def extract_szdq(content):
#所在地区的判断调用了cpca，保存结果的省、市、区；结果中无法判断省、市、区的为空；有多种匹配结果的，可能出现warning
retList = []
contentList = [content]
szdq = cpca.transform(contentList)
if content == szdq['地址'].tolist()[0] or (len(content) > len(szdq['省'].tolist()[0]) and szdq['市'].tolist()[0] == ''):
szdq = cpca.transform(contentList, cut=False)
retList += szdq['省'].tolist()
retList += szdq['市'].tolist()
retList += szdq['区'].tolist()
return retList
def test_transform():
addr_list = ["徐汇区虹漕路461号58号楼5楼", "泉州市洛江区万安塘西工业区", "福建省鼓楼区鼓楼医院",
"天津市",
"我家的地址是江苏淮安清江浦区人民路111号",
'我家的地址是江苏淮安清江浦区上海路111号',
"上海市浦东新区东明路街道三林路15号",
"贵州省黔南布依族苗族自治州长顺县长寨街道和平中路28号",
"宁夏",
"淮安市市辖区"]
transed = cpca.transform(addr_list)
assert_addr(transed)
# 测试pos_sensitive
transed = cpca.transform(addr_list, pos_sensitive=True)
assert_addr(transed, pos_sensitive=True)
def main():
duration = TimeDuration()
duration.start()
rows = get_corp_list()
corp_count = len(rows)
corp_index = 0
sql_list = []
for row in rows:
corp_id = row[0]
stock_code = row[1]
stock_name = row[2]
register_address = row[3]
print('{} 准备处理第{}支股票，股票代码：{}， 股票名称：{}，地址：{}'.format(
datetime.datetime.now(), corp_index, stock_code, stock_name, register_address))
df = cpca.transform([register_address])
data = json.loads(df.to_json(orient="records",force_ascii=False))
province = data[0]['省']
city = data[0]['市']
corp_index += 1
remain_count = corp_count - corp_index
print('{} 解析成功！省：{}，市：{}，当前剩余{}支股票，已耗时：{}'.format(datetime.datetime.now(), province, city, remain_count, duration.getTillNow()))
sql_list.append(get_sql(corp_id, province, city))
# print(sql_list)
batch_execute(sql_list)
duration.stop()
duration.printDurationInfo()
def earthquake_event():
print("earthquake_event begin")
cursor = connection.cursor()
sql = f"select weibo_post.post_id,weibo_post.task_id,post_content,post_time from weibo_post,noise_judge where weibo_post.post_id=noise_judge.post_id and weibo_post.task_id=noise_judge.task_id and noise_judge.noise='0' and (noise_judge.task_id={'or noise_judge.task_id='.join(earthquake_id)}) and CONCAT(noise_judge.task_id,'_',noise_judge.post_id) not in (select CONCAT(task_id,'_',post_id) from event);"
cursor.execute(sql)
text = cursor.fetchall()
events = []
for content in text:
try:
text = content[2]
for i in cut_sent(text):
values = cpca.transform([i], open_warning=False).values
province = values[0][0]
city = values[0][1]
area = values[0][2]
if(len(province) != 0):
break
e = re.search("发生.{0,6}?地震", content[2])
# print(text, province, city, area)
if e != None:
e = e.group(0)
if len(province) == 0 and len(city) == 0 and len(area) == 0:
continue
else:
t = event(task_id=content[1], post_id=content[0], province=province, city=city, area=area,event=e,time=content[3])
events.append(t)
except:
continue
event.objects.bulk_create(events)
cursor.close()
print("earthquake_event end")
return
def analysis_data():
with open('result.csv', 'r') as f:
data = f.read().split('\n')
citys = []
total = 0
for l in data:
l = l.strip()
if l == '':
continue
total += 1
city = l.split(',')[2]
df = cpca.transform([city])
if (df.get_values()[0][1] == ''):
continue
city = df.get_values()[0][1].replace('市', '')
print(city)
no_citys = ['池州', '黔西南布依族苗族自治州'] # 去除pyecharts中未收录的城市
if city in no_citys: continue
#city = city.replace('省','')
flag = True
for c in citys:
if c['name'] == city:
c['num'] += 1
flag = False
break
if flag:
citys.append({'name': city, 'num': 1})
