sunli
/
SmartTrain-Speech
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Browse Source

初始化

master
孙莉 3 months ago
commit
38cb125a80
6 changed files with 451 additions and 0 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 2
      build.sh
  2. 203
      conf/preload.json
  3. 76
      conf/surname.json
  4. 2
      readme.txt
  5. 2
      requirements.txt
  6. 166
      ttsmain.py

2
build.sh
View File

@ -0,0 +1,2 @@
#!/bin/bash
nuitka3 --standalone --onefile --jobs=8 --include-module=cProfile,wave,cmath --show-memory --show-progress --include-package=httpx,PIL,sqlite3 --output-dir=build --include-data-dir=conf=conf --nofollow-import-to=numpy,os,time,matplotlib,scipy,nltk,datasets,huggingface_hub,fastapi,paddlespeech,paddlenlp,paddleslim,panda,numba,pygments,loguru,cffi ttsmain.py & >>packlog.txt

203
conf/preload.json
View File

@ -0,0 +1,203 @@
[
{
"name": "ready",
"content": "开始！"
},
{
"name": "start",
"content": "开始"
},
{
"name": "add_person",
"content": "请举手识别"
},
{
"name": "t-7",
"content": "仰卧起坐"
},
{
"name": "t-0",
"content": "俯卧撑"
},
{
"name": "t-2",
"content": "单杠引体向上"
},
{
"name": "t-3",
"content": "曲臂悬垂"
},
{
"name": "t-5",
"content": "蛇形跑"
},
{
"name": "t-4",
"content": "跑步"
},
{
"name": "t-1",
"content": "仰卧卷腹"
},
{
"name": "un_identification",
"content": "请伸直手臂"
},
{
"name": "startup",
"content": "欢迎使用中能鸿达智能训考终端！"
},
{
"name": "recognize",
"content": "已识别！"
},
{
"name": "re_identification",
"content": "重新识别！"
},
{
"name": "face_little",
"content": "人脸识别不清，请靠近！"
},
{
"name": "face_start",
"content": "人脸采集开始！"
},
{
"name": "face_stop",
"content": "人脸采集完成！"
},
{
"name": "face_timeout",
"content": "人脸采集超时！"
},
{
"name": "recognize_timeout",
"content": "识别超时,请重新举手识别"
},
{
"name": "readying",
"content": "请全体准备!"
},
{
"name": "ready_timeout",
"content": "准备超时"
},
{
"name": "train_stop",
"content": "已停止训考!"
},
{
"name" : "pass",
"content": "合格!"
},
{
"name": "fail",
"content": "不合格!"
},
{
"name": "task_stop",
"content": "训考结束!"
},
{
"name": "max_person",
"content": "当前识别人数已超过最大值!"
},
{
"name": "countdown",
"content": "倒计时，3,2,1,开始"
},
{
"name": "area1",
"content": "区域一"
},
{
"name": "area2",
"content": "区域二"
},
{
"name": "area3",
"content": "区域三"
},
{
"name": "area4",
"content": "区域四"
},
{
"name": "area5",
"content": "区域五"
},
{
"name": "bend_knees",
"content": "弯膝"
},
{
"name": "hip",
"content": "撅臀"
},
{
"name": "waist",
"content": "塌腰"
},
{
"name": "shoulder",
"content": "肩高于肘"
},
{
"name": "feet",
"content": "脚未并拢"
},
{
"name": "body",
"content": "侧身"
},
{
"name": "elbow",
"content": "手肘弯曲"
},
{
"name": "bend_knees_e",
"content": "弯膝不足"
},
{
"name": "feet_floor",
"content": "脚掌离地"
},
{
"name": "hand_floor",
"content": "手掌离地"
},
{
"name": "up_hip",
"content": "翘臀"
},
{
"name": "shoulder_floor",
"content" : "肩未着地"
},
{
"name": "hand_move",
"content": "手未移动"
},
{
"name": "body_up",
"content": "起身不足"
},
{
"name": "double_hand",
"content": "双手离肩"
},
{
"name" : "swing",
"content": "摆浪"
},
{
"name": "bar",
"content": "搭杠"
},
{
"name": "hand_stright",
"content": "手未伸直"
}
]

76
conf/surname.json
View File

@ -0,0 +1,76 @@
[{
"suname": "尉迟",
"speak": "<say-as pinyin='yu4'>尉</say-as>迟"
}, {
"suname": "皇甫",
"speak": "皇<say-as pinyin='fu3'>甫</say-as>"
}, {
"suname": "长孙",
"speak": "<say-as pinyin='zhang3'>长</say-as>孙"
},
{
"suname": "仇",
"speak": "<say-as pinyin='qiu2'>仇</say-as>"
},
{
"suname": "朴",
"speak": "<say-as pinyin='piao2'>朴</say-as>"
}, {
"suname": "单",
"speak": "<say-as pinyin='shan4'>单</say-as>"
}, {
"suname": "解",
"speak": "<say-as pinyin='xie4'>解</say-as>"
}, {
"suname": "区",
"speak": "<say-as pinyin='ou1'>区</say-as>"
}, {
"suname": "查",
"speak": "<say-as pinyin='zha1'>查</say-as>"
}, {
"suname": "繁",
"speak": "<say-as pinyin='po2'>繁</say-as>"
}, {
"suname": "瞿",
"speak": "<say-as pinyin='qu2'>瞿</say-as>"
}, {
"suname": "覃",
"speak": "<say-as pinyin='qin2'>覃</say-as>"
}, {
"suname": "曾",
"speak": "<say-as pinyin='zeng1'>曾</say-as>"
}, {
"suname": "员",
"speak": "<say-as pinyin='yun4'>员</say-as>"
}, {
"suname": "能",
"speak": "<say-as pinyin='nai4'>能</say-as>"
}, {
"suname": "阚",
"speak": "<say-as pinyin='kan4'>阚</say-as>"
}, {
"suname": "都",
"speak": "<say-as pinyin='du1'>都</say-as>"
}, {
"suname": "乜",
"speak": "<say-as pinyin='nie4'>乜</say-as>"
}, {
"suname": "缪",
"speak": "<say-as pinyin='miao4'>缪</say-as>"
}, {
"suname": "句",
"speak": "<say-as pinyin='gou1'>句</say-as>"
}, {
"suname": "阿",
"speak": "<say-as pinyin='e1'>阿</say-as>"
}, {
"suname": "盖",
"speak": "<say-as pinyin='ge3'>盖</say-as>"
}, {
"suname": "乐",
"speak": "<say-as pinyin='yue4'>乐</say-as>"
}, {
"suname": "种",
"speak": "<say-as pinyin='chong2'>种</say-as>"
}
]

2
readme.txt
View File

@ -0,0 +1,2 @@
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
生成命令：

2
requirements.txt
View File

@ -0,0 +1,2 @@
paddlepaddle-gpu==2.5.1
paddlespeech==1.4.1

166
ttsmain.py
View File

@ -0,0 +1,166 @@
# 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)
Write Preview
Loading…
Cancel
Save