You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
166 lines
6.6 KiB
166 lines
6.6 KiB
# coding:utf-8
|
|
|
|
import json
|
|
from paddlespeech.cli.tts.infer import TTSExecutor
|
|
import numpy as np
|
|
from scipy.io import wavfile
|
|
from pathlib import Path
|
|
import sys
|
|
import os
|
|
import re
|
|
from flask import Flask, request, jsonify
|
|
from concurrent.futures import ThreadPoolExecutor
|
|
|
|
executor =ThreadPoolExecutor(1)
|
|
app = Flask(__name__)
|
|
current_file_path = os.fspath(Path(__file__).resolve().parent)
|
|
save_file_path = f"{os.path.join(os.fspath(Path(__file__).resolve().parent.parent),'data','preload')}"
|
|
tts=TTSExecutor()
|
|
@app.route('/buildvoice', methods=['POST'])
|
|
def buildvoice():
|
|
# 获取 JSON 请求体
|
|
data = request.get_json()
|
|
|
|
# 从请求体中提取参数
|
|
voice_text = data.get('voice_text')
|
|
file_name = data.get('file_name')
|
|
# 检查参数是否存在
|
|
if not voice_text or not file_name:
|
|
response = {
|
|
'status': 'error',
|
|
'message': 'Missing voice_text or file_name',
|
|
'data': None
|
|
}
|
|
return jsonify(response), 400
|
|
|
|
current_directory = f"{os.path.join(current_file_path, 'conf', 'surname.json')}"
|
|
try:
|
|
if '@' in voice_text:
|
|
audio_data="<speak>"
|
|
text=voice_text.split('@')
|
|
for txt in text:
|
|
if '*' in txt:
|
|
audio_data += f"<say-as pinyin='{txt.split('*')[1]}'>{txt.split('*')[0]}</say-as>"
|
|
else:
|
|
audio_data+=txt
|
|
audio_data+='</speak>'
|
|
else :
|
|
if not os.path.isfile(current_directory):
|
|
audio_data=voice_text
|
|
else:
|
|
try:
|
|
with open(current_directory, 'r', encoding='utf-8') as file:
|
|
preload_data = json.load(file)
|
|
except json.JSONDecodeError as e:
|
|
#print(f"预加载姓氏读音配置文件格式错误: {str(e)}")
|
|
response = {
|
|
'status': 'error',
|
|
'message': f"预加载姓氏读音配置文件格式错误: {str(e)}",
|
|
'data': None
|
|
}
|
|
return jsonify(response), 400
|
|
audio_data=voice_text
|
|
for item in preload_data:
|
|
if voice_text.startswith(item['suname']):
|
|
audio_data ="<speak>"+ re.sub("^" + re.escape(item['suname']), item['speak'], voice_text)+'</speak>'
|
|
break
|
|
save_voice(audio_data,f"{save_file_path}/{file_name}.wav")
|
|
# 响应内容
|
|
response = {
|
|
'status': 'success',
|
|
'message': 'voice file generated successfully'
|
|
}
|
|
return jsonify(response), 200
|
|
except json.JSONDecodeError as e:
|
|
response = {
|
|
'status': 'error',
|
|
'message': f"build voice file error: {str(e)}",
|
|
'data': None
|
|
}
|
|
return jsonify(response), 400
|
|
|
|
|
|
@app.route('/initvoice', methods=['POST'])
|
|
def run_initvoice():
|
|
executor.submit(initvoice)
|
|
response = {
|
|
'status': 'success',
|
|
'message': 'build init voice running.....'
|
|
}
|
|
return jsonify(response), 200
|
|
|
|
def initvoice():
|
|
print(f"=====build voice file start=====")
|
|
current_directory = f"{os.path.join(current_file_path, 'conf', 'preload.json')}"
|
|
if not os.path.exists(save_file_path):
|
|
os.makedirs(save_file_path)
|
|
if not os.path.isfile(current_directory):
|
|
print("预加载语音配置文件不存在！")
|
|
sys.exit(1)
|
|
try:
|
|
with open(current_directory, 'r', encoding='utf-8') as file:
|
|
preload_data = json.load(file)
|
|
except json.JSONDecodeError as e:
|
|
print(f"预加载语音配置文件格式错误: {str(e)}")
|
|
sys.exit(1)
|
|
for item in preload_data:
|
|
save_voice(item['content'], f"{save_file_path}/{item['name']}.wav")
|
|
for i in range(150):
|
|
save_voice(f'{i}', f"{save_file_path}/{i}.wav")
|
|
|
|
print(f"=====build voice file success=====")
|
|
|
|
def save_voice(audio_data,file_path):
|
|
tts(text=audio_data, output=f"temp.wav",device='cpu', spk_id=0, am='speedyspeech_csmsc')
|
|
# 将增强音量的音频写入新文件
|
|
sample_rate,data=wavfile.read("temp.wav")
|
|
new_data=data*2
|
|
new_data=np.clip(new_data,-32768,32767).astype(np.int16)
|
|
wavfile.write(file_path,sample_rate,new_data)
|
|
# if __name__ == '__main__':
|
|
# args = sys.argv
|
|
# current_directory = f"{os.path.join(current_file_path, 'conf', 'surname.json')}"
|
|
|
|
# print('生成多音字的格式,注明拼音的文字和其它文字使用@间隔，文字和拼音用*间隔，示例如：覃*qin2@海洋')
|
|
# if len(args)!=3:
|
|
# print("参数不合法，请检查:需要2个参数 第1个参数为语音文本 第2个参数为语音文件保存路径")
|
|
# sys.exit(-1)
|
|
# tts=TTSExecutor()
|
|
# if '@' in args[1]:
|
|
# audio_data="<speak>"
|
|
# text=args[1].split('@')
|
|
# for txt in text:
|
|
# if '*' in txt:
|
|
# audio_data += f"<say-as pinyin='{txt.split('*')[1]}'>{txt.split('*')[0]}</say-as>"
|
|
# else:
|
|
# audio_data+=txt
|
|
# audio_data+='</speak>'
|
|
# else :
|
|
# if not os.path.isfile(current_directory):
|
|
# audio_data=args[1]
|
|
# else:
|
|
# try:
|
|
# with open(current_directory, 'r', encoding='utf-8') as file:
|
|
# preload_data = json.load(file)
|
|
# except json.JSONDecodeError as e:
|
|
# print(f"预加载姓氏读音配置文件格式错误: {str(e)}")
|
|
# sys.exit(1)
|
|
# audio_data=args[1]
|
|
# for item in preload_data:
|
|
# if args[1].startswith(item['suname']):
|
|
# audio_data ="<speak>"+ re.sub("^" + re.escape(item['suname']), item['speak'], args[1])+'</speak>'
|
|
# break
|
|
# output_file = f"{os.path.join(current_file_path, 'files',args[2])}.wav"
|
|
# print('即将生成:{} 的语音文件,保存位置为{}.....'.format(audio_data,output_file))
|
|
# tts(text=audio_data, output=f"temp.wav",device='cpu', spk_id=0, am='speedyspeech_csmsc')
|
|
# # 将增强音量的音频写入新文件
|
|
# print('生成成功，增强音量中')
|
|
# sample_rate,data=wavfile.read("temp.wav")
|
|
# new_data=data*2
|
|
# new_data=np.clip(new_data,-32768,32767).astype(np.int16)
|
|
# wavfile.write(output_file,sample_rate,new_data)
|
|
# print('语音文件已保存{}'.format(output_file))
|
|
|
|
|
|
if __name__ == '__main__':
|
|
app.run(host='0.0.0.0',port=5000,debug=True)
|