def findSimilarEntitys(entity, list):
dictionary = {} # 用于计算相似度后，把解决方案、对应id、相似度存入字典，根据相似度大小逆序输出
for desc in list:
similar = cosin.sentence_resemble(entity, desc)
if similar > 0.6:
dictionary.update({desc: round(similar * 100,
2)}) # 将id，描述，解决方法，以及对应的相似度加入字典
dictionary = sorted(dictionary.items(),
key=lambda item: item[1],
reverse=True) # 对字典根据相似度进行降序排序
#暂且全部输出
return dictionary
def huida(q_type, zhuyu):
print(zhuyu, ':')
daanlist = []
# 某故障原因 会引起哪些现象
if q_type == 0:
relationyuanyins = []
for yuanyindb in findEntitiesByType('Yuanyin'):
similar = cosin.sentence_resemble(zhuyu, yuanyindb)
if similar > 0.8:
relationyuanyins.append(yuanyindb)
print('relationyuanyins:', relationyuanyins)
for relationyuanyin in relationyuanyins:
daanlist += findEntities2(relationyuanyin, '间接原因')
templist = []
for temp in daanlist:
if temp.isalnum():
templist.append(temp + '报警')
daanlist = templist
# 做什么操作 会遇到什么错误
if q_type == 1:
relationcaozuos = []
for caozuodb in findEntitiesByType('Caozuo'):
similar = cosin.sentence_resemble(zhuyu, caozuodb)
if similar > 0.8:
relationcaozuos.append(caozuodb)
print('relationcaozuos:', relationcaozuos)
for caozuo in relationcaozuos:
daanlist += findEntities(caozuo, '引起')
# 某部位 常出现哪些故障
if q_type == 2:
daanlist = findEntities2(zhuyu, '故障部位')
# 报警的含义是什么
if q_type == 3:
daanlist = findEntities(zhuyu, '直接原因')
temp = {}
temp = temp.fromkeys(daanlist)
daanlist = list(temp.keys())
return daanlist
def insertPa(pinpai, xinghao, errorid, describe, selectedList, yuanyin,
answer):
jieba.load_userdict(os.getcwd() + '\\Shukongdashi\\demo\\fencidian.txt')
pattern = r'\.|;|。|；|！'
pattern2 = r',|，'
# 加载停用词
stopwords = []
with open(os.getcwd() + '\\Shukongdashi\\demo\\stopwords.txt',
'r',
encoding='utf-8') as f:
st = f.readlines()
for line in st:
line = line.strip().encode('utf-8').decode('utf-8-sig')
stopwords.append(line)
# 加载故障部位
buweizhuyu = []
with open(os.getcwd() + '\\Shukongdashi\\demo\\zhuyu.txt',
'r',
encoding='utf-8') as f:
st = f.readlines()
for line in st:
line = line.strip().encode('utf-8').decode('utf-8-sig') # 防止BOM现象
buweizhuyu.append(line)
miaoshu = describe.replace(' ', '')
print(u'正在过滤停用词......')
miaoshu_baocun = []
miaoshu_cut = jieba.cut(miaoshu)
for m in miaoshu_cut:
if m not in stopwords:
miaoshu_baocun.append(m)
miaoshu = ''.join(miaoshu_baocun)
miaoshu_jvzi = re.split(pattern, miaoshu)
miaoshu_jvzi = list(filter(None, miaoshu_jvzi))
# 保存描述中描述品牌型号、执行操作、报警现象的内容
miaoshu_caozuo = []
miaoshu_xianxiang = []
for jvzi in miaoshu_jvzi:
# 把句子进一步拆分
miaoshu_list = re.split(pattern2, jvzi)
miaoshu_list = list(filter(None, miaoshu_list))
# 判断类型,把相同的类型放到一起
for miaoshu in miaoshu_list:
miaoshu_type = cnnModel.predict(miaoshu)
if miaoshu_type == '执行操作':
miaoshu_caozuo.append(miaoshu)
elif miaoshu_type == '故障现象':
miaoshu_xianxiang.append(miaoshu)
# 到此，不同的描述放到了不同的list里
print('执行操作:', miaoshu_caozuo)
print('故障现象:', miaoshu_xianxiang)
yuanyin_this = ''
try:
yuanyin_this = miaoshu_xianxiang[len(miaoshu_xianxiang) - 1] + '的故障'
except:
yuanyin_this = '未知原因引起的故障'
print(yuanyin_this)
erroridlist = re.split(',|，', errorid)
buweilist = []
for sentence in miaoshu_xianxiang:
# 处理顿号“、”
sentence_seged = jieba.posseg.cut(sentence.strip())
for x in sentence_seged:
if (x.flag == 'n' or x.flag == 'x'):
if x.word in buweizhuyu:
