From: Neil Smith Date: Wed, 5 Mar 2014 21:37:40 +0000 (-0500) Subject: Changed the break routines, no to make them all work again X-Git-Url: https://git.njae.me.uk/?a=commitdiff_plain;h=27abb216333fda20dc857a8a501fbee4a4a962f4;p=cipher-tools.git Changed the break routines, no to make them all work again --- diff --git a/cipherbreak.py b/cipherbreak.py index 03165e5..06a66eb 100644 --- a/cipherbreak.py +++ b/cipherbreak.py @@ -86,10 +86,7 @@ def caesar_break(message, fitness=Pletters): caesar_decipher(sanitised_message, best_shift)[:50])) return best_shift, best_fit -def affine_break(message, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise): +def affine_break(message, fitness=Pletters): """Breaks an affine cipher using frequency analysis >>> affine_break('lmyfu bkuusd dyfaxw claol psfaom jfasd snsfg jfaoe ls ' \ @@ -104,19 +101,18 @@ def affine_break(message, best_multiplier = 0 best_adder = 0 best_one_based = True - best_fit = float("inf") + best_fit = float("-inf") for one_based in [True, False]: for multiplier in range(1, 26, 2): for adder in range(26): plaintext = affine_decipher(sanitised_message, multiplier, adder, one_based) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + fit = fitness(plaintext) logger.debug('Affine break attempt using key {0}x+{1} ({2}) ' 'gives fit of {3} and decrypt starting: {4}'. format(multiplier, adder, one_based, fit, plaintext[:50])) - if fit < best_fit: + if fit > best_fit: best_fit = fit best_multiplier = multiplier best_adder = adder @@ -128,11 +124,7 @@ def affine_break(message, best_adder, best_one_based)[:50])) return (best_multiplier, best_adder, best_one_based), best_fit -def keyword_break(message, - wordlist=keywords, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise): +def keyword_break(message, wordlist=keywords, fitness=Pletters): """Breaks a keyword substitution cipher using a dictionary and frequency analysis @@ -143,17 +135,16 @@ def keyword_break(message, """ best_keyword = '' best_wrap_alphabet = True - best_fit = float("inf") + best_fit = float("-inf") for wrap_alphabet in range(3): for keyword in wordlist: plaintext = keyword_decipher(message, keyword, wrap_alphabet) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + fit = fitness(plaintext) logger.debug('Keyword break attempt using key {0} (wrap={1}) ' 'gives fit of {2} and decrypt starting: {3}'.format( keyword, wrap_alphabet, fit, sanitise(plaintext)[:50])) - if fit < best_fit: + if fit > best_fit: best_fit = fit best_keyword = keyword best_wrap_alphabet = wrap_alphabet @@ -164,12 +155,7 @@ def keyword_break(message, best_wrap_alphabet))[:50])) return (best_keyword, best_wrap_alphabet), best_fit -def keyword_break_mp(message, - wordlist=keywords, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise, - chunksize=500): +def keyword_break_mp(message, wordlist=keywords, fitness=Pletters, chunksize=500): """Breaks a keyword substitution cipher using a dictionary and frequency analysis @@ -179,28 +165,22 @@ def keyword_break_mp(message, (('elephant', 1), 0.106645344886...) """ with Pool() as pool: - helper_args = [(message, word, wrap, metric, target_counts, - message_frequency_scaling) + helper_args = [(message, word, wrap, fitness) for word in wordlist for wrap in range(3)] # Gotcha: the helper function here needs to be defined at the top level # (limitation of Pool.starmap) breaks = pool.starmap(keyword_break_worker, helper_args, chunksize) - return min(breaks, key=lambda k: k[1]) + return max(breaks, key=lambda k: k[1]) -def keyword_break_worker(message, keyword, wrap_alphabet, metric, target_counts, - message_frequency_scaling): +def keyword_break_worker(message, keyword, wrap_alphabet, fitness): plaintext = keyword_decipher(message, keyword, wrap_alphabet) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + fit = fitness(plaintext) logger.debug('Keyword break attempt using key {0} (wrap={1}) gives fit of ' '{2} and decrypt starting: {3}'.format(keyword, wrap_alphabet, fit, sanitise(plaintext)[:50])) return (keyword, wrap_alphabet), fit -def scytale_break(message, - metric=norms.euclidean_distance, - target_counts=normalised_english_bigram_counts, - message_frequency_scaling=norms.normalise): +def scytale_break(message, fitness=Pbigrams): """Breaks a Scytale cipher >>> scytale_break('tfeulchtrtteehwahsdehneoifeayfsondmwpltmaoalhikotoere' \ @@ -209,18 +189,15 @@ def scytale_break(message, (6, 0.092599933059...) """ best_key = 0 - best_fit = float("inf") - ngram_length = len(next(iter(target_counts.keys()))) + best_fit = float("-inf") for key in range(1, 20): if len(message) % key == 0: plaintext = scytale_decipher(message, key) - counts = message_frequency_scaling(frequencies( - ngrams(sanitise(plaintext), ngram_length))) - fit = metric(target_counts, counts) + fit = fitness(sanitise(plaintext)) logger.debug('Scytale break attempt using key {0} gives fit of ' '{1} and decrypt starting: {2}'.format(key, fit, sanitise(plaintext)[:50])) - if fit < best_fit: + if fit > best_fit: best_fit = fit best_key = key logger.info('Scytale break best fit with key {0} gives fit of {1} and ' @@ -229,12 +206,8 @@ def scytale_break(message, return best_key, best_fit -def column_transposition_break_mp(message, - translist=transpositions, - metric=norms.euclidean_distance, - target_counts=normalised_english_bigram_counts, - message_frequency_scaling=norms.normalise, - chunksize=500): +def column_transposition_break_mp(message, translist=transpositions, + fitness=Pbigrams, chunksize=500): """Breaks a column transposition cipher using a dictionary and n-gram frequency analysis """ @@ -264,23 +237,21 @@ def column_transposition_break_mp(message, # target_counts=normalised_english_trigram_counts) # doctest: +ELLIPSIS # (((2, 0, 5, 3, 1, 4, 6), False), 0.0592259560...) # """ - ngram_length = len(next(iter(target_counts.keys()))) with Pool() as pool: - helper_args = [(message, trans, columnwise, metric, target_counts, ngram_length, - message_frequency_scaling) - for trans in translist.keys() for columnwise in [True, False]] + helper_args = [(message, trans, columnwise, fitness) + for trans in translist.keys() + for columnwise in [True, False]] # Gotcha: the helper function here needs to be defined at the top level # (limitation of Pool.starmap) - breaks = pool.starmap(column_transposition_break_worker, helper_args, chunksize) - return min(breaks, key=lambda k: k[1]) + breaks = pool.starmap(column_transposition_break_worker, + helper_args, chunksize) + return max(breaks, key=lambda k: k[1]) column_transposition_break = column_transposition_break_mp -def column_transposition_break_worker(message, transposition, columnwise, metric, target_counts, - ngram_length, message_frequency_scaling): +def column_transposition_break_worker(message, transposition, columnwise, + fitness): plaintext = column_transposition_decipher(message, transposition, columnwise=columnwise) - counts = message_frequency_scaling(frequencies( - ngrams(sanitise(plaintext), ngram_length))) - fit = metric(target_counts, counts) + fit = fitness(sanitise(plaintext)) logger.debug('Column transposition break attempt using key {0} ' 'gives fit of {1} and decrypt starting: {2}'.format( transposition, fit, @@ -288,36 +259,28 @@ def column_transposition_break_worker(message, transposition, columnwise, metric return (transposition, columnwise), fit -def transposition_break_exhaustive(message): +def transposition_break_exhaustive(message, fitness=Pbigrams): best_transposition = '' - best_pw = -float('inf') + best_pw = float('-inf') for keylength in range(1, 21): if len(message) % keylength == 0: for transposition in permutations(range(keylength)): for columnwise in [True, False]: plaintext = column_transposition_decipher(message, transposition, columnwise=columnwise) - # pw = Pwords(segment(plaintext)) - pw = sum([log10(bigram_likelihood(b, - normalised_english_bigram_counts, - normalised_english_counts)) - for b in ngrams(plaintext, 2)]) + fit=fitness(plaintext) logger.debug('Column transposition break attempt using key {0} {1} ' 'gives fit of {2} and decrypt starting: {3}'.format( transposition, columnwise, pw, sanitise(plaintext)[:50])) - if pw > best_pw: + if fit > best_fit: best_transposition = transposition best_columnwise = columnwise - best_pw = pw + best_fit = fit return (best_transposition, best_columnwise), best_pw -def vigenere_keyword_break(message, - wordlist=keywords, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise): +def vigenere_keyword_break(message, wordlist=keywords, fitness=Pletters): """Breaks a vigenere cipher using a dictionary and frequency analysis @@ -327,16 +290,15 @@ def vigenere_keyword_break(message, ('cat', 0.15965224935...) """ best_keyword = '' - best_fit = float("inf") + best_fit = float("-inf") for keyword in wordlist: plaintext = vigenere_decipher(message, keyword) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + fit = fitness(plaintext) logger.debug('Vigenere break attempt using key {0} ' 'gives fit of {1} and decrypt starting: {2}'.format( keyword, fit, sanitise(plaintext)[:50])) - if fit < best_fit: + if fit > best_fit: best_fit = fit best_keyword = keyword logger.info('Vigenere break best fit with key {0} gives fit ' @@ -345,11 +307,7 @@ def vigenere_keyword_break(message, vigenere_decipher(message, best_keyword))[:50])) return best_keyword, best_fit -def vigenere_keyword_break_mp(message, - wordlist=keywords, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise, +def vigenere_keyword_break_mp(message, wordlist=keywords, fitness=Pletters, chunksize=500): """Breaks