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