# 按照数量排个序
citys = sorted(citys, key=lambda e: e["num"], reverse=True)
print(citys)
c = [(city['name'], city['num']) for city in citys]
return (c, total)
def get_city_class(lines: List[str]) -> Dict:
import cpca
import jieba
import jieba.posseg as pseg
jieba.enable_paddle()
court_name = None
citylist = []
for line in lines: # 提取人民法院名称
if line == '\n':
continue
if "人民法院" in line:
line = re.sub(r'\s', '', line)
line = re.split(r'[，：；。\s+]', line)
for subline in line:
if "人民法院" in subline:
subline = re.sub(r'[\n\s]', '', subline)
seg_list = pseg.cut(subline, use_paddle=True)
for seg in seg_list:
if seg.flag == 'ns' or seg.flag == 'nt' or seg.flag == 'ORG':
court_name = re.sub(r'[，：；。]', '', seg.word)
break
if court_name is not None:
break
if court_name is not None: # 进行地市分级
city = re.sub("人民法院", '', court_name)
citylist.append(city)
city_class = cpca.transform(citylist)
break
return city_class.iloc[0].to_dict()
def stu_map_analysis(strlist):
print("开始清洗地址数据==============>>>>>>")
# json数组字符串解构
address = strlist.get('data')
# before_split = list[1:len(list)-1]
# address = before_split.split(',')
c = {"address": address}
df = pd.DataFrame(c, columns=['address'])
adlist = df['address'].tolist()
# 地址分词
cut = cpca.transform(adlist)
# 替换表中的空字符串为NaN
cut.replace(to_replace=r'^\s*$', value=np.nan, regex=True, inplace=True)
# 删除'省'有空值的行，在原表更改
cut.dropna(subset=['省'], inplace=True)
provincelist = cut['省'].tolist()
mplist = []
for i in range(len(provincelist)):
s = provincelist[i]
# 删除两端空白
strips = re.sub('D', '', s)
# 删除后缀名
res = re.search('(?P<province>[^省|市|壮族自治区|自治区|维吾尔自治区|回族自治区|特别行政区]+)',
strips)
mplist.append(res.groupdict().get('province'))
mplist = pd.DataFrame(mplist)
counts = mplist[0].value_counts()
data = []
# 组装响应数据
for i in range(len(counts)):
data.append({"name": counts.index[i], "value": counts[i]})
print("学生地址信息清理完成============>>>>>>")
return data
def getinfos(str):
querystr = "https://www.qichacha.com/search?key="
dict = {}
dict['原公司名称'] = str
r = requests.get(querystr + str, headers=headers)
doc = pq(r.text)
num = doc('#countOld > span').text()
if r.status_code != 200 or not num: raise ValueError(u'查询过多，本日IP已被企查查禁用')
if int(num.strip()) == 0: raise ValueError(u'没有找到相关公司名称')
infos = doc('#search-result > tr:nth-child(1) > td:nth-child(3)')
a = infos('a')
infos_str = infos.text() + chr(10) #把最后一行也加一个换行符，方便用正则匹配
dict['查找到公司名称'] = a.text().split(' ')[0]
for item in ['负责人', '法定代表人', '注册资本', '成立时间', '邮箱', '电话', '地址']:
resp = re.search(f'{item}：\s?(.+?)\s', infos_str)
if resp:
dict[item] = re.search(f'{item}：\s?(.+?)\s', infos_str).group(1)
else:
dict[item] = ''
if "贵安" in dict['地址']:
dict['城市'] = "贵安新区"
else:
df = cpca.transform([dict['地址']])
dict['城市'] = df.loc[0, '市'] + df.loc[0, '区']
return dict
def parse_address(text):
text_arr = {text}
df = cpca.transform(text_arr, cut=False)
province = df.iat[0, 0]
city = df.iat[0, 1]
district = df.iat[0, 2]
other = df.iat[0, 3]
#print(other)
# pattern = re.compile(district + '(?P<street>.*?(街道|道|站|局|处|街|大道|小区|场|坊|路|团|委员会|区|县|州)){0,1}(\w*){0,1}')
# result = pattern.findall(text)
# print(result)
addr_json = {
"province": province,
"city": city,
"district": district,
"street": "",
"other": ""
}
addr_json = get_street(addr_json, text)
addr_json = get_other(addr_json, text)
# print(addr_json)
return addr_json
def city_distinguish(str_):
ending = []#最后输出为三元列表，分别为出发地，目的地，时间，时间可能因输入多个时间而形式为列表，请识别类型后处理
