def encode_bits_as_words(chain: MarkovChain,
wt_dict: WordTypeDictionary,
bits: Bits,
pad_text=True) -> list:
"""
Given a bit stream, a Markov chain, and a word-type dictionary, retrieve a corresponding list of words.
Every state in the Markov chain, except the start state s0, must have a corresponding word-type in the given
dictionary with at least one word.
If the word-type dictionary does not have path bits to match the end of the input exactly, it will append 0s
until the function can complete.
:param chain: a Markov chain with states
:param wt_dict: a corresponding dictionary of word-types
:param bits: the input bits
:param pad_text: if true, generate cover text from random bits until the Markov chain reaches state s0
:return: an ordered list of words encoded by the system
"""
if bits is None or bits.__eq__(Bits()):
raise ValueError("Bits cannot be None or empty.")
words = []
prefix = bits
while len(prefix) > 0:
chain.transition()
if chain.current_state.__eq__(START_STATE_LABEL):
chain.transition()
word, word_bits, encode_spaces = _encode_next_word(
chain, wt_dict, prefix)
words.append((word, encode_spaces))
bit_length = len(word_bits)
prefix = prefix[bit_length:]
if pad_text:
# add filler bits until s0 reached
while not chain.current_state.__eq__(START_STATE_LABEL):
chain.transition()
if chain.current_state.__eq__(START_STATE_LABEL):
break
longest_word = get_longest_word_in_dictionary(wt_dict)
pseudo_random_bits = Bits(bin="".join(
random.choice(["0", "1"]) for _ in range(len(longest_word))))
word, word_bits, encode_spaces = _encode_next_word(
chain, wt_dict, pseudo_random_bits)
words.append((word, encode_spaces))
return words
class TestMarkovChainTransitions(unittest.TestCase):
def setUp(self):
self.states = {"s1", "s2", "s3", ("s4", "dict")}
self.markov_chain = MarkovChain(self.states)
self.transitions = set()
self.transitions.add(("s0", "s1", 2))
self.transitions.add(("s0", "s2", 3))
self.transitions.add(("s1", "s3", 0.5))
self.transitions.add(("s1", "s4", 0.5))
self.transitions.add(("s2", "s4", 1))
self.transitions.add(("s3", "s0", 1))
self.transitions.add(("s4", "s0", 2))
def test_set_transitions(self):
self.markov_chain.set_transitions(self.transitions)
self.assertEqual(5, len(self.markov_chain.markov_chain))
self.assertIsInstance(self.markov_chain.markov_chain.get("s0"),
StateTransitions)
self.assertDictEqual(
{
"s1": 2,
"s2": 3
},
self.markov_chain.markov_chain.get("s0").transitions)
def test_set_transitions_bad_outbound(self):
self.transitions.add(("s4", "s5", 1))
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_set_transitions_bad_inbound(self):
self.transitions.add(("s5", "s1", 1))
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_set_transitions_neg_probability(self):
self.transitions.add(("s3", "s4", -1))
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_set_transitions_zero_probability(self):
self.transitions.add(("s3", "s4", 0))
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_set_transitions_s0_no_outbound(self):
self.transitions = {
x
for x in self.transitions if not x[0].__eq__("s0")
}
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_set_transitions_s0_no_inbound(self):
self.transitions = {
x
for x in self.transitions if not x[1].__eq__("s0")
}
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_dict(self):
self.markov_chain.set_transitions(self.transitions)
serial_dict = self.markov_chain.__dict__()
self.assertEqual(2, len(serial_dict))
chain_dict = serial_dict.get("chain")
self.assertIsInstance(chain_dict, dict)
self.assertEqual(5, len(chain_dict))
self.assertIsInstance(chain_dict.get("s0"), dict)
self.assertDictEqual({"s1": 2, "s2": 3}, chain_dict.get("s0"))
def test_get_outbound_transitions(self):
self.markov_chain.set_transitions(self.transitions)
self.markov_chain.current_state = "s0"
outbound_transitions = self.markov_chain.get_outbound_transitions()
self.assertIsInstance(outbound_transitions, StateTransitions)
self.assertEqual(2, len(outbound_transitions.transitions))
self.assertEqual(2, outbound_transitions.transitions.get("s1"))
self.assertEqual(3, outbound_transitions.transitions.get("s2"))
self.markov_chain.current_state = "s4"
outbound_transitions = self.markov_chain.get_outbound_transitions()
self.assertIsInstance(outbound_transitions, StateTransitions)
self.assertEqual(1, len(outbound_transitions.transitions))
self.assertEqual(2, outbound_transitions.transitions.get("s0"))
def test_accumulate_transitions(self):
self.markov_chain.set_transitions(self.transitions)
self.markov_chain.current_state = "s0"
accum = self.markov_chain.accumulate_transitions()
self.assertIsInstance(accum, list)
self.assertEqual(2, len(accum))
self.assertTrue(any(x[1] == 5) for x in accum)
def test_transition(self):
self.markov_chain.set_transitions(self.transitions)
self.markov_chain.current_state = "s0"
self.markov_chain.transition()
self.assertIn(self.markov_chain.current_state, {"s1", "s2"})
self.markov_chain.transition()
self.assertIn(self.markov_chain.current_state, {"s3", "s4"})
self.markov_chain.transition()
self.assertEqual("s0", self.markov_chain.current_state)
def test_find_2_cycle(self):
self.transitions.add(("s4", "s2", 1))
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_find_3_cycle(self):
self.transitions = {
x
for x in self.transitions if not x[1].__eq__("s3")
}
self.transitions.add(("s4", "s3", 1))
self.transitions.add(("s3", "s1", 1))
self.assertRaises(MarkovError, self.markov_chain.set_transitions,
self.transitions)
def test_ignore_non_cycle(self):
self.transitions.add(("s1", "s2", 1))
self.markov_chain.set_transitions(self.transitions)
def test_get_number_of_paths(self):
self.markov_chain.set_transitions(self.transitions)
no_of_paths = markov.get_number_of_paths(self.markov_chain)
self.assertEqual(3, no_of_paths)
self.states.add("s5")
self.markov_chain = MarkovChain(self.states)
self.transitions.add(("s2", "s5", 1))
self.transitions.add(("s4", "s5", 1))
self.transitions.add(("s5", "s0", 1))
self.markov_chain.set_transitions(self.transitions)
no_of_paths = markov.get_number_of_paths(self.markov_chain)
self.assertEqual(6, no_of_paths)
@unittest.skip
def test_save(self):
self.markov_chain.set_transitions(self.transitions)
markov.save_markov_chain(self.markov_chain)