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
|
# SPDX-License-Identifier: Apache-2.0
#
# Copyright (c) 2023 Malik Talha
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import json
import vosk
import wave
from agl_service_voiceagent.utils.common import generate_unique_uuid
# import the whisper model
# import whisper
# for whisper timeout feature
from concurrent.futures import ThreadPoolExecutor
import subprocess
class STTModel:
"""
STTModel is a class for speech-to-text (STT) recognition using the Vosk speech recognition library.
"""
def __init__(self, vosk_model_path,whisper_model_path,whisper_cpp_path,whisper_cpp_model_path,sample_rate=16000):
"""
Initialize the STTModel instance with the provided model and sample rate.
Args:
model_path (str): The path to the Vosk speech recognition model.
sample_rate (int, optional): The audio sample rate in Hz (default is 16000).
"""
self.sample_rate = sample_rate
self.vosk_model = vosk.Model(vosk_model_path)
self.recognizer = {}
self.chunk_size = 1024
# self.whisper_model = whisper.load_model(whisper_model_path)
self.whisper_cpp_path = whisper_cpp_path
self.whisper_cpp_model_path = whisper_cpp_model_path
def setup_vosk_recognizer(self):
"""
Set up a Vosk recognizer for a new session and return a unique identifier (UUID) for the session.
Returns:
str: A unique identifier (UUID) for the session.
"""
uuid = generate_unique_uuid(6)
self.recognizer[uuid] = vosk.KaldiRecognizer(self.vosk_model, self.sample_rate)
return uuid
def init_recognition(self, uuid, audio_data):
"""
Initialize the Vosk recognizer for a session with audio data.
Args:
uuid (str): The unique identifier (UUID) for the session.
audio_data (bytes): Audio data to process.
Returns:
bool: True if initialization was successful, False otherwise.
"""
return self.recognizer[uuid].AcceptWaveform(audio_data)
# Recognize speech using the Vosk recognizer
def recognize_using_vosk(self, uuid, partial=False):
"""
Recognize speech and return the result as a JSON object.
Args:
uuid (str): The unique identifier (UUID) for the session.
partial (bool, optional): If True, return partial recognition results (default is False).
Returns:
dict: A JSON object containing recognition results.
"""
self.recognizer[uuid].SetWords(True)
if partial:
result = json.loads(self.recognizer[uuid].PartialResult())
else:
result = json.loads(self.recognizer[uuid].Result())
self.recognizer[uuid].Reset()
return result
# Recognize speech using the whisper model
# def recognize_using_whisper(self,filename,language = None,timeout = 5,fp16=False):
# """
# Recognize speech and return the result as a JSON object.
# Args:
# filename (str): The path to the audio file.
# timeout (int, optional): The timeout for recognition (default is 5 seconds).
# fp16 (bool, optional): If True, use 16-bit floating point precision, (default is False) because cuda is not supported.
# language (str, optional): The language code for recognition (default is None).
# Returns:
# dict: A JSON object containing recognition results.
# """
# def transcribe_with_whisper():
# return self.whisper_model.transcribe(filename, language = language,fp16=fp16)
# with ThreadPoolExecutor() as executor:
# future = executor.submit(transcribe_with_whisper)
# try:
# return future.result(timeout=timeout)
# except TimeoutError:
# return {"error": "Transcription with Whisper exceeded the timeout."}
def recognize_using_whisper_cpp(self,filename):
command = self.whisper_cpp_path
arguments = ["-m", self.whisper_cpp_model_path, "-f", filename, "-l", "en","-nt"]
# Run the executable with the specified arguments
result = subprocess.run([command] + arguments, capture_output=True, text=True)
if result.returncode == 0:
result = result.stdout.replace('\n', ' ').strip()
return {"text": result}
else:
print("Error:\n", result.stderr)
return {"error": result.stderr}
def recognize_from_file(self, uuid, filename,stt_framework="vosk"):
"""
Recognize speech from an audio file and return the recognized text.
Args:
uuid (str): The unique identifier (UUID) for the session.
filename (str): The path to the audio file.
stt_model (str): The STT model to use for recognition (default is "vosk").
Returns:
str: The recognized text or error messages.
"""
if not os.path.exists(filename):
print(f"Audio file '{filename}' not found.")
return "FILE_NOT_FOUND"
wf = wave.open(filename, "rb")
if wf.getnchannels() != 1 or wf.getsampwidth() != 2 or wf.getcomptype() != "NONE":
print("Audio file must be WAV format mono PCM.")
return "FILE_FORMAT_INVALID"
# audio_data = wf.readframes(wf.getnframes())
# we need to perform chunking as target AGL system can't handle an entire audio file
audio_data = b""
while True:
chunk = wf.readframes(self.chunk_size)
if not chunk:
break # End of file reached
audio_data += chunk
if audio_data:
# Perform speech recognition using the specified STT model
if stt_framework == "vosk":
if self.init_recognition(uuid, audio_data):
result = self.recognize_using_vosk(uuid)
return result['text']
else:
result = self.recognize_using_vosk(uuid, partial=True)
return result['partial']
elif stt_framework == "whisper":
result = self.recognize_using_whisper_cpp(filename)
if 'error' in result:
print(result['error'])
# If Whisper times out, fall back to Vosk
if self.init_recognition(uuid, audio_data):
result = self.recognize_using_vosk(uuid)
return result['text']
else:
result = self.recognize_using_vosk(uuid, partial=True)
return result['partial']
else:
return result.get('text', '')
else:
print("Voice not recognized. Please speak again...")
return "VOICE_NOT_RECOGNIZED"
def cleanup_recognizer(self, uuid):
"""
Clean up and remove the Vosk recognizer for a session.
Args:
uuid (str): The unique identifier (UUID) for the session.
"""
del self.recognizer[uuid]
|