c = [str_]
df1 = cpca.transform(c,cut=False,open_warning=False)
prov1 = df1.iloc[0,0]
city1 = df1.iloc[0,1]
city2 = df1.iloc[0,2]
if city2 != "":
city1 = city2
if prov1 == city1:
n = prov1.find("省")
str_ = str_.replace(prov1[:n],"")
str_ = str_.replace(prov1[:2],"")
elif prov1 != city1:
n = prov1.find("省")
m = city1.find("市")
str_ = str_.replace(prov1[:n],"").replace(city1[:m],"")
str_ = str_.replace(prov1[:2],"").replace(city1[:2],"")
c1 = [str_]
#print(c1)
df2 = cpca.transform(c1,cut=False,open_warning=False)
#print(df2)
if df2.iloc[0,0] != '':
if prov1 != city1:
start_city = city1
elif prov1 == city1:
start_city = prov1
if df2.iloc[0,0] != df2.iloc[0,1]:
destination = df2.iloc[0,1]
elif df2.iloc[0,0] == df2.iloc[0,1]:
destination = df2.iloc[0,0]
#print("出发地:",start_city,"目的地:",destination)
ending.extend([start_city,destination])
elif df2.iloc[0,0] == '':
start_city = "您好像没有输入出发地"
if prov1 != city1 and city1 != "":
destination = city1
elif prov1 == city1:
destination = prov1
elif prov1 != "" and city1 == "":
destination = prov1
#print(start_city,"目的地:",destination)
ending.extend(["北京市",destination])
time = time_extract(str_)
ending.extend(time)
return ending
def get_company_info(self, company_url, item):
"""
获取工商信息和变更信息
@param item:
@param company_url: 公司请求url
@return: 返回获取到的数据
"""
headers = {
'User-Agent': UserAgent(verify_ssl=False).random,
'cookie': self.cookie,
}
try:
response_com = requests.get(company_url, headers=headers, verify=False)
company_response = response_com.text.replace(' ', '').replace('\n', '')
except Exception as e:
print(f'获取失败! {self.company_name} --> {company_url}--> {e}')
return {}
else:
time.sleep(1)
# 公司名
item['company_name'] = self.merge_list(re.findall(r'<h1>(.*?)</h1>', company_response))
# 注册资本
register_capital = ''.join(
re.findall(r'<tdclass=["\']tb["\']>注册资本</td><td.*?>(.*?)</td>', company_response, re.DOTALL))
item['register_capital'], item['register_currency'] = self.money(
re.sub(r'[\n 注册资本：:]', '', register_capital, re.DOTALL))
# 实缴资本
register_capital = ''.join(
re.findall(r'<tdclass=["\']tb["\']>实缴资本</td><td.*?>(.*?)<', company_response, re.DOTALL))
item['real_capital'], item['real_currency'] = self.money(
re.sub(r'[\n 实缴资本：:]', '', register_capital, re.DOTALL))
# 统一社会信用代码
credit_code = ''.join(
re.findall(r'<tdclass=["\']tb["\']>统一社会信用代码</td><tdclass=["\']["\']>(.*?)</td>', company_response,
re.DOTALL))
item['credit_code'] = re.sub(r'[\n 统一社会信用代码：:-]', '', credit_code, re.DOTALL)
# 工商注册号
business_number = ''.join(
re.findall(r'<tdclass=["\']tb["\']>工商注册号</td><tdclass=["\']["\']>(.*?)</td>', company_response,
re.DOTALL))
item['business_number'] = re.sub(r'[\n 工商注册号：:-]', '', business_number, re.DOTALL)
# 登记机关
reg_addr = ''.join(re.findall(
r'<tdwidth=["\'][0-9]{0,}%["\']class=["\']tb["\']>登记机关</td><tdwidth=["\'][0-9]{0,}%["\']class=["\']["\']>(.*?)</td>',
company_response, re.DOTALL))
item['reg_addr'] = re.sub(r'[\n 登记机关：:-]', '', reg_addr, re.DOTALL)
# 所属行业
industry = ''.join(
re.findall(r'<tdclass=["\']tb["\']>所属行业</td><tdclass=["\']["\']>(.*?)</td>', company_response,
re.DOTALL)).strip()
item['industry'] = re.sub(r'[\n 所属行业：:-]', '', industry, re.DOTALL)
# 企业类型
business_type = ''.join(re.findall(
r'<tdwidth=["\'][0-9]{0,}%["\']class=["\']tb["\']>企业类型</td><tdwidth=["\'][0-9]{0,}%["\']class=["\']["\']>(.*?)</td>',
company_response, re.DOTALL)).strip()
