cel_seo_click.py 3.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126
  1. import time
  2. import traceback
  3. import dataset
  4. from selenium import webdriver
  5. from selenium.webdriver.support.ui import WebDriverWait
  6. from selenium.webdriver.common.by import By
  7. from selenium.webdriver.support import expected_conditions as EC
  8. from selenium.webdriver.common.keys import Keys
  9. from celery import Celery
  10. import redis
  11. from celery.schedules import crontab
  12. import celery
  13. from billiard import current_process
  14. app = Celery('tasks', backend ='redis://172.104.92.245', broker='redis://172.104.92.245')
  15. print('*****')
  16. #print(celery.utils.nodenames.gethostname())
  17. p = current_process()
  18. print(p.initargs[1].split('@')[1])
  19. print('*****')
  20. #rkey=app.request.delivery_info['routing_key']
  21. app.conf.beat_schedule = {
  22. # Executes every Monday morning at 7:30 a.m.
  23. 'add-every-monday-morning': {
  24. 'task': 'cel_seo_click.selenium_jared_click',
  25. # 'schedule': crontab(minute='*/1'),
  26. 'schedule': 30.0,
  27. 'args': (),
  28. 'options':{'queue': 'xxx'}
  29. },
  30. }
  31. #@app.on_after_configure.connect
  32. #def setup_periodic_tasks(sender, **kwargs):
  33. # print(app.request.delivery_info['routing_key'])
  34. # print(json_str)
  35. #sender task.request.hostnam
  36. # sender.add_periodic_task(100.0, selenium_jared_click.s(), name='add every 2 min')
  37. @app.task(queue='ABC')
  38. def selenium_jared_click():
  39. db = dataset.connect('postgresql://postgres:eyJhbGciOiJI@172.105.241.163:5432/postgres')
  40. options = webdriver.ChromeOptions()
  41. options.add_argument("--no-sandbox")
  42. options.add_argument("--headless")
  43. options.add_argument("--incognito")
  44. #options.add_argument('--blink-settings=imagesEnabled=false')
  45. mobile_emulation = {
  46. "deviceMetrics": { "width": 360, "height": 640, "pixelRatio": 3.0 },
  47. "userAgent": "Mozilla/5.0 (Linux; Android 4.2.1; en-us; Nexus 5 Build/JOP40D) AppleWebKit/535.19 (KHTML, like Gecko) Chrome/18.0.1025.166 Mobile Safari/535.19" }
  48. try:
  49. driver = webdriver.Chrome(options=options)
  50. # driver = webdriver.Chrome(ChromeDriverManager().install(),options=options)
  51. ## Create Undetected Chromedriver with Options
  52. # driver = uc.Chrome(options=options)
  53. # options = uc.ChromeOptions()
  54. except:
  55. traceback.print_exc()
  56. #kw='真理大學國際生'
  57. #domain='cia.au.edu.tw'
  58. cursor=db.query("SELECT cust,plan,prefix,domain,kw,positive FROM public.seo_jobs where cust='真理' order by random() limit 1")
  59. kw=None
  60. domain=None
  61. for c in cursor:
  62. cust=c['cust']
  63. kw=c['kw']
  64. plan=c['plan']
  65. prefix=c['prefix']
  66. domain=eval(c['domain'])[0]
  67. positive=eval(c['positive'])
  68. break
  69. #kw='真理大學教堂'
  70. #domain='udn.com'
  71. print(kw)
  72. print(domain)
  73. #kw='真理大學校友'
  74. #kw='真理大學國際生'
  75. #kw='真理大學張聰聯'
  76. #domain='au.edu.tw'
  77. #domain='pronews.tw'
  78. driver.get('https://www.google.com?num=100')
  79. time.sleep(3)
  80. print(driver.current_url)
  81. elmt = driver.find_element(By.XPATH, "//textarea[@name='q']")
  82. time.sleep(1)
  83. elmt.send_keys(kw)
  84. elmt.send_keys(Keys.ENTER)
  85. time.sleep(4)
  86. elmts = driver.find_elements(By.XPATH, "//a[@jsname='UWckNb']")
  87. numresults=len(elmts)
  88. print('搜尋結果數量',numresults)
  89. if numresults<=0:
  90. print(elmts)
  91. for elmt in elmts:
  92. href=elmt.get_attribute('href')
  93. txt=elmt.text
  94. # print(txt)
  95. # print(href)
  96. if domain in href:
  97. webdriver.ActionChains(driver).move_to_element(elmt).perform()
  98. webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
  99. print(domain)
  100. print(href)
  101. driver.quit()
  102. return txt
  103. break
  104. return '{empty}'
  105. # time.sleep(5)