百度语音API使用（python实现之二）

昨天是原始PCM数据上传，响应太慢，profiling一番，发现是curl耗时，估计是音频数据上传耗时

于是改用speex压缩编码格式

#encoding=utf-8

import os
import time
import urllib2, pycurl
import base64
import json

TOKEN_PATH = "/home/wang/.speech.token"

## get access token by api key & secret key
def has_token():
    try:
        stat_info = os.stat(TOKEN_PATH)
    except OSError:
        return False
    if stat_info.st_size < 10: #invalid if too small
        return False
    db_ctime = stat_info.st_ctime
    create_date = time.strftime("%m", time.localtime(db_ctime))
    current_date = time.strftime("%m", time.localtime(time.time()))
    if current_date != create_date:
        return False   #old beyond 1 day, need update
    else:
        return True

def get_token():
    if has_token():
        fp = open(TOKEN_PATH, "r")
        token = fp.readline().rstrip("
")
        fp.close()
        return token
    apiKey = "FzxxxxxwY0SS"
    secretKey = "66xx5axxxb882a"

    auth_url = "https://openapi.baidu.com/oauth/2.0/token?grant_type=client_credentials&client_id=" + apiKey + "&client_secret=" + secretKey;

    res = urllib2.urlopen(auth_url)
    json_data = res.read()
    token = json.loads(json_data)["access_token"]
    fp = open(TOKEN_PATH, "w")
    fp.write(token)
    fp.close()
    return token

def dump_res(buf):
    fp = open("log_spx.log", "w")
    fp.write(buf)
    fp.close()


## post audio to server
def use_cloud(token):
    fp = open("test.spx", "rb")
    fp.seek(os.SEEK_END)
    f_len = fp.tell()
    audio_data = fp.read(f_len)

    cuid = "acxxxxxx677" #my xiaomi phone MAC
    srv_url = "http://vop.baidu.com/server_api" + "?cuid=" + cuid + "&token=" + token
    http_header = [
        "Content-Type: audio/speex; rate=8000",
        "Content-Length: %d" % f_len
    ]

    c = pycurl.Curl()
    c.setopt(pycurl.URL, str(srv_url)) #curl doesn"t support unicode
    #c.setopt(c.RETURNTRANSFER, 1)
    c.setopt(c.HTTPHEADER, http_header)   #must be list, not dict
    c.setopt(c.POST, 1)
    c.setopt(c.CONNECTTIMEOUT, 30)
    c.setopt(c.TIMEOUT, 30)
    c.setopt(c.WRITEFUNCTION, dump_res)
    c.setopt(c.POSTFIELDS, audio_data)
    c.setopt(c.POSTFIELDSIZE, f_len)
    c.perform() #pycurl.perform() has no return val

if __name__ == "__main__":
    token = get_token()
    use_cloud(token)

但是百度“识别失败”，考虑到我的麦克风太差，换用样例代码中的test.pcm压成speex格式再试，仍然“识别错误”，求教！

{"err_msg":"recognition error.","err_no":3301,"sn":"320487617001432001015"}