spider_get_one_week_weather.py 2.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172
  1. # -*- coding: utf-8 -*-
  2. '''
  3. 获取天气预报
  4. '''
  5. import os
  6. import sys
  7. sys.path.append(os.path.join(os.getcwd().split('auto_news_scheduler')[0], 'auto_news_scheduler'))
  8. import time
  9. from datetime import datetime
  10. import httpx
  11. from bs4 import BeautifulSoup
  12. from tools_logs_handle import LogsHandle
  13. from tools_send_email import SendEmail
  14. class Weather():
  15. def __init__(self):
  16. self.email_subject = '天气预报'
  17. self.email_title = 'Weather forecast'
  18. self.email_text = '获取数据时间:\n{0}\n{1}\n\n\n\n'.format(datetime.now().strftime("%Y-%m-%d %H:%M:%S"), ('-' * 90))
  19. self.logs_handle = LogsHandle()
  20. self.now_day = time.strftime('%Y-%m-%d', time.localtime())
  21. def main(self):
  22. self.logs_handle.logs_write('Weather forecast', '开始获取天气预报数据', 'start', False)
  23. try:
  24. area_code = '59287'
  25. one_week = [
  26. '/tomorrow-%s.htm' % area_code,
  27. '/third-%s.htm' % area_code,
  28. '/fourth-%s.htm' % area_code,
  29. '/fifth-%s.htm' % area_code,
  30. '/sixth-%s.htm' % area_code,
  31. '/seventh-%s.htm' % area_code,
  32. ]
  33. url = "https://tianqi.2345.com/today-%s.htm" % area_code
  34. header = {
  35. 'User-Agent': 'Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; WOW64; Trident/4.0; SLCC2; Media Center PC 6.0; InfoPath.2; MS-RTC LM 8'
  36. }
  37. response = httpx.get(url=url, headers=header) # 增加headers参数,简单伪装UA
  38. response.encoding = "utf-8"
  39. bs = BeautifulSoup(response.text, 'html.parser') # 这里我们用html.parser解析器
  40. one_week_weather = []
  41. for week in one_week:
  42. a = bs.find_all('a', href=week) # 查找对应元素
  43. a = ' '.join(a[0].text.split())
  44. one_week_weather.append(a)
  45. except Exception as e:
  46. print(e)
  47. self.logs_handle.logs_write('Weather forecast', e, 'error', False)
  48. exit(0)
  49. subject = "天气预报"
  50. title = "weather"
  51. text = "天气预报: {}获取并发送\n\n".format(time.strftime('%Y-%m-%d %H:%M:%S', time.localtime())) + '\n\n'.join(one_week_weather)
  52. S = SendEmail(subject=subject, title=title, text=text)
  53. S.send()
  54. self.logs_handle.logs_write('Weather forecast', '天气预报数据已获取', 'done', False)
  55. # if __name__ == "__main__":
  56. # L = LogsHandle()
  57. # L.logs_write('Weather forecast', '开始获取天气预报数据', 'start', False)
  58. # W = Weather()
  59. # W.main()
  60. # L.logs_write('Weather forecast', '天气预报数据已获取', 'done', False)