-
Notifications
You must be signed in to change notification settings - Fork 2
/
tts_handler.py
193 lines (166 loc) · 6.95 KB
/
tts_handler.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
import json
import threading
import queue
import time
import os
import pyaudio
from RealtimeTTS import TextToAudioStream, CoquiEngine
from lib.sentencequeue import ThreadSafeSentenceQueue, Sentence
from lib.bufferstream import BufferStream
class TTSHandler:
def __init__(self, config_file='tts_config.json'):
with open(config_file, 'r') as f:
self.config = json.load(f)
self.references_folder = self.config['references_folder']
self.dbg_log = self.config['dbg_log']
self.stop_event = threading.Event()
self.sentence_queue = ThreadSafeSentenceQueue()
self.chunk_queue = queue.Queue()
self.chunk_lock = threading.Lock()
self.pyFormat = pyaudio.paInt16
self.pyChannels = 1
self.pySampleRate = 24000
self.pyOutput_device_index = None
print("Loading TTS")
if self.config['use_local_model']:
self.engine = CoquiEngine(
specific_model=self.config['specific_model'],
local_models_path=self.config['local_models_path']
)
else:
self.engine = CoquiEngine()
self.stream = TextToAudioStream(self.engine, muted=True)
if self.dbg_log:
print("Test Play TTS")
self.stream.feed("hi!") # only small warmup
self.stream.play(log_synthesized_text=True, muted=True)
def initialize_pyaudio(self):
self.stop_event = threading.Event()
self.sentence_queue = ThreadSafeSentenceQueue()
self.chunk_queue = queue.Queue()
self.pyaudio_instance = pyaudio.PyAudio()
self.pystream = self.pyaudio_instance.open(
format=self.pyFormat,
channels=self.pyChannels,
rate=self.pySampleRate,
output_device_index=self.pyOutput_device_index,
output=True)
self.pystream.start_stream()
def tts_play_worker_thread(self):
while not self.stop_event.is_set():
if self.chunk_queue.empty():
time.sleep(0.001)
continue
with self.chunk_lock:
chunk = self.chunk_queue.get()
self.pystream.write(chunk)
def start_tts(self):
def on_audio_chunk(chunk):
with self.chunk_lock:
self.chunk_queue.put(chunk)
self.stream.play_async(
fast_sentence_fragment=True,
log_synthesized_text=True,
muted=True,
on_audio_chunk=on_audio_chunk,
minimum_sentence_length=10,
minimum_first_fragment_length=10,
context_size=5,
sentence_fragment_delimiters=".?!;:,\n…)]}。",
force_first_fragment_after_words=999999,
)
def tts_play_sentence(self, sentence: Sentence):
if sentence.get_finished():
sentence_text = sentence.get_text()
if self.dbg_log:
print(f"tts_play_sentence complete sentence found, playing {sentence_text}")
self.stream.feed(sentence_text)
if self.dbg_log:
print("tts_play_sentence [STARTPLAY]")
if not self.stream.is_playing():
self.start_tts()
else:
if self.dbg_log:
print(f"tts_play_sentence running sentence found, realtime playing")
buffer = BufferStream()
last_text = ""
if self.dbg_log:
print(f"ID: {sentence.id}")
print(f"EMOTION: {sentence.emotion}")
while not sentence.get_finished():
current_text = sentence.get_text()
if len(current_text) > len(last_text):
new_text = current_text[len(last_text):]
buffer.add(new_text)
if not self.stream.is_playing():
self.stream.feed(buffer.gen())
self.start_tts()
last_text = current_text
time.sleep(0.01)
if self.dbg_log:
print(" - feed finished")
buffer.stop()
while self.stream.is_playing():
time.sleep(0.01)
def tts_sentence_worker_thread(self):
while not self.stop_event.is_set():
sentence = self.sentence_queue.get_sentence()
if sentence:
emotion = sentence.emotion
if not emotion or emotion == "None":
emotion = "neutral"
emotion_file = emotion + ".wav"
path = os.path.join(self.references_folder, emotion_file)
if os.path.exists(path):
if self.dbg_log:
print(f"Setting TTS Emotion path: {path}")
self.engine.set_cloning_reference(path)
else:
if self.dbg_log:
print(f"No emotion found for path: {path}")
path = os.path.join(self.references_folder, "neutral.wav")
if os.path.exists(path):
if self.dbg_log:
print(f"Setting neutral: {path}")
self.engine.set_cloning_reference(path)
else:
if self.dbg_log:
print(f"CANT FIND EMOTIONS")
if self.dbg_log:
print(f"TTS found a sentence, running: {sentence.get_finished()}")
print(f" - finished: {sentence.get_finished()}")
print(f" - retrieved: {sentence.retrieved}")
print(f" - popped: {sentence.popped}")
print(f" - id: {sentence.id}")
self.tts_play_sentence(sentence)
time.sleep(0.01)
def start_threads(self):
self.tts_sentence_thread = threading.Thread(target=self.tts_sentence_worker_thread)
self.tts_sentence_thread.daemon = True
self.tts_sentence_thread.start()
self.tts_play_thread = threading.Thread(target=self.tts_play_worker_thread)
self.tts_play_thread.daemon = True
self.tts_play_thread.start()
def add_text(self, text):
self.sentence_queue.add_text(text)
def add_emotion(self, emotion):
self.sentence_queue.add_emotion(emotion)
def finish_current_sentence(self):
self.sentence_queue.finish_current_sentence()
def is_empty(self):
return self.sentence_queue.is_empty()
def is_playing(self):
return self.stream.is_playing()
def shutdown_pyaudio(self):
if self.pystream is not None:
self.pystream.stop_stream()
self.pystream.close()
if self.pyaudio_instance is not None:
self.pyaudio_instance.terminate()
def shutdown(self):
self.stop_event.set()
print("Waiting for sentence thread finished")
self.tts_sentence_thread.join()
print("Waiting for play thread finished")
self.tts_play_thread.join()
self.engine.shutdown()