buweilist.append(x.word)
if '报警' in sentence:
if x.flag == 'eng' or x.flag == 'm':
if not (x.word >= u'\u4e00'
and x.word <= u'\u9fa5'): # 判断是否是汉字
erroridlist.append(x.word)
print(buweilist, erroridlist)
# 相关的操作
similarcaozuos = miaoshu_caozuo
for i, caozuo in enumerate(miaoshu_caozuo):
for caozuodb in findEntitiesByType('Caozuo'):
similar = cosin.sentence_resemble(caozuo, caozuodb)
if similar > 0.8:
similarcaozuos[i] = caozuodb
# 相关故障
relationguzhangs = []
# 相关操作引起的现象
for caozuo in similarcaozuos:
relationguzhangs += findEntities(caozuo, '引起')
# 根据故障部位和报警信息寻找现象
for entity2 in buweilist:
relationguzhangs += findEntities2(entity2, '故障部位')
for entity2 in erroridlist:
relationguzhangs += findEntities2(entity2, '报警信息')
print('relationguzhangs:', relationguzhangs)
similarxianxiang = miaoshu_xianxiang # selectedList
for i, xianxiang in enumerate(miaoshu_xianxiang):
for relationguzhang in relationguzhangs:
similar = cosin.sentence_resemble(xianxiang, relationguzhang)
if similar > 0.8:
similarxianxiang[i] = relationguzhang
print('similarxianxiang:', similarxianxiang)
# similarcaozuos,similarxianxiang,erroridlist,
# yuanyin,jiejuebanfa
# 初始化操作和故障，故障和原因之间的关系
for xianxiang in similarxianxiang:
for caozuo in similarcaozuos:
insertTwoNodes(caozuo, '引起', xianxiang, 'Caozuo', 'Xianxiang')
insertTwoNodes(xianxiang, '间接原因', yuanyin_this, 'Xianxiang', 'Yuanyin')
# 故障现象之间的关联关系
zuhe_xianxiaing = list(combinations(similarxianxiang, 2))
for zuhe in zuhe_xianxiaing:
insertTwoNodes(zuhe[0], '相关', zuhe[1], 'Xianxiang', 'Xianxiang')
insertTwoNodes(zuhe[1], '相关', zuhe[0], 'Xianxiang', 'Xianxiang')
for sentence in similarxianxiang:
# 处理顿号“、”
sentence_seged = jieba.posseg.cut(sentence.strip())
zhuyu = []
for x in sentence_seged:
if (x.flag == 'n' or x.flag == 'x'):
if x.word in buweizhuyu:
insertTwoNodes(sentence, '故障部位', x.word, 'Xianxiang',
'GuzhangBuwei')
if '报警' in sentence:
if x.flag == 'eng' or x.flag == 'm':
if not (x.word >= u'\u4e00'
and x.word <= u'\u9fa5'): # 判断是否是汉字
insertTwoNodes(sentence, '报警信息', x.word, 'Xianxiang',
'Errorid')
# 报警信息和所有现象关联
# 最好判断一下是否已经存在关系了
answer = answer.replace("'", "\\\'").replace('"', '\\\"')
t = (yuanyin_this, describe + ",".join(selectedList), answer)
sql = "insert into guzhanganli values('%s','%s','%s')" % t
print(sql)
# try:
# 执行SQL语句
cursor.execute(sql)
# 获取所有记录列表
db_mysql.commit() # 提交，不然无法保存插入或者修改的数据(这个一定不要忘记加上)
#cursor.close()
# 关闭数据库连接
#db_mysql.close() # 关闭连接
return True
def getTuili(pinpai, xinghao, errorid, describe, relationList, ret_dict):
print( errorid, describe, relationList, ret_dict)
if errorid!='':
describe += '，' + errorid + '报警'
print('目录：',os.path.dirname(os.getcwd()))
jieba.load_userdict(os.getcwd()+'\\Shukongdashi\\demo\\fencidian.txt')
pattern = r'\.|;|。|；|！'
pattern2 = r',|，'
# 加载停用词
stopwords = []
with open(os.getcwd()+'\\Shukongdashi\\demo\\stopwords.txt', 'r', encoding='utf-8') as f:
st = f.readlines()
for line in st:
line = line.strip().encode('utf-8').decode('utf-8-sig')
stopwords.append(line)
# 加载故障部位
buweizhuyu = []
with open(os.getcwd()+'\\Shukongdashi\\demo\\zhuyu.txt', 'r', encoding='utf-8') as f:
st = f.readlines()
for line in st:
line = line.strip().encode('utf-8').decode('utf-8-sig') # 防止BOM现象
buweizhuyu.append(line)
miaoshu = describe.replace(' ', '')
print(u'正在过滤停用词......')
