In [1]:
import numpy as np
import math
import matplotlib.pyplot as plt
%matplotlib inline
import random
from numpy.random import rand
from copy import copy
from __future__ import division
def read_text_words(filename, wordsnumber):
with open(filename) as f:
X = f.readlines()
X = X[:wordsnumber]
X = ''.join(X)
X = X.replace('\n', '')
return X
def read_text_whole(filename):
with open(filename) as f:
X = f.read()
X = X.replace('\n', '')
return X
def chop_text_to_size(text, size):
return text[:1024*1024*size]
def read_text_filesize(filename, size):
with open(filename) as f:
X = f.read(1024*1024*size)
X = X.replace('\n', '')
return X
def get_unicount(text):
length = len(text)
counts = np.zeros(26)
for i in xrange(length):
c = ord(text[i])
counts[c-97]+=1
#97-122
return counts
def get_bigram_stats_dic(text):
length = len(text)
dic = {}
for i in 'abcdefghijklmnopqrstuvwxyz':
for j in 'abcdefghijklmnopqrstuvwxyz':
dic[(i,j)]=0
for i in xrange(length-1):
if (text[i], text[i+1]) in dic:
dic[(text[i], text[i+1])] += 1
for k,v in dic.items():
dic[k] = v/(counts[ord(k[0])-97])
return dic
def quality(decrypted, original):
l = len(decrypted)
zipped = zip(decrypted, original)
return sum(1.0 for x,y in zipped if x == y)/l
def crypt(text):
p = range(26)
random.shuffle(p)
output=''
for ch in text:
try:
x = ord(ch) - ord('a')
output+=(chr(p[x] + ord('a')))
except:
pass
return output, p
def get_desiredPDF_bigram(permutation):
logp = 0
for i in xrange(len(encrypted)-1):
pr = stats[chr(permutation[ord(encrypted[i])-97]+97),
chr(permutation[ord(encrypted[i+1])-97]+97)]
if pr>0:
logp += math.log(pr)
else:
logp += -9 #penalty for non existant pairs
return logp
def uniform( n ):
#initialize permutation with identical
permutation = [ i for i in xrange( n ) ]
#swap ith object with random onject from i to n - 1 enclusively
for i in xrange( n ):
j = random.randint( i, n - 1 )
permutation[ i ], permutation[ j ] = permutation[ j ], permutation[ i ]
return permutation
def applyedTranspostions( basePermutation ):
n = len( basePermutation )
permutation = copy( basePermutation )
#apply n random transpositions (including identical) to base permutation
# for i in xrange( n ):
k, l = random.randint( 0, n - 1 ), random.randint( 0, n - 1 )
permutation[ k ], permutation[ l ] = permutation[ l ], permutation[ k ]
return permutation
density maximization
In [2]:
def densityMaximization( desiredPDF, initValue, computableRVS, skipIterations = 200 ):
"""
This function return a generator, which generates random variables
from some space S by trying to maximize givven density.
The algorithm is a modification of Metropolis-Hastings.
It rejects all objects, which decrease density.
Args:
desiredPDF (func) : PDF of desired distribution p( T ), where T from S
initValue : an object from S to initialize the starting point
of iterative proccess
computableRVS (func) : a generator of random value from space S
with given parameter T, which is also from S
skipIterations (int) : number of iterations to skip
(skipping more iterations leads to better accuracy?
but greater time consuming)
Returns: generator, which produce some values from S,
where each next value has no less density, and their denisity
"""
random_variable = initValue
random_variableDensityValue = desiredPDF( random_variable )
"""
A state of MCMC
"""
#ignore first iterations to let the iterative proccess to enter
#the high density regions
for i in xrange( skipIterations ):
candidate = computableRVS( random_variable )
candidateDensityValue = desiredPDF( candidate )
"""
next candidate for sample, generated by computableRVS
"""
if random_variableDensityValue < candidateDensityValue:
random_variable = candidate
random_variableDensityValue = candidateDensityValue
#now when the procces is in high density regions,
#return acceptable candidates
while True:
candidate = computableRVS( random_variable )
candidateDensityValue = desiredPDF( candidate )
"""
next candidate for sample, generated by computableRVS
"""
if random_variableDensityValue < candidateDensityValue:
random_variable = candidate
random_variableDensityValue = candidateDensityValue
yield random_variable, random_variableDensityValue
decrypt
In [3]:
def decrypt(permutation, encrypted):
decrypted = []
for i in encrypted:
decrypted.append(chr(permutation[ord(i)-97]+97))
return ''.join(decrypted)
In [4]:
#TEST TEXT
fname = 'main/oliver_twist.txt'
original = read_text_words(fname, 5000)[3:]
#3 first symbols in oliver twist are unsupported by encryption
encrypted, p = crypt(original)
#TRAIN TEXT
train_text = read_text_whole('main/war_and_peace.txt')
counts = get_unicount(train_text)
stats = get_bigram_stats_dic(train_text)
# print stats
print p
bp = np.zeros(26, dtype=int)
for i in p:
bp[p[i]] = i
q = get_desiredPDF_bigram(bp)
print 'inverse to permutation used in encryption ', bp
print 'its density ', q
ra = uniform(26)
q = get_desiredPDF_bigram(ra)
print 'random permutation density ', q
In [10]:
import time
iterations = [250,500,1000,2000]
qualities = np.zeros((5, 4))
qualities[1, :] = qs;
init_p = uniform(26)
for i in xrange(1, 5):
for j, it in enumerate(iterations):
st = time.time()
computableGen = lambda t: applyedTranspostions(t)
dmgenerator = \
densityMaximization(get_desiredPDF_bigram, init_p, computableGen, it)
for k in xrange( 500 ):
x,y = dmgenerator.next()
et = time.time() - st
print 'cold iterations: ', it
print 'dm time: ', et
print 'best density among 500 last iterations: ', y
print 'corresponding permutation: ', x
decrypted = decrypt(x, encrypted)
qualities[i, j] = quality(decrypted, original)
print 'quality: ', qualities[i, j]
In [17]:
means = np.mean(qualities, 0)
stds = np.std(qualities, 0)
print(means)
print(stds)
plt.title('Dependence quolity on cold iterations')
plt.xlabel('iterations')
plt.ylabel('quolity')
plt.plot(iterations, means - stds, 'r:')
plt.plot(iterations, means + stds, 'r:')
plt.plot(iterations, means, 'b-')
plt.savefig('task-5-daniel.png')
In [ ]: