This is a simple GUI program that shows Markov Encryption at work. It is built to be interactive and has all needed code embedded within itself. This version of ME library is not very efficient and represents an early attempt at developing and easily testing the code. Certain limits are built into the program, and the code is not meant to be robust at this stage. This program is a proof-of-concept design meant to ensure that the work being done was viable for future use and that the encryption process could be carried out both ways, both in encoding plaintext and decoding ciphertext.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 | from tkinter import *
import traceback
from tkinter.scrolledtext import ScrolledText
CODEC = 'utf8'
# XXX This should import the "markov" module.
# XXX All changes made here should be copied.
################################################################################
class MarkovDemo:
def __init__(self, master):
self.prompt_size = Label(master, anchor=W, text='Encode Word Size')
self.prompt_size.pack(side=TOP, fill=X)
self.size_entry = Entry(master)
self.size_entry.insert(0, '8')
self.size_entry.pack(fill=X)
self.prompt_plain = Label(master, anchor=W, text='Plaintext Characters')
self.prompt_plain.pack(side=TOP, fill=X)
self.plain_entry = Entry(master)
self.plain_entry.insert(0, '""')
self.plain_entry.pack(fill=X)
self.showframe = Frame(master)
self.showframe.pack(fill=X, anchor=W)
self.showvar = StringVar(master)
self.showvar.set("encode")
self.showfirstradio = Radiobutton(self.showframe,
text="Encode Plaintext",
variable=self.showvar,
value="encode",
command=self.reevaluate)
self.showfirstradio.pack(side=LEFT)
self.showallradio = Radiobutton(self.showframe,
text="Decode Cyphertext",
variable=self.showvar,
value="decode",
command=self.reevaluate)
self.showallradio.pack(side=LEFT)
self.inputbox = ScrolledText(master, width=60, height=10, wrap=WORD)
self.inputbox.pack(fill=BOTH, expand=1)
self.dynamic_var = IntVar()
self.dynamic_box = Checkbutton(master, variable=self.dynamic_var,
text='Dynamic Evaluation',
offvalue=False, onvalue=True,
command=self.reevaluate)
self.dynamic_box.pack()
self.output = Label(master, anchor=W, text="This is your output:")
self.output.pack(fill=X)
self.outbox = ScrolledText(master, width=60, height=10, wrap=WORD)
self.outbox.pack(fill=BOTH, expand=1)
self.inputbox.bind('<Key>', self.reevaluate)
def select_all(event=None):
event.widget.tag_add(SEL, 1.0, 'end-1c')
event.widget.mark_set(INSERT, 1.0)
event.widget.see(INSERT)
return 'break'
self.inputbox.bind('<Control-Key-a>', select_all)
self.outbox.bind('<Control-Key-a>', select_all)
self.inputbox.bind('<Control-Key-/>', lambda event: 'break')
self.outbox.bind('<Control-Key-/>', lambda event: 'break')
self.outbox.config(state=DISABLED)
def reevaluate(self, event=None):
if event is not None:
if event.char == '':
return
if self.dynamic_var.get():
text = self.inputbox.get(1.0, END)[:-1]
if len(text) < 10:
return
text = text.replace('\n \n', '\n\n')
mode = self.showvar.get()
assert mode in ('decode', 'encode'), 'Bad mode!'
if mode == 'encode':
# Encode Plaintext
try:
# Evaluate the plaintext characters
plain = self.plain_entry.get()
if plain:
PC = eval(self.plain_entry.get())
else:
PC = ''
self.plain_entry.delete(0, END)
self.plain_entry.insert(0, '""')
# Evaluate the word size
size = self.size_entry.get()
if size:
XD = int(size)
while grid_size(text, XD, PC) > 1 << 20:
XD -= 1
else:
XD = 0
grid = 0
while grid <= 1 << 20:
grid = grid_size(text, XD, PC)
XD += 1
XD -= 1
# Correct the size and encode
self.size_entry.delete(0, END)
self.size_entry.insert(0, str(XD))
cyphertext, key, prime = encrypt_str(text, XD, PC)
except:
traceback.print_exc()
else:
buffer = ''
for block in key:
buffer += repr(block)[2:-1] + '\n'
buffer += repr(prime)[2:-1] + '\n\n' + cyphertext
self.outbox.config(state=NORMAL)
self.outbox.delete(1.0, END)
self.outbox.insert(END, buffer)
self.outbox.config(state=DISABLED)
else:
# Decode Cyphertext
try:
header, cypher = text.split('\n\n', 1)
lines = header.split('\n')
for index, item in enumerate(lines):
try:
lines[index] = eval('b"' + item + '"')
except:
lines[index] = eval("b'" + item + "'")
plain = decrypt_str(cypher, tuple(lines[:-1]), lines[-1])
except:
traceback.print_exc()
else:
self.outbox.config(state=NORMAL)
self.outbox.delete(1.0, END)
self.outbox.insert(END, plain)
self.outbox.config(state=DISABLED)
else:
text = self.inputbox.get(1.0, END)[:-1]
text = text.replace('\n \n', '\n\n')
mode = self.showvar.get()
assert mode in ('decode', 'encode'), 'Bad mode!'
if mode == 'encode':
try:
XD = int(self.size_entry.get())
PC = eval(self.plain_entry.get())
size = grid_size(text, XD, PC)
assert size
except:
pass
else:
buffer = 'Grid size will be:\n' + convert(size)
self.outbox.config(state=NORMAL)
self.outbox.delete(1.0, END)
self.outbox.insert(END, buffer)
self.outbox.config(state=DISABLED)
################################################################################
import random
CRYPT = random.SystemRandom()
################################################################################
# This section includes functions that
# can test the required key and bootstrap.
# sudoko_key
# - should be a proper "markov" key
def _check_sudoku_key(sudoku_key):
# Ensure key is a tuple with more than one item.
assert isinstance(sudoku_key, tuple), '"sudoku_key" must be a tuple'
assert len(sudoku_key) > 1, '"sudoku_key" must have more than one item'
# Test first item.
item = sudoku_key[0]
assert isinstance(item, bytes), 'first item must be an instance of bytes'
assert len(item) > 1, 'first item must have more than one byte'
assert len(item) == len(set(item)), 'first item must have unique bytes'
# Test the rest of the key.
for obj in sudoku_key[1:]:
assert isinstance(obj, bytes), 'remaining items must be of bytes'
assert len(obj) == len(item), 'all items must have the same length'
assert len(obj) == len(set(obj)), \
'remaining items must have unique bytes'
assert len(set(item)) == len(set(item).union(set(obj))), \
'all items must have the same bytes'
# boot_strap
# - should be a proper "markov" bootstrap
# - we will call this a "primer"
# sudoko_key
# - should be a proper "markov" key
def _check_boot_strap(boot_strap, sudoku_key):
assert isinstance(boot_strap, bytes), '"boot_strap" must be a bytes object'
assert len(boot_strap) == len(sudoku_key) - 1, \
'"boot_strap" length must be one less than "sudoku_key" length'
item = sudoku_key[0]
assert len(set(item)) == len(set(item).union(set(boot_strap))), \
'"boot_strap" may only have bytes found in "sudoku_key"'
################################################################################
# This section includes functions capable
# of creating the required key and bootstrap.
# bytes_set should be any collection of bytes
# - it should be possible to create a set from them
# - these should be the bytes on which encryption will follow
# word_size
# - this will be the size of the "markov" chains program uses
# - this will be the number of dimensions the "grid" will have
# - one less character will make up bootstrap (or primer)
def make_sudoku_key(bytes_set, word_size):
key_set = set(bytes_set)
blocks = []
for block in range(word_size):
blocks.append(bytes(CRYPT.sample(key_set, len(key_set))))
return tuple(blocks)
# sudoko_key
# - should be a proper "markov" key
def make_boot_strap(sudoku_key):
block = sudoku_key[0]
return bytes(CRYPT.choice(block) for byte in range(len(sudoku_key) - 1))
################################################################################
# This section contains functions needed to
# create the multidimensional encryption grid.
# sudoko_key
# - should be a proper "markov" key
def make_grid(sudoku_key):
grid = expand_array(sudoku_key[0], sudoku_key[1])
for block in sudoku_key[2:]:
grid = expand_array(grid, block)
return grid
# grid
# - should be an X dimensional grid from make_grid
# block_size
# - comes from length of one block in a sudoku_key
def make_decode_grid(grid, block_size):
cache = []
for part in range(0, len(grid), block_size):
old = grid[part:part+block_size]
new = [None] * block_size
key = sorted(old)
for index, byte in enumerate(old):
new[key.index(byte)] = key[index]
cache.append(bytes(new))
return b''.join(cache)
# grid
# - should be an X dimensional grid from make_grid
# block
# - should be a block from a sudoku_key
# - should have same unique bytes as the expanding grid
def expand_array(grid, block):
cache = []
grid_size = len(grid)
block_size = len(block)
for byte in block:
index = grid.index(bytes([byte]))
for part in range(0, grid_size, block_size):
cache.append(grid[part+index:part+block_size])
cache.append(grid[part:part+index])
return b''.join(cache)
################################################################################
# The first three functions can be used to check an encryption
# grid. The eval_index function is used to evaluate a grid cell.
# grid
# - grid object to be checked
# - grid should come from the make_grid function
# - must have unique bytes along each axis
# block_size
# - comes from length of one block in a sudoku_key
# - this is the length of one edge along the grid
# - each axis is this many unit long exactly
# word_size
# - this is the number of blocks in a sudoku_key
# - this is the number of dimensions in a grid
# - this is the length needed to create a needed markon chain
def check_grid(grid, block_size, word_size):
build_index(grid, block_size, word_size, [])
# create an index to access the grid with
def build_index(grid, block_size, word_size, index):
for number in range(block_size):
index.append(number)
if len(index) == word_size:
check_cell(grid, block_size, word_size, index)
else:
build_index(grid, block_size, word_size, index)
index.pop()
# compares the contents of a cell along each grid axis
def check_cell(grid, block_size, word_size, index):
master = eval_index(grid, block_size, index)
for axis in range(word_size):
for value in range(block_size):
if index[axis] != value:
copy = list(index)
copy[axis] = value
slave = eval_index(grid, block_size, copy)
assert slave != master, 'Cell not unique along axis!'
# grid
# - grid object to be accessed and evaluated
# - grid should come from the make_grid function
# - must have unique bytes along each axis
# block_size
# - comes from length of one block in a sudoku_key
# - this is the length of one edge along the grid
# - each axis is this many unit long exactly
# index
# - list of coordinates to access the grid
# - should be of length word_size
# - should be of length equal to number of dimensions in the grid
def eval_index(grid, block_size, index):
offset = 0
for power, value in enumerate(reversed(index)):
offset += value * block_size ** power
return grid[int(offset)]
################################################################################
# The following functions act as a suite that can ultimately
# encrpyt strings, though other functions can be built from them.
# bytes_obj
# - the bytes to encode
# byte_map
# - byte tranform map for inserting into the index
# grid
# - X dimensional grid used to evaluate markov chains
# index
# - list that starts the index for accessing grid (primer)
# - it should be of length word_size - 1
# block_size
# - length of each edge in a grid
def _encode(bytes_obj, byte_map, grid, index, block_size):
cache = bytes()
index = [0] + index
for byte in bytes_obj:
if byte in byte_map:
index.append(byte_map[byte])
index = index[1:]
cache += bytes([eval_index(grid, block_size, index)])
else:
cache += bytes([byte])
return cache, index[1:]
# bytes_obj
# - the bytes to encode
# sudoko_key
# - should be a proper "markov" key
# - this key will be automatically checked for correctness
# boot_strap
# - should be a proper "markov" bootstrap
def encrypt(bytes_obj, sudoku_key, boot_strap):
_check_sudoku_key(sudoku_key)
_check_boot_strap(boot_strap, sudoku_key)
# make byte_map
array = sorted(sudoku_key[0])
byte_map = dict((byte, value) for value, byte in enumerate(array))
# create two more arguments for encode
grid = make_grid(sudoku_key)
index = list(map(byte_map.__getitem__, boot_strap))
# run the actual encoding algorithm and create reversed map
code, index = _encode(bytes_obj, byte_map, grid, index, len(sudoku_key[0]))
rev_map = dict(reversed(item) for item in byte_map.items())
# fix the boot_strap and return the results
boot_strap = bytes(rev_map[number] for number in index)
return code, boot_strap
# string
# - should be the string that you want encoded
# word_size
# - length you want the markov chains to be of
# plain_chars
# - characters that you do not want to encrypt
def encrypt_str(string, word_size, plain_chars=''):
byte_obj = string.encode(CODEC, 'ignore')
encode_on = set(byte_obj).difference(set(plain_chars.encode()))
sudoku_key = make_sudoku_key(encode_on, word_size)
boot_strap = make_boot_strap(sudoku_key)
cyphertext = encrypt(byte_obj, sudoku_key, boot_strap)[0]
# return encrypted string, key, and original bootstrap
return cyphertext.decode(CODEC, 'ignore'), sudoku_key, boot_strap
def grid_size(string, word_size, plain_chars):
encode_on = set(string.encode()).difference(set(plain_chars.encode()))
return len(encode_on) ** word_size
################################################################################
# The following functions act as a suite that can ultimately
# decrpyt strings, though other functions can be built from them.
# bytes_obj
# - the bytes to encode
# byte_map
# - byte tranform map for inserting into the index
# grid
# - X dimensional grid used to evaluate markov chains
# index
# - list that starts the index for accessing grid (primer)
# - it should be of length word_size - 1
# block_size
# - length of each edge in a grid
def _decode(bytes_obj, byte_map, grid, index, block_size):
cache = bytes()
index = [0] + index
for byte in bytes_obj:
if byte in byte_map:
index.append(byte_map[byte])
index = index[1:]
decoded = eval_index(grid, block_size, index)
index[-1] = byte_map[decoded]
cache += bytes([decoded])
else:
cache += bytes([byte])
return cache, index[1:]
# bytes_obj
# - the bytes to decode
# sudoko_key
# - should be a proper "markov" key
# - this key will be automatically checked for correctness
# boot_strap
# - should be a proper "markov" bootstrap
def decrypt(bytes_obj, sudoku_key, boot_strap):
_check_sudoku_key(sudoku_key)
_check_boot_strap(boot_strap, sudoku_key)
# make byte_map
array = sorted(sudoku_key[0])
byte_map = dict((byte, value) for value, byte in enumerate(array))
# create two more arguments for decode
grid = make_grid(sudoku_key)
grid = make_decode_grid(grid, len(sudoku_key[0]))
index = list(map(byte_map.__getitem__, boot_strap))
# run the actual decoding algorithm and create reversed map
code, index = _decode(bytes_obj, byte_map, grid, index, len(sudoku_key[0]))
rev_map = dict(reversed(item) for item in byte_map.items())
# fix the boot_strap and return the results
boot_strap = bytes(rev_map[number] for number in index)
return code, boot_strap
# string
# - should be the string that you want decoded
# word_size
# - length you want the markov chains to be of
# plain_chars
# - characters that you do not want to encrypt
def decrypt_str(string, sudoku_key, boot_strap):
byte_obj = string.encode(CODEC, 'ignore')
plaintext = decrypt(byte_obj, sudoku_key, boot_strap)[0]
# return encrypted string, key, and original bootstrap
return plaintext.decode(CODEC, 'ignore')
################################################################################
def convert(number):
"Convert bytes into human-readable representation."
assert 0 < number < 1 << 110, 'Number Out Of Range'
ordered = reversed(tuple(format_bytes(partition_number(number, 1 << 10))))
cleaned = ', '.join(item for item in ordered if item[0] != '0')
return cleaned
################################################################################
def partition_number(number, base):
"Continually divide number by base until zero."
div, mod = divmod(number, base)
yield mod
while div:
div, mod = divmod(div, base)
yield mod
def format_bytes(parts):
"Format partitioned bytes into human-readable strings."
for power, number in enumerate(parts):
yield '{} {}'.format(number, format_suffix(power, number))
def format_suffix(power, number):
"Compute the suffix for a certain power of bytes."
return (PREFIX[power] + 'byte').capitalize() + ('s' if number != 1 else '')
################################################################################
PREFIX = ' kilo mega giga tera peta exa zetta yotta bronto geop'.split(' ')
################################################################################
if __name__ == '__main__':
root = Tk()
root.title('Markov Demo 1')
demo = MarkovDemo(root)
root.mainloop()
|
A far more efficient version of this demonstration has been added as recipe 578076.