X-Git-Url: https://git.njae.me.uk/?a=blobdiff_plain;f=cipherbreak.py;h=02e63c5ffd4a8f0b375e2144d2d8278c8d648fd6;hb=388422468f7dcf309d03005822ba8be9d2709b6e;hp=7665d46de21b45d9dc3daf51b2314460668bb6ab;hpb=da4054473cb34e7b1b62532bcc452e66f455df5c;p=cipher-training.git diff --git a/cipherbreak.py b/cipherbreak.py index 7665d46..02e63c5 100644 --- a/cipherbreak.py +++ b/cipherbreak.py @@ -199,23 +199,27 @@ def keyword_break_worker(message, keyword, wrap_alphabet, fitness): return (keyword, wrap_alphabet), fit def monoalphabetic_break_hillclimbing(message, max_iterations=10000000, - fitness=Pletters): + alphabet=None, fitness=Pletters): ciphertext = unaccent(message).lower() - alphabet = list(string.ascii_lowercase) - random.shuffle(alphabet) - alphabet = ''.join(alphabet) + if not alphabet: + alphabet = list(string.ascii_lowercase) + random.shuffle(alphabet) + alphabet = cat(alphabet) return monoalphabetic_break_hillclimbing_worker(ciphertext, alphabet, max_iterations, fitness) def monoalphabetic_break_hillclimbing_mp(message, workers=10, - max_iterations = 10000000, fitness=Pletters, chunksize=1): + max_iterations = 10000000, alphabet=None, fitness=Pletters, chunksize=1): worker_args = [] ciphertext = unaccent(message).lower() for i in range(workers): - alphabet = list(string.ascii_lowercase) - random.shuffle(alphabet) - alphabet = ''.join(alphabet) - worker_args.append((ciphertext, alphabet, max_iterations, fitness)) + if alphabet: + this_alphabet = alphabet + else: + this_alphabet = list(string.ascii_lowercase) + random.shuffle(this_alphabet) + this_alphabet = cat(this_alphabet) + worker_args.append((ciphertext, this_alphabet, max_iterations, fitness)) with Pool() as pool: breaks = pool.starmap(monoalphabetic_break_hillclimbing_worker, worker_args, chunksize) @@ -244,6 +248,83 @@ def monoalphabetic_break_hillclimbing_worker(message, alphabet, return best_alphabet, best_fitness +def vigenere_keyword_break_mp(message, wordlist=keywords, fitness=Pletters, + chunksize=500): + """Breaks a vigenere cipher using a dictionary and frequency analysis. + + >>> vigenere_keyword_break_mp(vigenere_encipher(sanitise('this is a test ' \ + 'message for the vigenere decipherment'), 'cat'), \ + wordlist=['cat', 'elephant', 'kangaroo']) # doctest: +ELLIPSIS + ('cat', -52.947271216...) + """ + with Pool() as pool: + helper_args = [(message, word, fitness) + for word in wordlist] + # Gotcha: the helper function here needs to be defined at the top level + # (limitation of Pool.starmap) + breaks = pool.starmap(vigenere_keyword_break_worker, helper_args, + chunksize) + return max(breaks, key=lambda k: k[1]) +vigenere_keyword_break = vigenere_keyword_break_mp + +def vigenere_keyword_break_worker(message, keyword, fitness): + plaintext = vigenere_decipher(message, keyword) + fit = fitness(plaintext) + logger.debug('Vigenere keyword break attempt using key {0} gives fit of ' + '{1} and decrypt starting: {2}'.format(keyword, + fit, sanitise(plaintext)[:50])) + return keyword, fit + + +def vigenere_frequency_break(message, max_key_length=20, fitness=Pletters): + """Breaks a Vigenere cipher with frequency analysis + + >>> vigenere_frequency_break(vigenere_encipher(sanitise("It is time to " \ + "run. She is ready and so am I. I stole Daniel's pocketbook this " \ + "afternoon when he left his jacket hanging on the easel in the " \ + "attic. I jump every time I hear a footstep on the stairs, " \ + "certain that the theft has been discovered and that I will " \ + "be caught. The SS officer visits less often now that he is " \ + "sure"), 'florence')) # doctest: +ELLIPSIS + ('florence', -307.5473096791...) + """ + def worker(message, key_length, fitness): + splits = every_nth(sanitised_message, key_length) + key = cat([chr(caesar_break(s)[0] + ord('a')) for s in splits]) + plaintext = vigenere_decipher(message, key) + fit = fitness(plaintext) + return key, fit + sanitised_message = sanitise(message) + results = starmap(worker, [(sanitised_message, i, fitness) + for i in range(1, max_key_length+1)]) + return max(results, key=lambda k: k[1]) + + +def beaufort_frequency_break(message, max_key_length=20, fitness=Pletters): + """Breaks a Beaufort cipher with frequency analysis + + >>> beaufort_frequency_break(beaufort_encipher(sanitise("It is time to " \ + "run. She is ready and so am I. I stole Daniel's pocketbook this " \ + "afternoon when he left his jacket hanging on the easel in the " \ + "attic. I jump every time I hear a footstep on the stairs, " \ + "certain that the theft has been discovered and that I will " \ + "be caught. The SS officer visits less often now " \ + "that he is sure"), 'florence')) # doctest: +ELLIPSIS + ('florence', -307.5473096791...) + """ + def worker(message, key_length, fitness): + splits = every_nth(sanitised_message, key_length) + key = cat([chr(-caesar_break(s)[0] % 26 + ord('a')) + for s in splits]) + plaintext = beaufort_decipher(message, key) + fit = fitness(plaintext) + return key, fit + sanitised_message = sanitise(message) + results = starmap(worker, [(sanitised_message, i, fitness) + for i in range(1, max_key_length+1)]) + return max(results, key=lambda k: k[1]) + + def column_transposition_break_mp(message, translist=transpositions, fitness=Pbigrams, chunksize=500): """Breaks a column transposition cipher using a dictionary and @@ -278,7 +359,7 @@ def column_transposition_break_mp(message, translist=transpositions, with Pool() as pool: helper_args = [(message, trans, fillcolumnwise, emptycolumnwise, fitness) - for trans in translist.keys() + for trans in translist for fillcolumnwise in [True, False] for emptycolumnwise in [True, False]] # Gotcha: the helper function here needs to be defined at the top level @@ -339,83 +420,144 @@ def scytale_break_mp(message, max_key_length=20, scytale_break = scytale_break_mp -def vigenere_keyword_break_mp(message, wordlist=keywords, fitness=Pletters, - chunksize=500): - """Breaks a vigenere cipher using a dictionary and frequency analysis. - - >>> vigenere_keyword_break_mp(vigenere_encipher(sanitise('this is a test ' \ - 'message for the vigenere decipherment'), 'cat'), \ - wordlist=['cat', 'elephant', 'kangaroo']) # doctest: +ELLIPSIS - ('cat', -52.947271216...) - """ - with Pool() as pool: - helper_args = [(message, word, fitness) - for word in wordlist] - # Gotcha: the helper function here needs to be defined at the top level - # (limitation of Pool.starmap) - breaks = pool.starmap(vigenere_keyword_break_worker, helper_args, - chunksize) - return max(breaks, key=lambda k: k[1]) -vigenere_keyword_break = vigenere_keyword_break_mp - -def vigenere_keyword_break_worker(message, keyword, fitness): - plaintext = vigenere_decipher(message, keyword) - fit = fitness(plaintext) - logger.debug('Vigenere keyword break attempt using key {0} gives fit of ' - '{1} and decrypt starting: {2}'.format(keyword, - fit, sanitise(plaintext)[:50])) - return keyword, fit - +def railfence_break(message, max_key_length=20, + fitness=Pletters, chunksize=500): + """Breaks a hill cipher using a matrix of given rank and letter frequencies - -def vigenere_frequency_break(message, max_key_length=20, fitness=Pletters): - """Breaks a Vigenere cipher with frequency analysis - - >>> vigenere_frequency_break(vigenere_encipher(sanitise("It is time to " \ - "run. She is ready and so am I. I stole Daniel's pocketbook this " \ - "afternoon when he left his jacket hanging on the easel in the " \ - "attic. I jump every time I hear a footstep on the stairs, " \ - "certain that the theft has been discovered and that I will " \ - "be