ASRT_SpeechRecognition/asrserver.py

146 lines
3.2 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
@author: nl8590687
语音识别API的HTTP服务器程序
"""
import http.server
import urllib
import keras
2018-07-27 14:31:48 +08:00
from SpeechModel251 import ModelSpeech
2018-04-26 12:10:34 +08:00
from LanguageModel import ModelLanguage
2019-01-09 16:37:10 +08:00
datapath = './'
2018-05-11 16:56:59 +08:00
modelpath = 'model_speech/'
ms = ModelSpeech(datapath)
2019-01-19 11:27:57 +08:00
ms.LoadModel(modelpath + 'm251/speech_model251_e_0_step_12000.model')
2018-05-11 16:56:59 +08:00
ml = ModelLanguage('model_language')
ml.LoadModel()
class TestHTTPHandle(http.server.BaseHTTPRequestHandler):
def setup(self):
self.request.settimeout(10)
http.server.BaseHTTPRequestHandler.setup(self)
def _set_response(self):
self.send_response(200)
self.send_header('Content-type', 'text/html')
self.end_headers()
def do_GET(self):
buf = 'ASRT_SpeechRecognition API'
self.protocal_version = 'HTTP/1.1'
self._set_response()
buf = bytes(buf,encoding="utf-8")
self.wfile.write(buf)
def do_POST(self):
'''
处理通过POST方式传递过来并接收的语音数据
通过语音模型和语言模型计算得到语音识别结果并返回
'''
path = self.path
print(path)
#获取post提交的数据
datas = self.rfile.read(int(self.headers['content-length']))
#datas = urllib.unquote(datas).decode("utf-8", 'ignore')
datas = datas.decode('utf-8')
datas_split = datas.split('&')
token = ''
fs = 0
wavs = []
2018-05-08 11:35:18 +08:00
#type = 'wavfilebytes' # wavfilebytes or python-list
for line in datas_split:
[key, value]=line.split('=')
if('wavs' == key and '' != value):
wavs.append(int(value))
elif('fs' == key):
2018-04-26 12:10:34 +08:00
fs = int(value)
elif('token' == key ):
token = value
2018-05-08 11:35:18 +08:00
#elif('type' == key):
# type = value
2018-04-25 21:55:34 +08:00
else:
print(key, value)
if(token != 'qwertasd'):
buf = '403'
print(buf)
buf = bytes(buf,encoding="utf-8")
self.wfile.write(buf)
return
2018-05-08 11:35:18 +08:00
#if('python-list' == type):
if(len(wavs)>0):
r = self.recognize([wavs], fs)
else:
r = ''
2018-05-08 11:35:18 +08:00
#else:
# r = self.recognize_from_file('')
2018-04-25 21:55:34 +08:00
if(token == 'qwertasd'):
#buf = '成功\n'+'wavs:\n'+str(wavs)+'\nfs:\n'+str(fs)
2018-05-11 16:56:59 +08:00
buf = r
else:
buf = '403'
#print(datas)
self._set_response()
2018-04-25 21:55:34 +08:00
#buf = '<!DOCTYPE HTML> \n<html> \n<head>\n<title>Post page</title>\n</head> \n<body>Post Data:%s <br />Path:%s\n</body> \n</html>'%(datas,self.path)
print(buf)
2018-05-08 11:35:18 +08:00
buf = bytes(buf,encoding="utf-8")
2018-04-25 21:55:34 +08:00
self.wfile.write(buf)
2018-04-26 12:10:34 +08:00
def recognize(self, wavs, fs):
r=''
try:
r_speech = ms.RecognizeSpeech(wavs, fs)
print(r_speech)
str_pinyin = r_speech
r = ml.SpeechToText(str_pinyin)
except:
r=''
print('[*Message] Server raise a bug. ')
2018-04-25 21:55:34 +08:00
return r
pass
2018-05-08 11:35:18 +08:00
def recognize_from_file(self, filename):
pass
2019-01-09 16:37:10 +08:00
import socket
class HTTPServerV6(http.server.HTTPServer):
address_family = socket.AF_INET6
def start_server(ip, port):
2019-01-09 16:37:10 +08:00
if(':' in ip):
http_server = HTTPServerV6((ip, port), TestHTTPHandle)
else:
http_server = http.server.HTTPServer((ip, int(port)), TestHTTPHandle)
print('服务器已开启')
try:
http_server.serve_forever() #设置一直监听并接收请求
except KeyboardInterrupt:
pass
http_server.server_close()
print('HTTP server closed')
if __name__ == '__main__':
2019-01-09 16:37:10 +08:00
start_server('', 20000) # For IPv4 Network Only
#start_server('::', 20000) # For IPv6 Network