def test_simple_hmm(self):
s1 = hmm.state('S1', 0.5,
{ '1': 0.5, '2': 0.5 },
{ 'S1': 0.9, 'S2': 0.1 })
s2 = hmm.state('S2', 0.5,
{ '1': 0.25, '2': 0.75 },
{ 'S1': 0.8, 'S2': 0.2 })
model = hmm.hmm(['1', '2'], [s1, s2])
state_path, prob = model.viterbi_path('222')
self.assertEqual(state_path, ['S2', 'S1', 'S1'])
self.assertEqual(round(prob, 6), -1.170696)
def test_implied_terminal_state(self):
s1 = hmm.state('E', 1.0,
{'A': 0.25, 'C': 0.25, 'G': 0.25, 'T': 0.25},
{'E': 0.9, '5': 0.1})
s2 = hmm.state('5', 0.0,
{'A': 0.05, 'C': 0.0, 'G': 0.95, 'T': 0.0},
{'I': 1.0})
s3 = hmm.state('I', 0.0,
{'A': 0.4, 'C': 0.1, 'G': 0.1, 'T': 0.4},
{'I': 0.9},
0.1)
model = hmm.hmm(['A', 'C', 'G', 'T'], [s1, s2, s3])
state_path, prob = model.viterbi_path('CTTCATGTGAAAGCAGACGTAAGTCA')
self.assertEqual(state_path, ['E', 'E', 'E', 'E', 'E', 'E', 'E', 'E',
'E', 'E', 'E', 'E', 'E', 'E', 'E', 'E', 'E', 'E', '5',
'I', 'I', 'I', 'I', 'I', 'I', 'I'])
self.assertEqual(round(prob, 10), -17.9014785649)
def test_implied_terminal_state(self):
# The HMM shown in figure 1 of:
# Eddy SR. 2004. What is a hidden Markov model?
# Nature Biotechnology 22(10): 1315-1316.
s1 = hmm.state('E', 1.0,
{'A': 0.25, 'C': 0.25, 'G': 0.25, 'T': 0.25},
{'E': 0.9, '5': 0.1})
s2 = hmm.state('5', 0.0,
{'A': 0.05, 'C': 0.0, 'G': 0.95, 'T': 0.0},
{'I': 1.0})
s3 = hmm.state('I', 0.0,
{'A': 0.4, 'C': 0.1, 'G': 0.1, 'T': 0.4},
{'I': 0.9},
0.1)
model = hmm.hmm(['A', 'C', 'G', 'T'], [s1, s2, s3])
self.assertEqual(len(model.initial_states), 1)
self.assertTrue(model.terminal_state)
self.assertEqual(len(model.terminating_states), 1)
def test_simple_hmm(self):
# The simple HMM shown in section 12.1 of Ewens and Grant, "Statistical
# Methods in Bioinformatics: An Introduction, 2nd Ed.", 2005
s1 = hmm.state(
'S1', # name of the state
0.5, # probability of being the initial state
{ '1': 0.5, # probability of emitting a '1' at each visit
'2': 0.5 }, # probability of emitting a '2' at each visit
{ 'S1': 0.9, # probability of transitioning to itself
'S2': 0.1 }) # probability of transitioning to state 'S2'
s2 = hmm.state('S2', 0.5,
{ '1': 0.25, '2': 0.75 },
{ 'S1': 0.8, 'S2': 0.2 })
model = hmm.hmm(['1', '2'], # all symbols that can be emitted
[s1, s2]) # all of the states in this HMM
self.assertEqual(model.states['S1'].p_initial, 0.5)
self.assertEqual(model.states['S2'].p_transition['S1'], 0.8)
self.assertEqual(len(model.initial_states), 2)
self.assertFalse(model.terminal_state)
self.assertEqual(len(model.terminating_states), 0)
def test_repr(self):
s1 = hmm.state('E', 1.0,
{'A': 0.25, 'C': 0.25, 'G': 0.25, 'T': 0.25},
{'E': 0.9, '5': 0.1})
s2 = hmm.state('5', 0.0,
{'A': 0.05, 'C': 0.0, 'G': 0.95, 'T': 0.0},
{'I': 1.0})
s3 = hmm.state('I', 0.0,
{'A': 0.4, 'C': 0.1, 'G': 0.1, 'T': 0.4},
{'I': 0.9},
0.1)
model = hmm.hmm(['A', 'C', 'G', 'T'], [s1, s2, s3])
model2 = eval(repr(model))
self.assertEqual(model.alphabet, model2.alphabet)
self.assertEqual(model.states['E'].p_transition['5'],
model2.states['E'].p_transition['5'])
self.assertEqual(model.states['5'].p_emission['T'],
model2.states['5'].p_emission['T'])
self.assertEqual(model.states['I'].p_termination,
model2.states['I'].p_termination)
import hmm
s1 = hmm.state(
'S1', # name of the state
0.5, # probability of being the initial state
{ '1': 0.5, # probability of emitting a '1' at each visit
'2': 0.5 }, # probability of emitting a '2' at each visit
{ 'S1': 0.9, # probability of transitioning to itself
'S2': 0.1 } # probability of transitioning to state 'S2'
)
s2 = hmm.state(
'S2', 0.5,
{ '1': 0.25, '2': 0.75 },
{ 'S1': 0.8, 'S2': 0.2 }
)
model = hmm.hmm(['1', '2'], # all symbols that can be emitted
[s1, s2]) # all of the states in this HMM
model.enumerate('222')
path, prob = model.viterbi_path('222')
print path
print prob
with open("./%s" %sys.argv[2], "rb") as infile:
reader = csv.reader(infile, delimiter=" ")
for row in reader:
obsDef.append(row)
numSent = int(obsDef[0][0])
a = np.array(hmmDef[4:8]).astype(float)
b = np.array(hmmDef[9:13]).astype(float)
pi = np.array(hmmDef[-1]).astype(float)
symList = np.array(hmmDef[2])
stateList = []
for stateNum in range(int(hmmDef[0][0])):
stateList.append(hmm.state(hmmDef[1][stateNum], stateNum, pi[stateNum]))
connectList = []
for stateFrom in stateList:
#print stateFrom.name, stateFrom.index, stateFrom.init
for stateTo in stateList:
if a[stateFrom.index][stateTo.index] > 0.0:
connectList.append(hmm.connect(stateFrom, stateTo, a[stateFrom.index][stateTo.index]))
outputList = []
for origState in stateList:
temp = []
for i in range(int(hmmDef[0][1])):
temp.append(hmm.output(origState, symList[i], b[origState.index][i], i))
outputList.append(temp)