caught. The SS officer visits less often now that he is " \ - "sure"), 'florence')) # doctest: +ELLIPSIS - ('florence', -307.5473096791...) + """ - def worker(message, key_length, fitness): - splits = every_nth(sanitised_message, key_length) - key = ''.join([chr(caesar_break(s)[0] + ord('a')) for s in splits]) - plaintext = vigenere_decipher(message, key) - fit = fitness(plaintext) - return key, fit + sanitised_message = sanitise(message) results = starmap(worker, [(sanitised_message, i, fitness) - for i in range(1, max_key_length+1)]) + for i in range(2, max_key_length+1)]) return max(results, key=lambda k: k[1]) -def beaufort_frequency_break(message, max_key_length=20, fitness=Pletters): - """Breaks a Beaufort cipher with frequency analysis +def railfence_break(message, max_key_length=20, + fitness=Pbigrams, chunksize=500): + """Breaks a railfence cipher using a range of lengths and + n-gram frequency analysis - >>> beaufort_frequency_break(beaufort_encipher(sanitise("It is time to " \ - "run. She is ready and so am I. I stole Daniel's pocketbook this " \ - "afternoon when he left his jacket hanging on the easel in the " \ - "attic. I jump every time I hear a footstep on the stairs, " \ - "certain that the theft has been discovered and that I will " \ - "be caught. The SS officer visits less often now " \ - "that he is sure"), 'florence')) # doctest: +ELLIPSIS - ('florence', -307.5473096791...) + >>> railfence_break(railfence_encipher(sanitise( \ + "It is a truth universally acknowledged, that a single man in \ + possession of a good fortune, must be in want of a wife. However \ + little known the feelings or views of such a man may be on his \ + first entering a neighbourhood, this truth is so well fixed in \ + the minds of the surrounding families, that he is considered the \ + rightful property of some one or other of their daughters."), \ + 7)) # doctest: +ELLIPSIS + (7, -709.46467226...) + >>> railfence_break(railfence_encipher(sanitise( \ + "It is a truth universally acknowledged, that a single man in \ + possession of a good fortune, must be in want of a wife. However \ + little known the feelings or views of such a man may be on his \ + first entering a neighbourhood, this truth is so well fixed in \ + the minds of the surrounding families, that he is considered the \ + rightful property of some one or other of their daughters."), \ + 7), \ + fitness=Ptrigrams) # doctest: +ELLIPSIS + (7, -997.0129085...) """ - def worker(message, key_length, fitness): - splits = every_nth(sanitised_message, key_length) - key = ''.join([chr(-caesar_break(s)[0] % 26 + ord('a')) - for s in splits]) - plaintext = beaufort_decipher(message, key) + def worker(message, height, fitness): + plaintext = railfence_decipher(message, height) fit = fitness(plaintext) - return key, fit + return height, fit + sanitised_message = sanitise(message) results = starmap(worker, [(sanitised_message, i, fitness) - for i in range(1, max_key_length+1)]) + for i in range(2, max_key_length+1)]) return max(results, key=lambda k: k[1]) +def amsco_break(message, translist=transpositions, patterns = [(1, 2), (2, 1)], + fillstyles = [AmscoFillStyle.continuous, + AmscoFillStyle.same_each_row, + AmscoFillStyle.reverse_each_row], + fitness=Pbigrams, + chunksize=500): + """Breaks an AMSCO transposition cipher using a dictionary and + n-gram frequency analysis + + >>> amsco_break(amsco_transposition_encipher(sanitise( \ + "It is a truth universally acknowledged, that a single man in \ + possession of a good fortune, must be in want of a wife. However \ + little known the feelings or views of such a man may be on his \ + first entering a neighbourhood, this truth is so well fixed in \ + the minds of the surrounding families, that he is considered the \ + rightful