{{ message }}
forked from TheAlgorithms/Python
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathviterbi.py
More file actions
377 lines (347 loc) · 13.2 KB
/
Copy pathviterbi.py
File metadata and controls
377 lines (347 loc) · 13.2 KB
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
from typing import Any
def viterbi(
observations_space: list,
states_space: list,
initial_probabilities: dict,
transition_probabilities: dict,
emission_probabilities: dict,
) -> list:
"""
Viterbi Algorithm, to find the most likely path of
states from the start and the expected output.
https://en.wikipedia.org/wiki/Viterbi_algorithm
Wikipedia example
>>> observations = ["normal", "cold", "dizzy"]
>>> states = ["Healthy", "Fever"]
>>> start_p = {"Healthy": 0.6, "Fever": 0.4}
>>> trans_p = {
... "Healthy": {"Healthy": 0.7, "Fever": 0.3},
... "Fever": {"Healthy": 0.4, "Fever": 0.6},
... }
>>> emit_p = {
... "Healthy": {"normal": 0.5, "cold": 0.4, "dizzy": 0.1},
... "Fever": {"normal": 0.1, "cold": 0.3, "dizzy": 0.6},
... }
>>> viterbi(observations, states, start_p, trans_p, emit_p)
['Healthy', 'Healthy', 'Fever']
>>> viterbi((), states, start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: There's an empty parameter
>>> viterbi(observations, (), start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: There's an empty parameter
>>> viterbi(observations, states, {}, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: There's an empty parameter
>>> viterbi(observations, states, start_p, {}, emit_p)
Traceback (most recent call last):
...
ValueError: There's an empty parameter
>>> viterbi(observations, states, start_p, trans_p, {})
Traceback (most recent call last):
...
ValueError: There's an empty parameter
>>> viterbi("invalid", states, start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: observations_space must be a list
>>> viterbi(["valid", 123], states, start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: observations_space must be a list of strings
>>> viterbi(observations, "invalid", start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: states_space must be a list
>>> viterbi(observations, ["valid", 123], start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: states_space must be a list of strings
>>> viterbi(observations, states, "invalid", trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: initial_probabilities must be a dict
>>> viterbi(observations, states, {2:2}, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: initial_probabilities all keys must be strings
>>> viterbi(observations, states, {"a":2}, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: initial_probabilities all values must be float
>>> viterbi(observations, states, start_p, "invalid", emit_p)
Traceback (most recent call last):
...
ValueError: transition_probabilities must be a dict
>>> viterbi(observations, states, start_p, {"a":2}, emit_p)
Traceback (most recent call last):
...
ValueError: transition_probabilities all values must be dict
>>> viterbi(observations, states, start_p, {2:{2:2}}, emit_p)
Traceback (most recent call last):
...
ValueError: transition_probabilities all keys must be strings
>>> viterbi(observations, states, start_p, {"a":{2:2}}, emit_p)
Traceback (most recent call last):
...
ValueError: transition_probabilities all keys must be strings
>>> viterbi(observations, states, start_p, {"a":{"b":2}}, emit_p)
Traceback (most recent call last):
...
ValueError: transition_probabilities nested dictionary all values must be float
>>> viterbi(observations, states, start_p, trans_p, "invalid")
Traceback (most recent call last):
...
ValueError: emission_probabilities must be a dict
>>> viterbi(observations, states, start_p, trans_p, None)
Traceback (most recent call last):
...
