郭嘉统计局虽然经常更新地区数据，但是其数据结构糟糕透顶，plain
HTML有没有！都不提供SQL或者是XML数据类型！ 都还得写个解析器来加载这个结构，用LXML解析的过程我就不写了
去除各种table之后的数据库
110000 北京市 110100 市辖区 110101 东城区 110102 西城区 110105 朝阳区 110106 丰台区 110107
石景山区 110108 海淀区
转换以后：
省:北京
├─ 市辖区
│ ├─ 东城区
│ ├─ 西城区
│ ├─ 朝阳区
│ ├─ 石景山区
│ ├─ 海淀区
│ ╰─ 平谷区
╰─ 县
├─ 密云县
╰─ 延庆县
规律就是邮政编码！用了groupby和defaultdict这些基本的Python东西 下面是程序啦，用法是直接打省份的名称即可。
当然，函数已经是单独实现的，所以在其他地方用也行啦～
#!/usr/bin/env python
# encoding: utf-8
from itertools import groupby
from utils import ScaleTree
def make_city_tree(path):
"""
make a tree of province, city, county of China
:path: path to database
:returns: tree
"""
def strip_zipcode(item):
"""
Strip all zipcode from stats.gov.cn
:returns: city_name
"""
return item[7:].decode('utf8').strip()
with file(path, 'r') as db:
tree = ScaleTree()
provinces = groupby(db, key=lambda x: x[:2])
for pid, province_data in provinces:
province_data = list(province_data)
province_name, cities = strip_zipcode(province_data[0]), province_data[1:]
for cid, city_data in groupby(cities, key=lambda x: x[:4]):
city_data = list(city_data)
city_name, counties = strip_zipcode(city_data[0]), city_data[1:]
tree[province_name][city_name] = map(strip_zipcode, counties)
return tree
if __name__ == '__main__':
tree = make_city_tree('db.txt')
pro = unicode(raw_input('省:'), 'utf8')
for x in tree.keys():
if pro in x:
cities = tree.get(x)
for t, city in enumerate(cities):
if t+1 != len(cities):
st = u"├"
else:
st = u'╰'
print u"\u2004\u2004\u2004%s─\u2004%s" % (st, city)
for d, county in enumerate(tree.get(x).get(city)):
if d+1 != len(tree.get(x).get(city)):
st = u"├"
else:
st = u'╰'
if t+1 != len(cities):
ct = u"│"
else:
ct = u"\u2004"
print u"\u2004\u2004\u2004%s\u2004\u2004%s─\u2004%s" % (ct, st, county)