item['business_type'] = re.sub(r'[\n 企业类型：:-]', '', business_type, re.DOTALL)
item['province'], item['city'], item['area'] = cpca.transform([item['address']]).loc[0, ['省', '市', '区']]
return item
def insertToMongoDB(set1):
# 打开文件teacher.csv
with open('D:\\bysj-workspace\\education_data\\2_student_info.csv',
'r',
encoding='gbk') as csvfile:
# 调用csv中的DictReader函数直接获取数据为字典形式
reader = csv.DictReader(csvfile)
# 创建一个counts计数一下 查看一共添加了了多少条数据以验证是否正确无重复添加
counts = 0
demo = {}
place = ['demo']
for each in reader:
# 将数据中需要转换类型的数据转换类型。
demo['stu_id'] = int(each['bf_StudentID'])
demo['stu_name'] = each['bf_Name']
demo['stu_sex'] = each['bf_sex']
demo['stu_nation'] = each['bf_nation']
demo['stu_borndate'] = int(each['bf_BornDate'])
demo['cla_name'] = each['cla_Name']
# 判断住址是否登记
if each['bf_NativePlace'].strip() == '':
demo['stu_nativeplace'] = "未登记"
else:
place[0] = each['bf_NativePlace']
df = cpca.transform(place)
# 首先将pandas读取的数据转化为array
data_array = np.array(df)
# 然后转化为list形式
data_list = data_array.tolist()
#浙江省内细分到市，省外到省
if data_list[0][0] == '浙江省':
demo['stu_nativeplace'] = data_list[0][0] + data_list[0][1]
else:
demo['stu_nativeplace'] = data_list[0][0]
demo['stu_residencetype'] = each['Bf_ResidenceType']
demo['stu_policy'] = each['bf_policy']
demo['cla_id'] = int(each['cla_id'])
demo['cla_term'] = each['cla_term']
#判断学生是否处于休学
if each['bf_leaveSchool'].strip() == '':
demo['stu_leaveschool'] = "否"
if each['bf_zhusu'].strip() == '':
demo['stu_zhusu'] = "不住宿"
demo['stu_qinshihao'] = 0
else:
demo['stu_zhusu'] = "住宿"
demo['stu_qinshihao'] = int(each['bf_qinshihao'])
else:
demo['stu_leaveschool'] = "是"
demo['stu_zhusu'] = "不住宿"
demo['stu_qinshihao'] = 0
# 每次使用同一个变量存储不同的数据，导致数据库认为每次存储的是同一条数据，最终生成同一个_id值
# 手动添加_id值,当插入的数据带有_id的字段时,mongodb就不再自动生成_id
demo['_id'] = int(each['bf_StudentID'])
print(demo)
set1.insert_one(demo)
counts += 1
print('成功添加了' + str(counts) + '条数据 ')
def readfile():
file_object = open('douban.txt', 'r')
try:
for line in file_object:
item = json.loads(line)
if item == None:
continue
author = item['author']
text = item['text']
images = item['images']
id = item['id']
addr_transform = cpca.transform([text])
addr = None
if addr_transform['省'].str.split(' ')[0] != None:
addr = addr_transform['省'].str.split(' ')[0][0].rstrip('省')
if addr is None and author['loc'] is not None:
cpca.transform([author['loc']['name']])
if addr_transform['省'].str.split(' ')[0] != None:
addr = addr_transform['省'].str.split(' ')[0][0].rstrip('省')
if addr is not None:
if addr == '广西壮族自治区':
addr = '广西'
if addr == '香港特别行政区':
addr = '香港'
if addr == '澳门特别行政区':
addr = '澳门'
addr_dic[addr] = addr_dic.get(addr, 0) + 1
seg_list = jieba.cut(text, cut_all=False)
text_list.extend(seg_list)
index = 0
for i in images:
index = index + 1
url = i.get('large').get('url')
r = requests.get(url)
with open('./image/{}-{}.jpg'.format(id, index), 'wb') as f:
f.write(r.content)
finally:
file_object.close()
async def run(self, dispatcher: CollectingDispatcher,
tracker: Tracker,
domain: Dict[Text, Any]) -> List[Dict[Text, Any]]:
user_in = tracker.latest_message.get("text")
province, city = cpca.transform([user_in]).loc[0, ["省", "市"]]
city = province if city in ["市辖区", None] else city
text = await self.get_weather(await self.get_location_id(city))
dispatcher.utter_message(text=text)
return []
def match_name_tel(s: str): # get name and phone number
s = s.rstrip(".")