property of some one or other of their daughters."), \ + 'encipher'), \ + translist={(2, 0, 5, 3, 1, 4, 6): ['encipher'], \ + (5, 0, 6, 1, 3, 4, 2): ['fourteen'], \ + (6, 1, 0, 4, 5, 3, 2): ['keyword']}, \ + patterns=[(1, 2)]) # doctest: +ELLIPSIS + (((2, 0, 5, 3, 1, 4, 6), (1, 2), ), -709.4646722...) + >>> amsco_break(amsco_transposition_encipher(sanitise( \ + "It is a truth universally acknowledged, that a single man in \ + possession of a good fortune, must be in want of a wife. However \ + little known the feelings or views of such a man may be on his \ + first entering a neighbourhood, this truth is so well fixed in \ + the minds of the surrounding families, that he is considered the \ + rightful property of some one or other of their daughters."), \ + 'encipher', fillpattern=(2, 1)), \ + translist={(2, 0, 5, 3, 1, 4, 6): ['encipher'], \ + (5, 0, 6, 1, 3, 4, 2): ['fourteen'], \ + (6, 1, 0, 4, 5, 3, 2): ['keyword']}, \ + patterns=[(1, 2), (2, 1)], fitness=Ptrigrams) # doctest: +ELLIPSIS + (((2, 0, 5, 3, 1, 4, 6), (2, 1), ), -997.0129085...) + """ + with Pool() as pool: + helper_args = [(message, trans, pattern, fillstyle, fitness) + for trans in translist + for pattern in patterns + for fillstyle in fillstyles] + # Gotcha: the helper function here needs to be defined at the top level + # (limitation of Pool.starmap) + breaks = pool.starmap(amsco_break_worker, helper_args, chunksize) + return max(breaks, key=lambda k: k[1]) + +def amsco_break_worker(message, transposition, + pattern, fillstyle, fitness): + plaintext = amsco_transposition_decipher(message, transposition, + fillpattern=pattern, fillstyle=fillstyle) + fit = fitness(sanitise(plaintext)) + logger.debug('AMSCO transposition break attempt using key {0} and pattern' + '{1} ({2}) gives fit of {3} and decrypt starting: ' + '{4}'.format( + transposition, pattern, fillstyle, fit, + sanitise(plaintext)[:50])) + return (transposition, pattern, fillstyle), fit + + +def hill_break(message, matrix_size=2, fitness=Pletters, + number_of_solutions=1, chunksize=500): + + all_matrices = [np.matrix(list(m)) + for m in itertools.product([list(r) + for r in itertools.product(range(26), repeat=matrix_size)], + repeat=matrix_size)] + valid_matrices = [m for m, d in + zip(all_matrices, (int(round(linalg.det(m))) for m in all_matrices)) + if d != 0 + if d % 2 != 0 + if d % 13 != 0 ] + with Pool() as pool: + helper_args = [(message, matrix, fitness) + for matrix in valid_matrices] + # Gotcha: the helper function here needs to be defined at the top level + # (limitation of Pool.starmap) + breaks = pool.starmap(hill_break_worker, helper_args, chunksize) + if number_of_solutions == 1: + return max(breaks, key=lambda k: k[1]) + else: + return sorted(breaks, key=lambda k: k[1], reverse=True)[:number_of_solutions] + +def hill_break_worker(message, matrix, fitness): + plaintext = hill_decipher(matrix, message) + fit = fitness(plaintext) + logger.debug('Hill cipher break attempt using key {0} gives fit of ' + '{1} and decrypt starting: {2}'.format(matrix, + fit, sanitise(plaintext)[:50])) + return matrix, fit + def pocket_enigma_break_by_crib(message, wheel_spec, crib, crib_position): """Break a pocket enigma using a crib (some plaintext that's expected to @@ -446,13 +588,13 @@ def pocket_enigma_break_by_crib(message, wheel_spec, crib, crib_position): def plot_frequency_histogram(freqs, sort_key=None): - x = range(len(freqs.keys())) - y = [freqs[l] for l in sorted(freqs.keys(), key=sort_key)] + x = range(len(freqs)) + y = [freqs[l] for l in sorted(freqs, key=sort_key)] f = plt.figure() ax = f.add_axes([0.1, 0.1, 0.9, 0.9]) ax.bar(x, y, align='center') ax.set_xticks(x) - ax.set_xticklabels(sorted(freqs.keys(), key=sort_key)) + ax.set_xticklabels(sorted(freqs, key=sort_key)) f.show()