File size: 4,815 Bytes
a203d35 |
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 |
from flask import Flask
from flask import Flask, request, jsonify
import yaml
from vietTTS.hifigan.mel2wave import mel2wave
from vietTTS.nat.text2mel import text2mel
from vietTTS.synthesizer import nat_normalize_text
import numpy as np
import gradio as gr
import re
from vietnam_number import n2w
from vietnam_number import n2w_single
from synthesize import synthesizer
import noisereduce as nr
import os
import scipy.io.wavfile as wavf
from scipy.io import wavfile
app = Flask(__name__)
@app.route('/hello/', methods=['GET', 'POST'])
def welcome():
return "Hello World!"
@app.route('/tts',methods = ['POST'])
def tts():
text = request.json['data'][0]
model = request.json['data'][1]
duration = request.json['data'][2]
if(model == "Tacotron2"):
return using_tacotron(text)
else:
return '1'
def text_to_speech(text,stop_duration):
print("starting")
# prevent too long text
if len(text) > 500:
text = text[:500]
# text_to_speech_tacotron(text)
# stop_duration_float = float(stop_duration_text)
text = clean_text(text)
text = nat_normalize_text(text)
mel = text2mel(
text,
"lexicon.txt",
stop_duration,
"acoustic_latest_ckpt.pickle",
"duration_latest_ckpt.pickle",
)
wave = mel2wave(mel, "config.json", "hk_hifi.pickle")
return (wave * (2**15)).astype(np.int16)
def text_to_speech_tacotron(text):
print("starting")
# prevent too long text
if len(text) > 500:
text = text[:500]
wav = synthesizer.tts(text)
output = './out.wav'
synthesizer.save_wav(wav,output)
return output
def using_viettts(text,stop_duration):
y = text_to_speech(text,stop_duration)
fs = 16000
output = './out.wav'
output_denoise = './output_denoise.wav'
wavf.write(output, fs, y)
rate, data = wavfile.read(output)
# perform noise reduction
reduced_noise = nr.reduce_noise(y=data, sr=rate)
wavfile.write(output_denoise, rate, reduced_noise)
return (output,output_denoise)
def using_tacotron(text):
y = text_to_speech_tacotron(text)
output_denoise = "./output_denoise.wav"
rate, data = wavfile.read(y)
# perform noise reduction
reduced_noise = nr.reduce_noise(y=data, sr=rate)
wavfile.write(output_denoise, rate, reduced_noise)
return (y,output_denoise)
def clean_text(test_string):
list_word = test_string.split()
# print(list_word)
regex = r"\d{2}(?P<sep>[-/])\d{1,2}(?P=sep)\d{4}"
for word in list_word :
try:
# print(word)
searchbox_result = re.match(regex, word)
day = searchbox_result.group(0)
day2 = day
day2 = day2.replace('/',' ').replace('-',' ')
list_date = day2.split(' ')
date_result = 'Ngày ' + n2w(list_date[0]) + ' tháng ' + n2w(list_date[1].replace('0','') if list_date[1].startswith('0') else list_date[1]) + ' năm ' + n2w(list_date[2])
# print(date_result)
test_string = test_string.replace(word, date_result)
except AttributeError:
# print(word)
# print("can't make a group")
continue
# print(test_string)
regex2 = r"\d{2}(?P<sep>[-/])\d{1,2}"
for word in list_word :
try:
# print(word)
searchbox_result = re.match(regex2, word)
day = searchbox_result.group(0)
day2 = day
day2 = day2.replace('/',' ').replace('-',' ')
list_date = day2.split(' ')
date_result = 'Ngày ' + n2w(list_date[0]) + ' tháng ' + n2w(list_date[1].replace('0','') if list_date[1].startswith('0') else list_date[1])
# print(date_result)
test_string = test_string.replace(word, date_result)
except AttributeError:
# print(word)
# print("can't make a group")
continue
# print(test_string)
regex3 = r"\d{1,2}(?P<sep>[h:])\d{1,2}"
for word in list_word :
try:
# print(word)
searchbox_result = re.match(regex3, word)
day = searchbox_result.group(0)
day2 = day
day2 = day2.replace('h',' ').replace(':',' ')
list_date = day2.split(' ')
date_result = n2w(list_date[0]) + ' giờ ' + n2w(list_date[1].replace('0','') if list_date[1].startswith('0') else list_date[1]) + ' phút '
# print(date_result)
test_string = test_string.replace(word, date_result)
except AttributeError:
# print(word)
# print("can't make a group")
continue
print(test_string)
for word in list_word :
try:
if word.isdigit() :
# print(word)
text_result = n2w_single(word)
# print(text_result)
test_string = test_string.replace(word, text_result, 1)
except AttributeError:
# print(word)
print("can't make a group")
continue
return test_string
if __name__ == '__main__':
app.run(host='0.0.0.0', port=105) |