python 语音websocket,Azure语音到文本WebSocket[Python]

我面临的问题是语音.config头在打开事件监听器，它显示了我在下面捕捉到的错误，似乎web套接字无法成功构建。我遵循Micrsoft的官方指南，在建立连接后是否可以发送头？

非常感谢你的帮助！

这是我的代码：import speech_recognition as sr

import json

import requests

import ssl

import urllib3

import xmltodict

import time

import websocket

import thread

import uuid

from googletrans import Translator

from urllib import urlencode

from auth import AzureAuthClient

import datetime

import struct

import pyaudio

import sys

import StringIO

import platform

import os

def get_wave_header(frame_rate):

"""

Generate WAV header that precedes actual audio data sent to the speech translation service.

:param frame_rate: Sampling frequency (8000 for 8kHz or 16000 for 16kHz).

:return: binary string

"""

if frame_rate not in [8000, 16000]:

raise ValueError("Sampling frequency, frame_rate, should be 8000 or 16000.")

nchannels = 1

bytes_per_sample = 2

output = StringIO.StringIO()

output.write('RIFF')

output.write(struct.pack('

output.write('WAVE')

output.write('fmt ')

output.write(struct.pack('

output.write(struct.pack('

output.write(struct.pack('

output.write(struct.pack('

output.write(struct.pack('

output.write(struct.pack('

output.write(struct.pack('

output.write(struct.pack('

output.write('data')

output.write(struct.pack('

data = output.getvalue()

output.close()

return data

def get_wave_empty():

output = StringIO.StringIO()

# for num in range(1, 801):

for num in range(1, 41):

output.write(struct.pack('

data = output.getvalue()

output.close()

return data

def on_open(ws):

timestap = datetime.datetime.utcnow().isoformat()+'Z'

speechConfigBody = {

"context": {

"system": {

"version": "2.0.12341",

},

"os": {

"platform": "Linux",

"name": "Debian",

"version": "2.14324324"

},

"device": {

"manufacturer": "Contoso",

"model": "Fabrikan",

"version": "7.341"

}

},

}

speechConfigHeaders = ['Path: speech.config\n', 'X-Timestamp: '+timestap+'\n', 'Content-Type: application/json; charset=utf-8\n']

print(speechConfigHeaders[0].encode('utf8'))

print(speechConfigHeaders[0].encode('utf8')+speechConfigHeaders[1].encode('utf8')+speechConfigHeaders[2].encode('utf8')+json.dumps(speechConfigBody))

ws.send(speechConfigHeaders[0].encode('utf8')+speechConfigHeaders[1].encode('utf8')+speechConfigHeaders[2].encode('utf8')+json.dumps(speechConfigBody))

#print("Web Socket is opened")

data = get_wave_header(16000)

#ws.send(data, websocket.ABNF.OPCODE_BINARY)

def run(*args):

print("Websocket Running")

while True:

r = sr.Recognizer()

with sr.Microphone() as source:

print("Say something!")

audio = r.listen(source)

print("After Say")

print(audio)

#ws.send(audio, websocket.ABNF.OPCODE_BINARY)

ws.close()

thread.start_new_thread(run, ())

def on_data(ws, message, message_type, fin):

print("on_data")

print(message)

def on_error(ws, error):

print("Websocket Error")

print(error)

def on_close(ws):

print("Web Socket is closed")

if __name__ == "__main__":

urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

translator1 = Translator()

websocket.enableTrace(True)

#print(translator1.translate('이 문장은 한글로 쓰여졌습니다.').text)

# recognize speech using Microsoft Bing Voice Recognition

BING_KEY = "confidential key" # Microsoft Bing Voice Recognition API keys 32-character lowercase hexadecimal strings

auth_client = AzureAuthClient(BING_KEY)

# post_auth_url = "https://api.cognitive.microsoft.com/sts/v1.0/issueToken"

# post_auth_header = {"Ocp-Apim-Subscription-Key": BING_KEY}

# auth_access_code = requests.post(post_auth_url, headers=post_auth_header)

# print(auth_access_code.content)

language = "en-US"

client_trace_id = str(uuid.uuid4().hex)

print("client_trace_id: "+client_trace_id)

timestap = datetime.datetime.utcnow().isoformat()+'Z'

print(timestap)

azureHost = "wss://speech.platform.bing.com/speech/recognition/conversation/cognitiveservices/v1?language=en-US"

print("Ready to start web Socket")

ws_client = websocket.WebSocketApp(

azureHost,

header=[

'Authorization: Bearer '+ auth_client.get_access_token(),

'X-ConnectionId: ' + client_trace_id

],

on_open = on_open,

on_data=on_data,

on_error=on_error,

on_close=on_close

)

#ws_client.on_open = on_open

ws_client.run_forever(sslopt={"cert_reqs": ssl.CERT_NONE})