调用百度语音识别接口实现语音识别与语音合成

目标：

1.学习使用百度AI开放平台进行语音识别与语音合成

       百度AI有两种开发方式：REST, SDK;    本文使用的是REST api+python进行开发

        文档地址：https://ai.baidu.com/docs#/ASR-Guide/top

2.将学到的内容：JSON数据，HTTP API调用，base64文件编码与解码

3.生成exe格式可执行文件

  pyinstaller -F xxx.py 参考文档 https://blog.csdn.net/qq_35203425/article/details/78568141

4.python-Tkinter图形界面开发

问题点：

1.界面上的模块不会随窗口放大产生变化

2.未添加异常捕获

界面：

代码：

from tkinter import *
from tkinter import ttk
from tkinter.filedialog import askopenfilename,askdirectory
import json
import requests
import base64
from urllib.parse import urlencode,quote_plus
import os

class Audio():
    def __init__(self):
        #参数初始化化
        # self.width,self.height,self.xoffset,self.yoffset=800,400,350,200  #定义窗口大小
        self.myWindow = Tk()
        self.path =  StringVar()          #搜索路径变量
        self.path1=StringVar()            #保存路径变量
        self.audio_formats = StringVar() #定义音频识别格式

        self.API_Key = 'AN6TAou6hNd5UEMxBGxUFXGC'             #百度AI秘钥
        self.Secret_Key = 'rTb6zkjVWEnuXoEVyv6cSaiblzNfxP6l'
        self.url_1 = 'https://aip.baidubce.com/oauth/2.0/token?grant_type=' \
                     'client_credentials&client_id=AN6TAou6hNd5UEMxBGxUFXGC&client_secret=rTb6zkjVWEnuXoEVyv6cSaiblzNfxP6l& '
        self.url_2='http://vop.baidu.com/server_api'
        self.url_3 = 'http://tsn.baidu.com/text2audio'
    def selectPath(self):
        # 路径搜索
        self.name = askopenfilename()
        self.path.set(self.name)
    def savepath(self):
        self.name1=askdirectory()
        self.path1.set(self.name1)

    def audio_recognition(self):
        #语言识别模块
        audio_path=self.path.get()
        audio_formats=self.audio_formats.get()
        if len(audio_path) == 0 or len(audio_formats) == 0:
            self.text.insert(INSERT, '未选定识别文件')
        else:
            response = requests.get(url=self.url_1)
            self.token = response.json()['access_token']
            # 打开文件
            with open(audio_path,'rb') as f:
                audio=f.read()
            # 识别文件
            length = len(audio)
            speech = base64.b64encode(audio)
            speech = str(speech, 'utf-8')
            params = {
                "format": audio_formats,
                "rate": 16000,

                "channel": '1',
                "token": self.token,
                "cuid": 'D8-F2-CA-06-B4-59',
                'len': length,
                "speech": speech}
            post_data = json.dumps(params, sort_keys=False)
            text = requests.post(self.url_2, post_data.encode('utf-8'))
            # 输出文件内容
            if text.json()['err_msg']=='success.':
                self.text.insert(INSERT,text.json()[ 'result'])
            else:
                self.text.insert(INSERT,text.json()['err_msg']+text.json()['err_no'])

    def audi_composite(self):
        #语言合成模块
        response = requests.get(url=self.url_1)
        token = response.json()['access_token']
        string=self.text1.get(1.0,2.0)
        tex = quote_plus(string)

        params = {'tok': token, 'tex': tex, 'per': 1, 'spd': 5, 'pit': 5, 'vol': 5, 'aue': 3,
                  'cuid': 'D8-F2-CA-06-B4-59',
                  'lan': 'zh', 'ctp': 1}
        data = urlencode(params)
        audio = requests.post(self.url_3, data.encode('utf-8'))
        if len(self.path1.get())!=0:
            path=os.path.join(self.path1.get(),'audio.mp3')

            with open(path, 'wb') as f:
                f.write(audio.content)

    def interface(self):
        # 制作界面

        self.myWindow.title('语音演示程序')  # 修改窗口标题
        self.myWindow.geometry("800x400+350+200")  # 修改窗口大小 width*height+xoffset+yoffset
        # 创建两个标签容器
        frm_l = LabelFrame(self.myWindow, text='语言识别', width=380, height=380)
        frm_l.grid(row=0, column=0, ipadx=20, ipady=10, padx=10, pady=10)
        frm_r = LabelFrame(self.myWindow, text='语言合成', width=380, height=380)
        frm_r.grid(row=0, column=1, padx=10, pady=10)

        entry1 = Entry(frm_l, textvariable=self.path, width=35).grid(row=0, column=0, rowspan=2, columnspan=4, sticky='EW', pady=5, padx=5)
        button1 = Button(frm_l, text='选择文件', command=self.selectPath, width=10, height=1).grid(row=0, column=4, sticky='EW',pady=5, padx=5)

        label_1 = Label(frm_l, text="音频格式").grid(row=2, column=0, sticky='EW', pady=5, padx=5)
        com = ttk.Combobox(frm_l, textvariable=self.audio_formats, value=['PCM', 'WAV', 'AMR'], width=10, height=1)
        com.current(0)
        com.grid(row=2, column=1, sticky='EW', pady=5, padx=5)
        button2 = Button(frm_l, text='开始识别', command=self.audio_recognition, width=10, height=1).grid(row=2, column=4,sticky='EW', pady=5,padx=5)

        label_2 = Label(frm_l, text='识别结果').grid(row=4, column=0, sticky='EW', pady=5, padx=5)

        self.text = Text(frm_l, width=40, height=15)
        self.text.grid(row=5, column=0, rowspan=2, columnspan=5, sticky='EW', pady=5, padx=5)

        # 语音合成模块

        self.text1 = Text(frm_r, width=50, height=19)
        self.text1.grid(row=0, column=0, rowspan=2, columnspan=4, sticky='EW', pady=5, padx=5)


        entry1 = Entry(frm_r, textvariable=self.path1).grid(row=2, column=0, rowspan=1, columnspan=3,
                                                                     sticky='EW', pady=5, padx=5)
        button4=Button(frm_r, text='保存路径',command=self.savepath, width=10, height=1).grid(row=2, column=3, sticky='EW',pady=5, padx=5)

        button3 = Button(frm_r, text='开始合成',command=self.audi_composite, width=10, height=1)
        button3.grid(row=3, column=3, sticky='EW', pady=5, padx=5)

        # 启动主窗口的消息循环
        self.myWindow.mainloop()

if __name__ == '__main__':
    audio=Audio()
    audio.interface()