a vigenere cipher using a dictionary and frequency analysis @@ -360,19 +318,16 @@ def vigenere_keyword_break_mp(message, ('cat', 0.159652249358...) """ with Pool() as pool: - helper_args = [(message, word, metric, target_counts, - message_frequency_scaling) + 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 min(breaks, key=lambda k: k[1]) -def vigenere_keyword_break_worker(message, keyword, metric, target_counts, - message_frequency_scaling): +def vigenere_keyword_break_worker(message, keyword, fitness): plaintext = vigenere_decipher(message, keyword) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + 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])) @@ -380,30 +335,29 @@ def vigenere_keyword_break_worker(message, keyword, metric, target_counts, -def vigenere_frequency_break(message, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise): +def vigenere_frequency_break(message, 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."), 'florence')) # doctest: +ELLIPSIS + "attic. I jump every time I hear a footstep on the stairs, " \ + "certain that the theft has been discovered and that I will " \ + "and that I will be caught. The SS officer visits less often now " \ + "that he is sure"), 'florence')) # doctest: +ELLIPSIS ('florence', 0.077657073...) """ - best_fit = float("inf") + best_fit = float("-inf") best_key = '' sanitised_message = sanitise(message) for trial_length in range(1, 20): splits = every_nth(sanitised_message, trial_length) key = ''.join([chr(caesar_break(s)[0] + ord('a')) for s in splits]) plaintext = vigenere_decipher(sanitised_message, key) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + fit = fitness(plaintext) logger.debug('Vigenere key length of {0} ({1}) gives fit of {2}'. format(trial_length, key, fit)) - if fit < best_fit: + if fit > best_fit: best_fit = fit best_key = key logger.info('Vigenere break best fit with key {0} gives fit ' @@ -412,30 +366,29 @@ def vigenere_frequency_break(message, vigenere_decipher(message, best_key))[:50])) return best_key, best_fit -def beaufort_frequency_break(message, - metric=norms.euclidean_distance, - target_counts=normalised_english_counts, - message_frequency_scaling=norms.normalise): +def beaufort_frequency_break(message, 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."), 'florence')) # doctest: +ELLIPSIS + "attic. I jump every time I hear a footstep on the stairs, " \ + "certain that the theft has been discovered and that I will " \ + "and that I will be caught. The SS officer visits less often now " \ + "that he is sure"), 'florence')) # doctest: +ELLIPSIS ('florence', 0.077657073...) """ - best_fit = float("inf") + best_fit = float("-inf") best_key = '' sanitised_message = sanitise(message) for trial_length in range(1, 20): splits = every_nth(sanitised_message, trial_length) key = ''.join([chr(caesar_break(s)[0] + ord('a')) for s in splits]) plaintext = beaufort_decipher(sanitised_message, key) - counts = message_frequency_scaling(frequencies(plaintext)) - fit = metric(target_counts, counts) + fit = fitness(plaintext) logger.debug('Beaufort key length of {0} ({1}) gives fit of {2}'. format(trial_length, key, fit)) - if fit < best_fit: + if fit > best_fit: best_fit = fit best_key = key logger.info('Beaufort break best fit with key {0} gives fit ' diff --git a/language_models.py b/language_models.py index d457386..1b90ac2 100644 --- a/language_models.py +++ b/language_models.py @@ -88,6 +88,19 @@ def random_english_letter(): return weighted_choice(normalised_english_counts) +def ngrams(text, n): + """Returns all n-grams of a text + + >>> ngrams(sanitise('the quick brown fox'), 2) # doctest: +NORMALIZE_WHITESPACE + ['th', 'he', 'eq', 'qu', 'ui', 'ic', 'ck', 'kb', 'br', 'ro', 'ow', 'wn', + 'nf', 'fo', 'ox'] + >>> ngrams(sanitise('the quick brown fox'), 4) # doctest: +NORMALIZE_WHITESPACE + ['theq', 'hequ', 'equi', 'quic', 'uick', 'ickb', 'ckbr', 'kbro', 'brow', + 'rown', 'ownf', 'wnfo', 'nfox'] + """ + return [text[i:i+n] for i in range(len(text)-n+1)] + + class Pdist(dict): """A probability distribution estimated from counts in datafile. Values are stored and returned as log probabilities. @@ -108,6 +121,7 @@ def log_probability_of_unknown_word(key, N): Pw = Pdist(datafile('count_1w.txt'), log_probability_of_unknown_word) Pl = Pdist(datafile('count_1l.txt'), lambda _k, _N: 0) +Pl2 = Pdist(datafile('count_2l.txt'), lambda _k, _N: 0) def Pwords(words): """The Naive Bayes log probability of a sequence of words. @@ -119,6 +133,11 @@ def Pletters(letters): """ return sum(Pl[l.lower()] for l in letters) +def Pbigrams(letters): + """The Naive Bayes log probability of the bigrams formed from a sequence + of letters. + """ + return sum(P2l[p] for p in ngrams(letters, 2)) def cosine_distance_score(text):