百度英语中文语音识别为文字服务 python demo【使用requests重写官方demo】

# -*- coding: utf-8 -*-

# @Time : 2021/4/24 20:50

# @File : baidu_voice_service.py

# @Author : Rocky C@www.30daydo.com



import os

import requests

import sys

import pickle

sys.path.append('..')

from config import API_KEY,SECRET_KEY

from base64 import b64encode

from pathlib import PurePath





BASE = PurePath(__file__).parent



# 需要识别的文件

AUDIO_FILE = r'C:\OtherGit\speech-demo\rest-api-asr\python\audio\2.m4a'  # 只支持 pcm/wav/amr 格式，极速版额外支持m4a 格式

# 文件格式

FORMAT = AUDIO_FILE[-3:]  # 文件后缀只支持 pcm/wav/amr 格式，极速版额外支持m4a 格式



CUID = '24057753'

# 采样率

RATE = 16000  # 固定值





ASR_URL = 'http://vop.baidu.com/server_api'



#测试自训练平台需要打开以下信息， 自训练平台模型上线后，您会看见 第二步：“”获取专属模型参数pid:8001，modelid:1234”，按照这个信息获取 dev_pid=8001，lm_id=1234

'''

http://vop.baidu.com/server_api

1537	普通话(纯中文识别)	输入法模型	有标点	支持自定义词库

1737	英语	英语模型	无标点	不支持自定义词库

1637	粤语	粤语模型	有标点	不支持自定义词库

1837	四川话	四川话模型	有标点	不支持自定义词库

1936	普通话远场

'''

DEV_PID = 1737



SCOPE = 'brain_enhanced_asr'  # 有此scope表示有asr能力，没有请在网页里开通极速版





class DemoError(Exception):

    pass





TOKEN_URL = 'http://openapi.baidu.com/oauth/2.0/token'



def fetch_token():



    params = {'grant_type': 'client_credentials',

              'client_id': API_KEY,

              'client_secret': SECRET_KEY}

    r = requests.post(

        url=TOKEN_URL,

        data=params

    )



    result = r.json()

    if ('access_token' in result.keys() and 'scope' in result.keys()):

        if SCOPE and (not SCOPE in result['scope'].split(' ')):  # SCOPE = False 忽略检查

            raise DemoError('scope is not correct')



        return result['access_token']



    else:

        raise DemoError('MAYBE API_KEY or SECRET_KEY not correct: access_token or scope not found in token response')





"""  TOKEN end """



def dump_token(token):

    with open(os.path.join(BASE,'token.pkl'),'wb') as fp:

        pickle.dump({'token':token},fp)



def load_token(filename):



    if not os.path.exists(filename):

        token=fetch_token()

        dump_token(token)

        return token

    else:

        with open(filename,'rb') as fp:

            token = pickle.load(fp)

            return token['token']



def recognize_service(token,filename):



    with open(filename, 'rb') as speech_file:

        speech_data = speech_file.read()



    length = len(speech_data)

    if length == 0:

        raise DemoError('file %s length read 0 bytes' % AUDIO_FILE)



    b64_data = b64encode(speech_data)

    params = {'cuid': CUID, 'token': token, 'dev_pid': DEV_PID,'speech':b64_data,'len':length,'format':FORMAT,'rate':RATE,'channel':1}



    headers = {

        'Content-Type':'application/json',

    }

    r = requests.post(url=ASR_URL,json=params,headers=headers)

    return r.json()





def main():

    filename = 'token.pkl'

    token = load_token(filename)

    result = recognize_service(token,AUDIO_FILE)

    print(result['result'])



if __name__ == '__main__':

    main()