Blame view
egs/vystadial_cz/online_demo/live-demo.py
4.88 KB
8dcb6dfcb
first commit
|
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 |
#!/usr/bin/env python
# encoding: utf-8
# Copyright (c) 2013, Ondrej Platek, Ufal MFF UK <oplatek@ufal.mff.cuni.cz>
#
# 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
#
# THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
# WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
# MERCHANTABLITY OR NON-INFRINGEMENT.
# See the Apache 2 License for the specific language governing permissions and
# limitations under the License. #
from __future__ import unicode_literals
from __future__ import print_function
import pyaudio
from kaldi.decoders import PyOnlineLatgenRecogniser
from kaldi.utils import wst2dict, lattice_to_nbest
import sys
import time
import select
import tty
import termios
import wave
CHANNELS, RATE, FORMAT = 1, 16000, pyaudio.paInt16
class LiveDemo(object):
def __init__(self, audio_batch_size, wst, dec_args):
self.batch_size = audio_batch_size
self.wst = wst
self.args = dec_args
self.d = PyOnlineLatgenRecogniser()
self.pin, self.stream = None, None
self.frames = []
self.utt_frames, self.new_frames = 0, 0
self.utt_end, self.dialog_end = False, False
def setup(self):
self.d.reset()
self.d.setup(argv)
self.pin = pyaudio.PyAudio()
self.stream = self.pin.open(format=FORMAT, channels=CHANNELS,
rate=RATE, input=True, frames_per_buffer=self.batch_size,
stream_callback=self.get_audio_callback())
self.utt_frames, self.new_frames = 0, 0
self.utt_end, self.dialog_end = False, False
self.frames = []
def tear_down(self):
if self.stream is not None:
self.stream.stop_stream()
self.stream.close()
if self.pin is not None:
self.pin.terminate()
p, stream = None, None
self.frames = []
def get_audio_callback(self):
def frame_in(in_data, frame_count, time_info, status):
self.d.frame_in(in_data)
self.frames.append(in_data)
return in_data, pyaudio.paContinue
return frame_in
def _user_control(self):
'''Simply stupid sollution how to control state of recogniser.'''
self.utt_end, self.dialog_end = False, False
old_settings = termios.tcgetattr(sys.stdin)
try:
tty.setcbreak(sys.stdin.fileno())
# if is data on input
while (select.select([sys.stdin], [], [], 1) == ([sys.stdin], [], [])):
c = sys.stdin.read(1)
if c == 'u':
print('
Marked end of utterance
')
self.utt_end = True
elif c == 'c':
self.dialog_end = True
print('
Marked end of dialogue
')
finally:
termios.tcsetattr(sys.stdin, termios.TCSADRAIN, old_settings)
print("""Chunks: %d ; Utterance %d ; end %d : press 'u'
For terminating press 'c'
""" % (len(self.frames), self.utt_frames, self.utt_end))
def run(self):
while True:
time.sleep(0.1)
self._user_control()
new_frames = self.d.decode(max_frames=10)
while new_frames > 0:
self.utt_frames += new_frames
new_frames = self.d.decode(max_frames=10)
if self.utt_end or self.dialog_end:
start = time.time()
self.d.prune_final()
prob, lat = self.d.get_lattice()
# lat.write('live-demo-recorded.fst')
nbest = lattice_to_nbest(lat, n=10)
if nbest:
best_prob, best_path = nbest[0]
decoded = ' '.join([wst[w] for w in best_path])
else:
decoded = 'Empty hypothesis'
print("%s secs, frames: %d, prob: %f, %s " % (
str(time.time() - start), self.utt_frames, prob, decoded))
self.utt_frames = 0
self.d.reset(keep_buffer_data=False)
if self.dialog_end:
self.save_wav()
break
def save_wav(self):
wf = wave.open('live-demo-record.wav', 'wb')
wf.setnchannels(CHANNELS)
wf.setframerate(RATE)
wf.setsampwidth(self.pin.get_sample_size(FORMAT))
wf.writeframes(b''.join(self.frames))
wf.close()
if __name__ == '__main__':
audio_batch_size, wst_path = int(sys.argv[1]), sys.argv[2]
argv = sys.argv[3:]
print('Python args: %s' % str(sys.argv), file=sys.stderr)
wst = wst2dict(wst_path)
demo = LiveDemo(audio_batch_size, wst, argv)
demo.setup()
demo.run()
|