-
Notifications
You must be signed in to change notification settings - Fork 0
/
automaton.py
296 lines (245 loc) · 9.19 KB
/
automaton.py
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
from typing import List, Optional, Dict, Callable, Union
from subprocess import Popen
import logging
import os
import random
import speech_recognition as sr
from custom_recognizer import CustomRecognizer
from keywords import KeywordModel
DEBUG = 1
__all__ = ["State", "VoiceControlledAutomaton", "Exit"]
class State:
# all subclass VCA FSA's states must inherit from this
enter = 0
exit = 1
class Exit(Exception):
def __init__(self, *args, **kwargs):
super().__init__(self, *args, **kwargs)
if self.args:
self.text = args[0]
else:
self.text = ""
def __bool__(self):
return bool(self.text)
class VoiceControlledAutomaton:
"""
A Voice Controlled
Finite State Automaton.
States may be VCAs as well.
NOTE
A singular underscore ("_") at beginning of method
denotes not privacy/visibility, but that this method should be
overwritten by subclasses.
###########################
TODO FIXME NOTE
This object has more methods/attributes than necessary
for an FSA. TODO clean up structure/make simpler
###########################
"""
keyword_dir="/home/pi/audio/hal/keywords/",
kw_model_dir="/home/pi/audio/hal/keywords/models/",
kw_model_ext= ".pt"
def __init__(
self,
_super = None,
name = None,
mic_index=1,
sound_dir="./wavs/",
kw_model_path: str="/home/pi/audio/hal/models/audio_model_fp32.pt",
n_keywords: int=8,
sampling_rate: int=16000,
log_automaton_utterances: bool=True,
log_user_utterances: bool=True,
**kwargs
):
kwargs["_super"] = self
"""Subclass inits should first call super init, then append to self.keywords"""
self.mic_index = int(mic_index)
self.sound_dir = str(sound_dir)
self.super = None
self.keywords: List[List[str]] = []
if _super is not None:
self.super = _super
assert isinstance(_super, VoiceControlledAutomaton)
self.keywords += _super.keywords
self.kw_model = _super.kw_model
self.R = _super.R
self.name = name
self.logger = _super.logger
self.log_user_utterances = _super.log_user_utterances
self.log_automaton_utterances = _super.log_automaton_utterances
else:
self.R = CustomRecognizer()
self.name = name
self.logger = logging.getLogger(name=self.name)
self.logger.setLevel(logging.INFO)
if DEBUG: kw_model_path = None
self.kw_model = KeywordModel(model_path=kw_model_path)
self.log_user_utterances = log_user_utterances
self.log_automaton_utterances = log_automaton_utterances
# initial state, may be overwritten by subclasses
self.set_state(State.enter)
# which function to call depending on own state
self.SideEffectTransitionMatrix: Dict[State, Callable[str, Union[State, Exit]]] = NotImplemented
self.state_keywords: Dict[State, List[List[str]]] = NotImplemented
# def load_keyword_model(self):
# """
# load model, dependent on self.state and self.keywords
# """
# raise NotImplementedError
# model_path = VoiceControlledAutomaton.kw_model_dir + self.name + VoiceControlledAutomaton.kw_model_ext
# model = ...
# return model
def play_sound(self, soundfile: str):
f = os.path.join(self.sound_dir, soundfile+".mp3")
if not os.path.exists(f):
f = os.path.join(self.sound_dir, soundfile+".wav")
Popen(["aplay", f])
else:
Popen(["mpg123", f]) #, close_fds=1)
def speak(self, s, wait=True) -> None:
say_process = Popen(["say", s], shell=False)
if self.log_automaton_utterances:
self.logger.info(s)
if wait:
say_process.wait()
def keyword_transition(self) -> None:
# listen to input and make state transition with side effects
text = self.get_utterance(keyword=True)
if type(text) != int:
# within respond to input, we may enter a lower FSA
self.transition(text)
def transition(self, text):
new_state = self.respond_to_input(text)
self.set_state(new_state)
# self.kw_model = self.load_keyword_model()
return new_state
def __call__(self, text):
# start self.run after responding to text
# (used by higher level FSA)
self.transition(text)
return self.run()
def run(self):
while True:
try:
self.keyword_transition()
except Exit as couldnt_handle:
print(f"{self} received Exit")
# self.keyword_transition()
self.reset()
if self.super is not None:
print(f"{self} passes to super")
raise couldnt_handle
else:
print(f"{self} passes")
pass
def reset(self):
self.state = State.enter
def respond_to_input(self, text):
"""
Rule based side effects during state transition
"""
new_state = self.SideEffectTransitionMatrix[self.state](text)
return new_state
def listen(self, for_keyword=True):
with sr.Microphone(device_index=self.mic_index) as source:
self.R.adjust_for_ambient_noise(source)
self.logger.info("Waiting for voice input")
if for_keyword:
if not DEBUG:
# TODO implement custom recognizer keyword model
audio = self.R.listen_from_keyword_on(
source,
keyword_model=self.kw_model
)
else:
# while listen_from_keyword_on not impl, use this:
audio = self.R.listen(source)
else:
audio = self.R.listen(source)
# got possibly empty/garbage audio by this point
self.play_sound("blang")
return audio
def recognize(self, audio):
try:
text = self.R.recognize_google(audio).lower()
# text = self.R.recognize_sphinx(audio).lower()
self.play_sound("blung")
# self.speak(random.choice([
# "alrighty",
# "alrighty",
# "okay",
# "ok",
# "gotcha",
# "gotchu",
# "k",
# "i am at your service"
# ]))
self.logger.info("got non-empty input: "+text)
return text
except sr.UnknownValueError:
# empty/non-parseable utterance
# err_msg = f"Repeat that, please."
# self.speak(err_msg)
# print(err_msg)
return 1
except sr.RequestError as e:
self.logger.warn("Could not request results")
return 2
def get_utterance(self, keyword=True) -> str:
audio = self.listen(for_keyword=keyword)
text = self.recognize(audio)
return text
def exit(self, text: str) -> State:
self._exit_effects(text)
raise Exit(text)
def _exit_effects(self, text):
# overwrite me
return
def remind_options(self):
if len(self.keywords) < 4:
s = " or ".join(self.keywords) + "?"
else:
self.speak(f"{self} has the following options:")
# self.speak("You can currently say one of the following:")
s = ", ".join(self.keywords)
self.speak(s)
def react_to_choice(self, choice: str):
"""
After hearing the first keyword while waiting for a command,
The FSA can use this method to get the next state from
the choice on what to do,
and then delegate the response to another function/method
In other words, this may get called by VCA.respond_to_input
(e.g. in the idle state), change the state, and call
VCA.respond_to_input again
"""
# see if initial utterance already contains valid choice
state = self._parse_choice(choice, must_understand=False)
if state is None:
# in this case, ask about desired activity
# self.remind_options()
# while True:
# choice = self.get_utterance(keyword=False)
# if type(choice) == int:
# pass
# else:
# state = self._parse_choice(choice, must_understand=True)
# # got a next state? -> exit loop
# if state is not None:
# break
return self.state
# self.play_sound("pling")
self.set_state(state)
return self.transition(choice)
def set_state(self, state):
self._set_state_effects(state)
self.state = state
def _set_state_effects(self, state):
# overwrite me
return
def _parse_choice(self, text: str, must_understand:bool = False) -> Optional[State]:
# overwrite me
raise NotImplementedError
def __str__(self):
return self.name