miaoshu_baocun = []
miaoshu_cut = jieba.cut(miaoshu)
for m in miaoshu_cut:
if m not in stopwords:
miaoshu_baocun.append(m)
miaoshu = ''.join(miaoshu_baocun)
miaoshu_jvzi = re.split(pattern, miaoshu)
miaoshu_jvzi = list(filter(None, miaoshu_jvzi))
# 保存描述中描述品牌型号、执行操作、报警现象的内容
miaoshu_xinghao = []
miaoshu_caozuo = []
miaoshu_xianxiang = []
for jvzi in miaoshu_jvzi:
# 把句子进一步拆分
miaoshu_list = re.split(pattern2, jvzi)
miaoshu_list = list(filter(None, miaoshu_list))
# 判断类型,把相同的类型放到一起
for miaoshu in miaoshu_list:
miaoshu_type = cnnModel.predict(miaoshu)
if miaoshu_type == '机床类型':
miaoshu_xinghao.append(miaoshu)
elif miaoshu_type == '执行操作':
miaoshu_caozuo.append(miaoshu)
elif miaoshu_type == '故障现象':
miaoshu_xianxiang.append(miaoshu)
# 到此，不同的描述放到了不同的list里
print(miaoshu_caozuo, miaoshu_xianxiang)
# 处理品牌型号描述
miaoshu_xinghao = ''.join(miaoshu_xinghao)
pinpai_xinghao = ''.join(re.findall(u'[0-9a-zA-Z]+', miaoshu_xinghao))
if pinpai_xinghao != '':
seg_list = list(jieba.cut(pinpai_xinghao, cut_all=False))
pinpai = seg_list[0]
xinghao = ''
if len(seg_list) > 1:
xinghao = seg_list[1]
print(pinpai, xinghao)
# 处理操作
# 处理故障描述,寻找部位和故障代码
buweilist = []
erroridlist = []
for sentence in miaoshu_xianxiang:
# 处理顿号“、”
sentence_seged = jieba.posseg.cut(sentence.strip())
for x in sentence_seged:
if (x.flag == 'n' or x.flag == 'x'):
if x.word in buweizhuyu:
buweilist.append(x.word)
if '报警' in sentence:
if x.flag == 'eng' or x.flag == 'm':
if not (x.word >= u'\u4e00' and x.word <= u'\u9fa5'): # 判断是否是汉字
erroridlist.append(x.word)
# print(buweilist, erroridlist)
# 根据执行了的操作推理导致的现象
# 相关的操作
similarcaozuos = []
for caozuo in miaoshu_xianxiang:
for caozuodb in findEntitiesByType('Caozuo'):
similar = cosin.sentence_resemble(caozuo, caozuodb)
if similar > 0.8:
similarcaozuos.append(caozuodb)
# print('similarcaozuos:', similarcaozuos)
# 相关故障
relationguzhangs = []
# 相关操作引起的现象
for caozuo in similarcaozuos:
relationguzhangs += findEntities(caozuo, '引起')
# 根据故障部位和报警信息寻找现象
for entity2 in buweilist:
relationguzhangs += findEntities2(entity2, '故障部位')
for entity2 in erroridlist:
relationguzhangs += findEntities2(entity2, '报警信息')
# print('relationguzhangs:', relationguzhangs)
# 寻找相似现象
similarxianxiang = miaoshu_xianxiang
for i,xianxiang in enumerate(miaoshu_xianxiang):
for relationguzhang in relationguzhangs:
similar = cosin.sentence_resemble(xianxiang, relationguzhang)
if similar > 0.8:
similarxianxiang[i] = relationguzhang
print('similarxianxiang:', similarxianxiang)
# 找出所有相似现象
if len(buweilist) == 0 or len(erroridlist) == 0:
for i,xianxiang in enumerate(similarxianxiang):
for xianxiangdb in findEntitiesByType('Xianxiang'):
similar = cosin.sentence_resemble(xianxiang, xianxiangdb)
if similar > 0.8:
similarxianxiang[i] = xianxiangdb
print('similar:',similarxianxiang)
print('relationList:', relationList)
similarxianxiang += relationList
