Commit 534799ff authored by 张彦钊's avatar 张彦钊

add

parent 0522e469
...@@ -32,41 +32,50 @@ def name_short(): ...@@ -32,41 +32,50 @@ def name_short():
print(name_list[:10]) print(name_list[:10])
names = ["门诊部","医疗", "门诊", "研究所", "有限", "公司", "医学", "诊所", "中心","美容","医院","整形","外科", names = ["门诊部","医疗", "门诊", "研究所", "有限", "公司", "医学", "诊所", "中心","美容","医院","整形","外科",
"医美", "集团", "卫生", "机构", "专业","皮肤","管理", "医美", "集团", "卫生", "机构", "专业","皮肤","管理",
"光学", "国际", "连锁", "综合", "专科","市","测试","测试1","ceshi","ceshi1","*"] "光学", "国际", "连锁", "综合", "专科","市","*"]
location = ["街道", "社区",] location = ["街道", "社区",]
city_list = get_city_names() city_list = get_city_names()
city_list += ["香港","澳门","韩国"] city_list += ["香港","澳门","韩国"]
for word in ["市中心","丰台","科","东方"]:
if word in city_list:
city_list.remove(word)
stop_words = names + location stop_words = names + location
new_names = [] new_names = []
third_names = [] third_names = []
for name in name_list: for name in name_list:
city_tag = "flag" for test_word in ["测试","测试1","ceshi","ceshi1"]:
for city in city_list: if test_word in name:
if city in name: pass
city_tag = city else:
name = re.sub(city, '', name) city_tag = "flag"
for word in stop_words: for city in city_list:
name = re.sub(word, '', name) if city in name:
# 去除\t city_tag = city
name = re.sub(r'\t', "", name) name = re.sub(city, '', name)
# 去除 中文括号( ) for word in stop_words:
name = re.sub(r'\(.*?\)', '', name) name = re.sub(word, '', name)
# 去除 英文括号( ) # 去除\t
name = re.sub(r'\(.*?\)', '', name) name = re.sub(r'\t', "", name)
# 去除 左英文括号,右中文括号 # 去除 中文括号( )
name = re.sub(r'\(.*?\)', '', name) name = re.sub(r'\(.*?\)', '', name)
if city_tag != "flag": # 去除 英文括号( )
third_names.append(city_tag+name) name = re.sub(r'\(.*?\)', '', name)
else: # 去除 左英文括号,右中文括号
third_names.append("无") name = re.sub(r'\(.*?\)', '', name)
new_names.append(name) if city_tag != "flag":
third_names.append(city_tag+name)
else:
third_names.append("无")
new_names.append(name)
df = pd.DataFrame() df = pd.DataFrame()
df['old_name'] = name_list df['old_name'] = name_list
df['new_name'] = new_names df['new_name'] = new_names
df['core_name'] = third_names df['core_name'] = third_names
print(df.head(6)) print(df.head(6))
df.to_csv("/tmp/"+"name_22.csv",index=None) df.to_csv("/tmp/"+"name_28.csv",index=None)
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment