Python爬虫-获取天气城市code

这里的城市包括全球各个国家，可以根据需要剔除，当然国外的在天气网站上也是可以查天气的～

#!/usr/bin/env python3
# -*- coding : utf-8 -*-

'全国天气城市code'

import re
import requests
from bs4 import BeautifulSoup
import datetime
import pymysql

# 打开数据库连接
db = pymysql.connect(host="localhost", user="dev", passwd="000000", db="test", use_unicode=True, charset="utf8")
# 获得一个游标
cursor = db.cursor()

city_rul = 'https://my.oschina.net/joanfen/blog/140364'

header = {
        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
        'Accept-Encoding': 'gzip, deflate, br',
        'Accept-Language': 'zh-CN,zh;q=0.8',
        'Connection': 'keep-alive',
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 '
                      '(KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.36'
    }


def get_data(city_rul,header):
    response = requests.get(city_rul, headers=header)
    response.encoding = 'utf-8'
    html = response.text
    data = BeautifulSoup(html, 'html.parser')
    content = data.find('div', {'class': 'BlogContent'})
    # 获取省列表
    # h4 = content.find_all('h4')
    # for h in h4:
    #     print(h.contents)

    p = content.find_all('p')
    city = re.split('\n', content.get_text().strip())
    city_end = [['北京', '101010100']]
    for c in city:
        if c != '':
            cd = re.split(r'\s', c.strip())
            cic = []
            for ct in cd:
                if ct != '':
                    # 去除单个的省名：
                    if len(ct) > 3:
                        code = re.search(r'[0-9]{9}', ct)
                        st = re.search(r'[\u4e00-\u9fa5]{2,9}', ct)
                        if code.group() == '101010100':
                            pass
                        else:
                            cic.append(st.group())
                            cic.append(code.group())
                        city_end.append(cic)
    return city_end # 返回一个list


def write_sql(city_list, file='G:/python/city_code.sql'):
    with open(file, 'w') as f: # 如果直接保存到数据库，则不需要打开文件
        now = datetime.datetime.now()
        i = 1
        for city in city_list:
            sql = 'insert into city_code(id, city_name, city_code, gmt_created, gmt_modified)' \
                  ' values (%d,\'%s\',\'%s\',\'%s\',\'%s\')' % (i, city[0], city[1], now, now)+';'
            # f.write(sql) # 保存到sql
            i += 1
            # print(sql)
            # 直接写入到数据库
            try:
                # 执行sql语句
                cursor.execute(sql)
                # 提交到数据库执行
                db.commit()
            except:
                # 如果发生错误则回滚
                db.rollback()
            # 关闭数据库连接
        db.close()
if __name__ == '__main__':
    city_list = get_data(city_rul, header)
    write_sql(city_list,)