match = s.split(",")
name = match[0] #get name
tel = re.search('\d{11}', match[1]).group(0) #get phone number
mat = re.split('\d{11}', match[1])
rest_addr = mat[0] + mat[1] # join two string
DataFrame = cpca.transform([rest_addr]) # use the ilst
return (name, tel, DataFrame)
def is_loc(loc):
d = cpca.transform([loc])
if str(d['省'][0]):
return True
if str(d['市'][0]):
return True
if str(d['区'][0]):
return True
return False
def is_loc(loc):
d = cpca.transform([loc], open_warning=False)
if str(d['省'][0]):
return True
if str(d['市'][0]):
return True
if str(d['区'][0]):
return True
return False
def process_item(self, item, spider):
if isinstance(item, HuocheDealerItem):
if item.get('address') and (not item.get('province')
or not item['city']):
dataframe = transform([item['address']])
item['province'] = dataframe['省'].values[0]
item['city'] = dataframe['市'].values[0]
if item.get('tags') and isinstance(item['tags'], list):
item['tags'] = ','.join(item['tags'])
return item
def judge_cpca(senta, sentb):
'''
判断省市区信息是否一致
'''
# logger.info('judge_cpca...')
df = cpca.transform([senta, sentb],open_warning=False,cut=False, lookahead=3)
for i, j in zip(df.loc[0][:-1], df.loc[1][:-1]):
if i != j:
return False
return True
def judge_source(_location):
result = cpca.transform(_location, cut=False)
result = result[['省', '市', '区']]
maps = MapAreas()
for var in result.columns:
result[var] = result[var].apply(lambda x: maps.trans_map(x))
if result.loc[(result['省'] == '') & (result['市'] == '') &
(result['区'] == '')].shape[0] > 0:
result.loc[(result['省'] == '') & (result['市'] == '') &
(result['区'] == ''), ['省']] = '000000'
return result
def get_area_out_txt():
file = open('../data/deal/titles.txt', mode='r', encoding='utf-8')
area_out = open('../data/deal/area_out.txt', mode='w+', encoding='utf-8')
titles = []
for title in file.readlines():
titles.append(str(title).splitlines()[0])
df = cpca.transform(titles, cut=False, lookahead=3)
area_out.writelines('ID' + '\t' + '省' + '\t' + '市' + '\t' + '区' + '\t' + '地址' + '\n')
for index, rows in df.iterrows():
row = str(index) + '\t' + rows['省'] + '\t' + rows['市'] + '\t' + rows['区'] + '\t' + rows['地址'] + '\n'
area_out.writelines(row)
print('地点数据获取完成')
def place(newPlace):
originalText = [newPlace]
placeKeyVlue = cpca.transform(originalText, cut=False)
placeVlue = placeKeyVlue.values
place = placeVlue[0][0] + placeVlue[0][1] + placeVlue[0][2]
if place != '':
province = placeVlue[0][0]
country = '中国'
else:
province = ''
country = ''
return province, country
