one or other of their daughters."), \
453 7), \
454 fitness=Ptrigrams) # doctest: +ELLIPSIS
455 (7, -997.0129085...)
456 """
457 def worker(message, height, fitness):
458 plaintext = railfence_decipher(message, height)
459 fit = fitness(plaintext)
460 return height, fit
461
462 sanitised_message = sanitise(message)
463 results = starmap(worker, [(sanitised_message, i, fitness)
464 for i in range(2, max_key_length+1)])
465 return max(results, key=lambda k: k[1])
466
467 def amsco_break(message, translist=transpositions, patterns = [(1, 2), (2, 1)],
468 fitness=Pbigrams,
469 chunksize=500):
470 """Breaks an AMSCO transposition cipher using a dictionary and
471 n-gram frequency analysis
472
473 >>> amsco_break(amsco_transposition_encipher(sanitise( \
474 "It is a truth universally acknowledged, that a single man in \
475 possession of a good fortune, must be in want of a wife. However \
476 little known the feelings or views of such a man may be on his \
477 first entering a neighbourhood, this truth is so well fixed in \
478 the minds of the surrounding families, that he is considered the \
479 rightful property of some one or other of their daughters."), \
480 'encipher'), \
481 translist={(2, 0, 5, 3, 1, 4, 6): ['encipher'], \
482 (5, 0, 6, 1, 3, 4, 2): ['fourteen'], \
483 (6, 1, 0, 4, 5, 3, 2): ['keyword']}, \
484 patterns=[(1, 2)]) # doctest: +ELLIPSIS
485 (((2, 0, 5, 3, 1, 4, 6), (1, 2)), -709.4646722...)
486 >>> amsco_break(amsco_transposition_encipher(sanitise( \
487 "It is a truth universally acknowledged, that a single man in \
488 possession of a good fortune, must be in want of a wife. However \
489 little known the feelings or views of such a man may be on his \
490 first entering a neighbourhood, this truth is so well fixed in \
491 the minds of the surrounding families, that he is considered the \
492 rightful property of some one or other of their daughters."), \
493 'encipher', fillpattern=(2, 1)), \
494 translist={(2, 0, 5, 3, 1, 4, 6): ['encipher'], \
495 (5, 0, 6, 1, 3, 4, 2): ['fourteen'], \
496 (6, 1, 0, 4, 5, 3, 2): ['keyword']}, \
497 patterns=[(1, 2), (2, 1)], fitness=Ptrigrams) # doctest: +ELLIPSIS
498 (((2, 0, 5, 3, 1, 4, 6), (2, 1)), -997.0129085...)
499 """
500 with Pool() as pool:
501 helper_args = [(message, trans, pattern, fitness)
502 for trans in translist.keys()
503 for pattern in patterns]
504 # Gotcha: the helper function here needs to be defined at the top level
505 # (limitation of Pool.starmap)
506 breaks = pool.starmap(amsco_break_worker, helper_args, chunksize)
507 return max(breaks, key=lambda k: k[1])
508
509 def amsco_break_worker(message, transposition,
510 pattern, fitness):
511 plaintext = amsco_transposition_decipher(message, transposition,
512 fillpattern=pattern)
513 fit = fitness(sanitise(plaintext))
514 logger.debug('AMSCO transposition break attempt using key {0} and pattern'
515 '{1} gives fit of {2} and decrypt starting: {3}'.format(
516 transposition, pattern, fit,
517 sanitise(plaintext)[:50]))
518 return (transposition, pattern), fit
519
520
521 def hill_break(message, matrix_size=2, fitness=Pletters,
522 number_of_solutions=1, chunksize=500):
523
524 all_matrices = [np.matrix(list(m))
525 for m in itertools.product([list(r)
526 for r in itertools.product(range(26), repeat=matrix_size)],
527 repeat=matrix_size)]
528 valid_matrices = [m for m, d in
529 zip(all_matrices, (int(round(linalg.det(m))) for m in all_matrices))
530 if d != 0
531 if d % 2 != 0
532 if d % 13 != 0 ]
533 with Pool() as pool:
534 helper_args = [(message, matrix, fitness)
535 for matrix in valid_matrices]
536 # Gotcha: the helper function here needs to be defined at the top level
537 # (limitation of Pool.starmap)
538 breaks = pool.starmap(hill_break_worker, helper_args, chunksize)
539 if number_of_solutions == 1:
540 return max(breaks, key=lambda k: k[1])
541 else:
542 return sorted(breaks, key=lambda k: k[1], reverse=True)[:number_of_solutions]
543
544 def hill_break_worker(message, matrix, fitness):
545 plaintext = hill_decipher(matrix, message)
546 fit = fitness(plaintext)
547 logger.debug('Hill cipher break attempt using key {0} gives fit of '
548 '{1} and decrypt starting: {2}'.format(matrix,
549 fit, sanitise(plaintext)[:50]))
550 return matrix, fit
551
552
553 def pocket_enigma_break_by_crib(message, wheel_spec, crib, crib_position):
554 """Break a pocket enigma using a crib (some plaintext that's expected to
555 be in a certain position). Returns a list of possible starting wheel
556 positions that could produce the crib.
557
558 >>> pocket_enigma_break_by_crib('kzpjlzmoga', 1, 'h', 0)
559 ['a', 'f', 'q']
560 >>> pocket_enigma_break_by_crib('kzpjlzmoga', 1, 'he', 0)
561 ['a']
562 >>> pocket_enigma_break_by_crib('kzpjlzmoga', 1, 'll', 2)
563 ['a']
564 >>> pocket_enigma_break_by_crib('kzpjlzmoga', 1, 'l', 2)
565 ['a']
566 >>> pocket_enigma_break_by_crib('kzpjlzmoga', 1, 'l', 3)
567 ['a', 'j', 'n']
568 >>> pocket_enigma_break_by_crib('aaaaa', 1, 'l', 3)
569 []
570 """
571 pe = PocketEnigma(wheel=wheel_spec)
572 possible_positions = []
573 for p in string.ascii_lowercase:
574 pe.set_position(p)
575 plaintext = pe.decipher(message)
576 if plaintext[crib_position:crib_position+len(crib)] == crib:
577 possible_positions += [p]
578 return possible_positions
579
580
581 def plot_frequency_histogram(freqs, sort_key=None):
582 x = range(len(freqs.keys()))
583 y = [freqs[l] for l in sorted(freqs.keys(), key=sort_key)]
584 f = plt.figure()
585 ax = f.add_axes([0.1, 0.1, 0.9, 0.9])
586 ax.bar(x, y, align='center')
587 ax.set_xticks(x)
588 ax.set_xticklabels(sorted(freqs.keys(), key=sort_key))
589 f.show()
590
591
592 if __name__ == "__main__":
593 import doctest
594 doctest.testmod()