ValueError: There's an empty parameter
"""
_validation(
observations_space,
states_space,
initial_probabilities,
transition_probabilities,
emission_probabilities,
)
# Creates data structures and fill initial step
probabilities: dict = {}
pointers: dict = {}
for state in states_space:
observation = observations_space[0]
probabilities[(state, observation)] = (
initial_probabilities[state] * emission_probabilities[state][observation]
)
pointers[(state, observation)] = None
# Fills the data structure with the probabilities of
# different transitions and pointers to previous states
for o in range(1, len(observations_space)):
observation = observations_space[o]
prior_observation = observations_space[o - 1]
for state in states_space:
# Calculates the argmax for probability function
arg_max = ""
max_probability = -1
for k_state in states_space:
probability = (
probabilities[(k_state, prior_observation)]
* transition_probabilities[k_state][state]
* emission_probabilities[state][observation]
)
if probability > max_probability:
max_probability = probability
arg_max = k_state
# Update probabilities and pointers dicts
probabilities[(state, observation)] = (
probabilities[(arg_max, prior_observation)]
* transition_probabilities[arg_max][state]
* emission_probabilities[state][observation]
)
pointers[(state, observation)] = arg_max
# The final observation
final_observation = observations_space[len(observations_space) - 1]
# argmax for given final observation
arg_max = ""
max_probability = -1
for k_state in states_space:
probability = probabilities[(k_state, final_observation)]
if probability > max_probability:
max_probability = probability
arg_max = k_state
last_state = arg_max
# Process pointers backwards
previous = last_state
result = []
for o in range(len(observations_space) - 1, -1, -1):
result.append(previous)
previous = pointers[previous, observations_space[o]]
result.reverse()
return result
def _validation(
observations_space: Any,
states_space: Any,
initial_probabilities: Any,
transition_probabilities: Any,
emission_probabilities: Any,
) -> None:
"""
>>> observations = ["normal", "cold", "dizzy"]
>>> states = ["Healthy", "Fever"]
>>> start_p = {"Healthy": 0.6, "Fever": 0.4}
>>> trans_p = {
... "Healthy": {"Healthy": 0.7, "Fever": 0.3},
... "Fever": {"Healthy": 0.4, "Fever": 0.6},
... }
>>> emit_p = {
... "Healthy": {"normal": 0.5, "cold": 0.4, "dizzy": 0.1},
... "Fever": {"normal": 0.1, "cold": 0.3, "dizzy": 0.6},
... }
>>> _validation(observations, states, start_p, trans_p, emit_p)
>>> _validation([], states, start_p, trans_p, emit_p)
Traceback (most recent call last):
...
ValueError: There's an empty parameter
"""
_validate_not_empty(
observations_space,
states_space,
initial_probabilities,
transition_probabilities,
emission_probabilities,
)
_validate_lists(observations_space, states_space)
_validate_dicts(
initial_probabilities, transition_probabilities, emission_probabilities
)
def _validate_not_empty(
observations_space: Any,
states_space: Any,
initial_probabilities: Any,
transition_probabilities: Any,
emission_probabilities: Any,
) -> None:
"""
>>> _validate_not_empty(["a"], ["b"], {"c":0.5},
... {"d": {"e": 0.6}}, {"f": {"g": 0.7}})
>>> _validate_not_empty(["a"], ["b"], {"c":0.5}, {}, {"f": {"g": 0.7}})
Traceback (most recent call last):
...
ValueError: There's an empty parameter
>>> _validate_not_empty(["a"], ["b"], None, {"d": {"e": 0.6}}, {"f": {"g": 0.7}})
Traceback (most recent call last):
...
ValueError: There's an empty parameter
"""
if not all(
[
observations_space,
states_space,
initial_probabilities,
transition_probabilities,
emission_probabilities,
]
):
raise ValueError("There's an empty parameter")
def _validate_lists(observations_space: Any, states_space: Any) -> None:
"""
>>> _validate_lists(["a"], ["b"])
>>> _validate_lists(1234, ["b"])
Traceback (most recent call last):
...
ValueError: observations_space must be a list
>>> _validate_lists(["a"], [3])
Traceback (most recent call last):
...
ValueError: states_space must be a list of strings
"""
_validate_list(observations_space, "observations_space")
_validate_list(states_space, "states_space")
def _validate_list(_object: Any, var_name: str) -> None:
"""
>>> _validate_list(["a"], "mock_name")
>>> _validate_list("a", "mock_name")
Traceback (most recent call last):
...
