From a2588dec1af02e106a910286c692354ca4fe626a Mon Sep 17 00:00:00 2001
From: nl8590687 <3210346136@qq.com>
Date: Thu, 7 Jun 2018 12:48:31 +0800
Subject: [PATCH] =?UTF-8?q?=E4=BF=AE=E5=A4=8D=E4=BA=86=E8=AF=AD=E8=A8=80?=
=?UTF-8?q?=E6=A8=A1=E5=9E=8B=E7=9A=84=E6=A6=82=E7=8E=87=E8=AE=A1=E7=AE=97?=
=?UTF-8?q?=E9=94=99=E8=AF=AF=E7=9A=84Bug=EF=BC=8C=E5=B9=B6=E6=8F=90?=
=?UTF-8?q?=E9=AB=98API=E6=9C=8D=E5=8A=A1=E5=99=A8=E7=9A=84=E5=81=A5?=
=?UTF-8?q?=E5=A3=AE=E6=80=A7?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
---
LanguageModel.py | 4 +++-
asrserver.py | 30 +++++++++++++++++++++++-------
2 files changed, 26 insertions(+), 8 deletions(-)
diff --git a/LanguageModel.py b/LanguageModel.py
index 48eba22..0b4774e 100644
--- a/LanguageModel.py
+++ b/LanguageModel.py
@@ -137,8 +137,9 @@ class ModelLanguage(): # 语音模型类
#print('tmp_words: ',tmp_words,tmp_words in self.model2)
if(tmp_words in self.model2): # 判断它们是不是再状态转移表里
#print(tmp_words,tmp_words in self.model2)
- tuple_word[1] = tuple_word[1] * float(self.model2[tmp_words]) / float(self.model1[tmp_words[-1]])
+ tuple_word[1] = tuple_word[1] * float(self.model2[tmp_words]) / float(self.model1[tmp_words[-2]])
# 核心！在当前概率上乘转移概率，公式化简后为第n-1和n个字出现的次数除以第n-1个字出现的次数
+ #print(self.model2[tmp_words],self.model1[tmp_words[-2]])
else:
tuple_word[1] = 0.0
continue
@@ -238,6 +239,7 @@ if(__name__=='__main__'):
#str_pinyin = ['wo3','qu4','a4','mei2','shi4','er2','la1']
#str_pinyin = ['wo3', 'men5', 'qun2', 'li3', 'xiong1', 'di4', 'jian4', 'mei4', 'dou1', 'zai4', 'shuo1']
#str_pinyin = ['su1', 'an1', 'ni3', 'sui4', 'li4', 'yun4', 'sui2', 'cong2', 'jiao4', 'ming2', 'tao2', 'qi3', 'yu2', 'peng2', 'ya4', 'yang4', 'chao1', 'dao3', 'jiang1', 'li3', 'yuan2', 'kang1', 'zhua1', 'zou3']
+ str_pinyin = ['da4', 'jia1', 'hao3']
#r = ml.decode(str_pinyin)
r=ml.SpeechToText(str_pinyin)
print('语音转文字结果：\n',r)
diff --git a/asrserver.py b/asrserver.py
index 180388c..c3fea57 100644
--- a/asrserver.py
+++ b/asrserver.py
@@ -14,7 +14,7 @@ from LanguageModel import ModelLanguage
datapath = 'data/'
modelpath = 'model_speech/'
ms = ModelSpeech(datapath)
-ms.LoadModel(modelpath + 'speech_model24_e_0_step_216500.model')
+ms.LoadModel(modelpath + 'm24/speech_model24_e_0_step_411000.model')
ml = ModelLanguage('model_language')
ml.LoadModel()
@@ -55,7 +55,7 @@ class TestHTTPHandle(http.server.BaseHTTPRequestHandler):
for line in datas_split:
[key, value]=line.split('=')
- if('wavs' == key):
+ if('wavs' == key and '' != value):
wavs.append(int(value))
elif('fs' == key):
fs = int(value)
@@ -66,8 +66,18 @@ class TestHTTPHandle(http.server.BaseHTTPRequestHandler):
else:
print(key, value)
+ if(token != 'qwertasd'):
+ buf = '403'
+ print(buf)
+ buf = bytes(buf,encoding="utf-8")
+ self.wfile.write(buf)
+ return
+
#if('python-list' == type):
- r = self.recognize([wavs], fs)
+ if(len(wavs)>0):
+ r = self.recognize([wavs], fs)
+ else:
+ r = ''
#else:
# r = self.recognize_from_file('')
@@ -82,14 +92,20 @@ class TestHTTPHandle(http.server.BaseHTTPRequestHandler):
self._set_response()
#buf = ' \n \n\nPost page\n \nPost Data:%s
Path:%s\n \n'%(datas,self.path)
+ print(buf)
buf = bytes(buf,encoding="utf-8")
self.wfile.write(buf)
def recognize(self, wavs, fs):
- r_speech = ms.RecognizeSpeech(wavs, fs)
-
- str_pinyin = r_speech
- r = ml.SpeechToText(str_pinyin)
+ 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. ')
return r
pass