環境省花粉観測システム(はなこさん)から花粉データを取得

hanako.py

#!/usr/bin/env python
# coding:utf-8

import re
import requests

class Hanako:
    COOKIE_URL = 'http://kafun.taiki.go.jp/Hyou0.aspx?MstCode={0}&AreaCode={1}'
    INFO_URL = 'http://kafun.taiki.go.jp/Hyou2.aspx'

    def __init__(self, area_code, mst_code):
        self._area_code = area_code
        self._mst_code = mst_code
        self._info = []
        self._get_info()

    def _get_info(self):
        url = self.COOKIE_URL.format(self._mst_code, self._area_code)
        cookie = self._send_request(url).cookies
        content = self._send_request(self.INFO_URL, cookie).content
        matcher = re.compile(r'<td><font size="2">(?P<hour>[^<]*)</font></td><td align="Right"><font size="2">(?P<pollen>[^<]*)</font></td><td align="Right"><font size="2">(?P<wd>[^<]*)</font></td><td align="Right"><font size="2">(?P<ws>[^<]*)</font></td><td align="Right"><font size="2">(?P<temp>[^<]*)</font></td><td align="Right"><font size="2">(?P<prec>[^<]*)</font></td><td align="Right"><font size="2">(?P<prec_bool>[^<]*)</font></td>')
        self._info = [m.groupdict() for m in matcher.finditer(content)]

    def _send_request(self, url, cookie=None):
        r = requests.get(url, cookies=cookie)
        if r.status_code != 200:
            raise Exception(u'HTTP status is ' + r.status_code)
        return r

    def get(self, hour):
        if 1 > hour or 24 < hour:
            raise ValueError('Hour is betweet 1 and 24')
        if hour > len(self._info):
            raise Exception(u'That time is not measured')
        return self._info[hour - 1]

    def now(self):
        return self.get(len(self._info))

    def today(self):
        return self._info

if __name__ == '__main__':
    hanako = Hanako('05', '52310100')
    print hanako.now() #最新のデータ
    print hanako.get(10) #10時のデータ
    print hanako.today() #今日のデータ全て