def test_hmm_transition_matrix(self):
hmm = HMM({INITIAL_STATE: ['q1'],
'q1': (['q2', 'q1', FINAL_STATE], ['dag', 'kot']),
'q2': (['q1', FINAL_STATE], ['z'])})
log_hmm(hmm)
transition_matrix = hmm.get_transition_matrix()
print(transition_matrix)
def test_get_totally_random_hmm(self):
configurations['MIN_NUM_OF_INNER_STATES'] = 1
configurations['MAX_NUM_OF_INNER_STATES'] = 5
configurations['RANDOM_HMM_MAX_EMISSIONS_PER_STATE'] = 5
self.initialise_segment_table("plural_english_segment_table.txt")
hmm = HMM.get_random_hmm_totally_random(data=['dag', 'dagzook', 'kook', 'kooz'])
log_hmm(hmm)
self.write_to_dot_to_file(hmm, "random_hmm")
def test_remove_states(self):
hmm = HMM({INITIAL_STATE: ['q1'],
'q1': (['q4', FINAL_STATE], ['a']),
'q4': (['q8', 'q8', FINAL_STATE], ['d']),
'q7': (['q8'], ['d']),
'q8': (['q2', FINAL_STATE], ['g']),
'q2': (['q1', FINAL_STATE], ['z'])})
hmm.remove_states(['q1'])
print(hmm.inner_states)
self.write_to_dot_to_file(hmm, 'after_remove')
log_hmm(hmm)
def test_defragment_states(self):
hmm = HMM({INITIAL_STATE: ['q1'],
'q1': (['q4', FINAL_STATE], ['a']),
'q4': (['q10', 'q10', FINAL_STATE], ['d']),
'q7': ([], ['d']),
'q10': (['q2', FINAL_STATE], ['g']),
'q2': (['q1', FINAL_STATE], ['z'])})
self.write_to_dot_to_file(hmm, 'before_defragment')
hmm.ensure_contiguous_state_indices()
print(hmm.inner_states)
self.write_to_dot_to_file(hmm, 'after_defragment')
log_hmm(hmm)
def test_remove_random(self):
self.initialise_segment_table("plural_english_segment_table.txt")
for _ in range(100000):
hmm = HMM.get_random_hmm()
try:
hmm.make_mutation()
# hmm.remove_random_state()
hmm.get_transducer()
except Exception:
from traceback import print_exc
log_hmm(hmm)
print_exc()
break
def test_hmm_connected_components(self):
hmm = HMM({INITIAL_STATE: ['q1'],
'q1': (['q2', FINAL_STATE], ['dag', 'kot']),
'q2': (['q1', FINAL_STATE], ['z'])})
log_hmm(hmm)
component_states = hmm.get_connected_components(ignore_initial_and_final_states=False)
self.write_to_dot_to_file(hmm, 'connected_hmm')
print(component_states)
assert component_states[0][0] == FINAL_STATE
assert component_states[2][0] == INITIAL_STATE
assert 'q1' in component_states[1]
assert 'q2' in component_states[1]
component_states = hmm.get_connected_components(ignore_initial_and_final_states=True)
assert 'q1' in component_states[0]
assert 'q2' in component_states[0]
assert [INITIAL_STATE] not in component_states
assert [FINAL_STATE] not in component_states
def test_random_hmms_crossover(self):
self.initialise_segment_table("plural_english_segment_table.txt")
for _ in range(1):
hmm_1 = HMM.get_random_hmm(data=['dag', 'zook', 'kook', 'kooz'])
hmm_2 = HMM.get_random_hmm(data=['dag', 'zook', 'kook', 'kooz'])
print('HMM 1')
log_hmm(hmm_1)
print('HMM 2')
log_hmm(hmm_2)
offspring_1, offspring_2 = HMM.crossover(hmm_1, hmm_2)
print('Offspring 1')
print('HMM', offspring_1.inner_states, 'TRANSITIONS', offspring_1.transitions, 'EMISSIONS',
offspring_1.emissions)
print('Offspring 2')
print('HMM', offspring_2.inner_states, 'TRANSITIONS', offspring_2.transitions, 'EMISSIONS',
offspring_2.emissions)
self.write_to_dot_to_file(hmm_1, 'random_parent_1')
self.write_to_dot_to_file(hmm_2, 'random_parent_2')
self.write_to_dot_to_file(offspring_1, 'random_offspring_1')
self.write_to_dot_to_file(offspring_2, 'random_offspring_2')
offspring_1.get_transducer()
offspring_2.get_transducer()
def test_crossover_by_matrix(self):
hmm_1 = HMM({INITIAL_STATE: ['q1', 'q2'],
'q1': (['q2', FINAL_STATE], ['koko', 'dogo']),
'q2': ([FINAL_STATE], ['zz'])})
hmm_2 = HMM({INITIAL_STATE: ['q1'],
'q1': (['q2'], ['dog', 'kat']),
'q2': (['q3'], ['s']),
'q3': ([FINAL_STATE], ['z'])})
print("Parent 1")
log_hmm(hmm_1)
print("Parent 2")
log_hmm(hmm_2)
offspring_1, offspring_2 = HMM.crossover_by_matrix(hmm_1, hmm_2)
print("Offspring 1")
log_hmm(offspring_1)
print("Offspring 2")
log_hmm(offspring_2)
self.write_to_dot_to_file(hmm_1, 'matrix_parent_1')
self.write_to_dot_to_file(hmm_2, 'matrix_parent_2')
self.write_to_dot_to_file(offspring_1, 'matrix_offspring_1')
self.write_to_dot_to_file(offspring_2, 'matrix_offspring_2')
def test_crossover(self):
hmm_1 = HMM({INITIAL_STATE: ['q1'],
'q1': (['q2', FINAL_STATE], ['koko', 'dogo']),
'q2': ([FINAL_STATE], ['zz'])})
hmm_2 = HMM({INITIAL_STATE: ['q1'],
'q1': (['q2'], ['dog', 'kat']),
'q2': (['q3'], ['s']),
'q3': ([FINAL_STATE], ['z'])})
print("Parent 1")
log_hmm(hmm_1)
print()
print("Parent 2")
log_hmm(hmm_2)
offspring_1, offspring_2 = HMM.crossover(hmm_1, hmm_2)
print()
print("Offspring 1")
log_hmm(offspring_1)
print()
print("Offspring 2")
log_hmm(offspring_2)