ValueError: mock_name must be a list
>>> _validate_list([0.5], "mock_name")
Traceback (most recent call last):
...
ValueError: mock_name must be a list of strings
"""
if not isinstance(_object, list):
msg = f"{var_name} must be a list"
raise ValueError(msg)
else:
for x in _object:
if not isinstance(x, str):
msg = f"{var_name} must be a list of strings"
raise ValueError(msg)
def _validate_dicts(
initial_probabilities: Any,
transition_probabilities: Any,
emission_probabilities: Any,
) -> None:
"""
>>> _validate_dicts({"c":0.5}, {"d": {"e": 0.6}}, {"f": {"g": 0.7}})
>>> _validate_dicts("invalid", {"d": {"e": 0.6}}, {"f": {"g": 0.7}})
Traceback (most recent call last):
...
ValueError: initial_probabilities must be a dict
>>> _validate_dicts({"c":0.5}, {2: {"e": 0.6}}, {"f": {"g": 0.7}})
Traceback (most recent call last):
...
ValueError: transition_probabilities all keys must be strings
>>> _validate_dicts({"c":0.5}, {"d": {"e": 0.6}}, {"f": {2: 0.7}})
Traceback (most recent call last):
...
ValueError: emission_probabilities all keys must be strings
>>> _validate_dicts({"c":0.5}, {"d": {"e": 0.6}}, {"f": {"g": "h"}})
Traceback (most recent call last):
...
ValueError: emission_probabilities nested dictionary all values must be float
"""
_validate_dict(initial_probabilities, "initial_probabilities", float)
_validate_nested_dict(transition_probabilities, "transition_probabilities")
_validate_nested_dict(emission_probabilities, "emission_probabilities")
def _validate_nested_dict(_object: Any, var_name: str) -> None:
"""
>>> _validate_nested_dict({"a":{"b": 0.5}}, "mock_name")
>>> _validate_nested_dict("invalid", "mock_name")
Traceback (most recent call last):
...
ValueError: mock_name must be a dict
>>> _validate_nested_dict({"a": 8}, "mock_name")
Traceback (most recent call last):
...
ValueError: mock_name all values must be dict
>>> _validate_nested_dict({"a":{2: 0.5}}, "mock_name")
Traceback (most recent call last):
...
ValueError: mock_name all keys must be strings
>>> _validate_nested_dict({"a":{"b": 4}}, "mock_name")
Traceback (most recent call last):
...
ValueError: mock_name nested dictionary all values must be float
"""
_validate_dict(_object, var_name, dict)
for x in _object.values():
_validate_dict(x, var_name, float, True)
def _validate_dict(
_object: Any, var_name: str, value_type: type, nested: bool = False
) -> None:
"""
>>> _validate_dict({"b": 0.5}, "mock_name", float)
>>> _validate_dict("invalid", "mock_name", float)
Traceback (most recent call last):
...
ValueError: mock_name must be a dict
>>> _validate_dict({"a": 8}, "mock_name", dict)
Traceback (most recent call last):
...
ValueError: mock_name all values must be dict
>>> _validate_dict({2: 0.5}, "mock_name",float, True)
Traceback (most recent call last):
...
ValueError: mock_name all keys must be strings
>>> _validate_dict({"b": 4}, "mock_name", float,True)
Traceback (most recent call last):
...
ValueError: mock_name nested dictionary all values must be float
"""
if not isinstance(_object, dict):
msg = f"{var_name} must be a dict"
raise ValueError(msg)
if not all(isinstance(x, str) for x in _object):
msg = f"{var_name} all keys must be strings"
raise ValueError(msg)
if not all(isinstance(x, value_type) for x in _object.values()):
nested_text = "nested dictionary " if nested else ""
msg = f"{var_name} {nested_text}all values must be {value_type.__name__}"
raise ValueError(msg)
if __name__ == "__main__":
from doctest import testmod
testmod()
You can’t perform that action at this time.