# print('similarxianxiang:',similarxianxiang)
temp = {}
temp = temp.fromkeys(similarxianxiang)
selectedlist = list(temp.keys())
# 推理原因，推理相关现象
tuilixianxiang = []
for xianxiang in selectedlist:
tuilixianxiang += findEntities(xianxiang, '相关')
# 删除重复
tuilixianxiangtemp = []
for chongfu in tuilixianxiang:
if (chongfu not in selectedlist) and (chongfu not in tuilixianxiangtemp):
tuilixianxiangtemp.append(chongfu)
tuilixianxiang = tuilixianxiangtemp
# 根据这个error_list查找一个和里面元素相关的元素，作数组返回
ret_dict['selectedlist'] = selectedlist
ret_dict['hiddenlist'] = tuilixianxiang
count = {}
# 根据现象查找可能的原因，将可能的原因作为键，将契合的次数作为值
for xianxiang in selectedlist:
jainjieyuanyindb = db.findOtherEntities(xianxiang, "间接原因")
selected_index = [i for i in range(len(jainjieyuanyindb))]
for i in selected_index:
jianjieyuanyin = jainjieyuanyindb[i]['n2']['title']
count.setdefault(jianjieyuanyin, 0)
count[jianjieyuanyin] = count[jianjieyuanyin] + 1
# 根据可能性重新对最终原因进行排序
for i, j in count.items():
# print(str(i) + ":", j)
relathionCount = db.findNumberOfEntities1(i, "间接原因")[0]["relathionCount"]
count[i] = round(j / relathionCount, 2)
# 对相关原因按匹配次数排序
list1 = sorted(count.items(), key=lambda x: x[1], reverse=True)
for yuanyinItem in list1:
# 查解决方法
jiejuelist = []
db_mysql = pymysql.connect(host='localhost', user='******', password='******', db='sg_faq')
cursor = db_mysql.cursor() # 获取指针以操作数据库
cursor.execute('set names utf8')
print(yuanyinItem)
sql = "select guzhangfenxi FROM guzhanganli where guzhangyuanyin = '%s'" % (yuanyinItem[0])
try:
# 执行SQL语句
cursor.execute(sql)
# 获取所有记录列表
results = cursor.fetchall()
for row in results:
jiejuelist.append(row[0])
except:
print("Error: unable to fetch data")
# 关闭数据库连接
db_mysql.close()
# 查找间接导致原因的现象个数#计算可能性
xianxiangdb = db.findOtherEntities2(yuanyinItem[0], "间接原因")
relathionCount = len(xianxiangdb)
selected_index2 = [i for i in range(len(xianxiangdb))]
tuililist = []
for i in selected_index2:
xianxiang = xianxiangdb[i]['n1']['title']
if xianxiang in selectedlist:
zhijieyuanyindb = db.findOtherEntities(xianxiang, "直接原因")
if len(zhijieyuanyindb) > 0:
zhijieyuanyi = zhijieyuanyindb[0]['n2']['title']
tuililist.append(
{"entity1": xianxiang, "rel": "含义", "entity2": zhijieyuanyi, "entity1_type": "故障代码",
"entity2_type": "含义"})
tuililist.append(
{"entity1": zhijieyuanyi, "rel": "间接原因", "entity2": yuanyinItem[0], "entity1_type": "含义",
"entity2_type": "最终原因"})
else:
tuililist.append(
{"entity1": xianxiang, "rel": "间接原因", "entity2": yuanyinItem[0], "entity1_type": "现象",
"entity2_type": "最终原因"})
for jiejue in jiejuelist:
tuililist.append({"entity1": yuanyinItem[0], "rel": "解决办法", "entity2": jiejue, "entity1_type": "最终原因",
"entity2_type": "解决办法"})
if (ret_dict.get('list') is None):
ret_dict['list'] = []
ret_dict['list'].append(
{"yuanyin": yuanyinItem[0], "answer": jiejuelist, "possibility": yuanyinItem[1], "list": tuililist})
print(ret_dict)
return ret_dict