Worked on Enigma, mainly changing how the notch positions are handled

[cipher-training.git] / cipherbreak.py

diff --git a/cipherbreak.py b/cipherbreak.py
index 1a589c2ef1c5d1720f44d82651eda6b6a6cfe434..f1ab58e50a4f293368ef9d5382aef9dabf590420 100644 (file)
--- a/cipherbreak.py
+++ b/cipherbreak.py
@@ -13,11 +13,22 @@ from multiprocessing import Pool
 
 import matplotlib.pyplot as plt
 
-logger = logging.getLogger(__name__)
-logger.addHandler(logging.FileHandler('cipher.log'))
+# logging.basicConfig(filename="cipher.log", level=logging.INFO)
+# logger = logging.getLogger(__name__)
+
+logger = logging.getLogger('cipherbreak')
 logger.setLevel(logging.WARNING)
-#logger.setLevel(logging.INFO)
-#logger.setLevel(logging.DEBUG)
+# logger.setLevel(logging.INFO)
+# logger.setLevel(logging.DEBUG)
+
+# create the logging file handler
+fh = logging.FileHandler("cipher.log")
+formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
+fh.setFormatter(formatter)
+
+# add handler to logger object
+logger.addHandler(fh)
+
 
 from cipher import *
 from language_models import *
@@ -204,7 +215,7 @@ def monoalphabetic_break_hillclimbing(message, max_iterations=10000000,
     if not alphabet:
         alphabet = list(string.ascii_lowercase)
         random.shuffle(alphabet)
-        alphabet = ''.join(alphabet)
+        alphabet = cat(alphabet)
     return monoalphabetic_break_hillclimbing_worker(ciphertext, alphabet,
                                                     max_iterations, fitness)
 
@@ -218,7 +229,7 @@ def monoalphabetic_break_hillclimbing_mp(message, workers=10,
         else:
             this_alphabet = list(string.ascii_lowercase)
             random.shuffle(this_alphabet)
-            this_alphabet = ''.join(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,
@@ -290,7 +301,7 @@ def vigenere_frequency_break(message, max_key_length=20, fitness=Pletters):
     """
     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])
+        key = cat([chr(caesar_break(s)[0] + ord('a')) for s in splits])
         plaintext = vigenere_decipher(message, key)
         fit = fitness(plaintext)
         return key, fit
@@ -314,7 +325,7 @@ def beaufort_frequency_break(message, max_key_length=20, fitness=Pletters):
     """
     def worker(message, key_length, fitness):
         splits = every_nth(sanitised_message, key_length)
-        key = ''.join([chr(-caesar_break(s)[0] % 26 + ord('a'))
+        key = cat([chr(-caesar_break(s)[0] % 26 + ord('a'))
                        for s in splits])
         plaintext = beaufort_decipher(message, key)
         fit = fitness(plaintext)
@@ -558,6 +569,43 @@ def hill_break_worker(message, matrix, fitness):
                      fit, sanitise(plaintext)[:50]))
     return matrix, fit
 
+def bifid_break_mp(message, wordlist=keywords, fitness=Pletters, max_period=10,
+                     number_of_solutions=1, chunksize=500):
+    """Breaks a keyword substitution cipher using a dictionary and
+    frequency analysis
+
+    >>> bifid_break_mp(bifid_encipher('this is a test message for the ' \
+          'keyword decipherment', 'elephant', wrap_alphabet=KeywordWrapAlphabet.from_last), \
+          wordlist=['cat', 'elephant', 'kangaroo']) # doctest: +ELLIPSIS
+    (('elephant', <KeywordWrapAlphabet.from_last: 2>, 0), -52.834575011...)
+    >>> bifid_break_mp(bifid_encipher('this is a test message for the ' \
+          'keyword decipherment', 'elephant', wrap_alphabet=KeywordWrapAlphabet.from_last), \
+          wordlist=['cat', 'elephant', 'kangaroo'], \
+          number_of_solutions=2) # doctest: +ELLIPSIS, +NORMALIZE_WHITESPACE
+    [(('elephant', <KeywordWrapAlphabet.from_last: 2>, 0), -52.834575011...), 
+    (('elephant', <KeywordWrapAlphabet.from_largest: 3>, 0), -52.834575011...)]
+    """
+    with Pool() as pool:
+        helper_args = [(message, word, wrap, period, fitness)
+                       for word in wordlist
+                       for wrap in KeywordWrapAlphabet
+                       for period in range(max_period+1)]
+        # Gotcha: the helper function here needs to be defined at the top level
+        #   (limitation of Pool.starmap)
+        breaks = pool.starmap(bifid_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 bifid_break_worker(message, keyword, wrap_alphabet, period, fitness):
+    plaintext = bifid_decipher(message, keyword, wrap_alphabet, period=period)
+    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, period), 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