cel_seo_click.py 4.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143
  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. from celery import signals
  15. from click import Option
  16. app = Celery('tasks', backend ='redis://172.104.92.245', broker='redis://172.104.92.245')
  17. qname=None
  18. app.user_options['preload'].add(Option(('-Z', '--template'),
  19. default='default',
  20. help='Configuration template to use.'))
  21. @signals.user_preload_options.connect
  22. def on_preload_parsed(options, **kwargs):
  23. print(options['template'])
  24. print(options)
  25. @celeryd_init.connect
  26. def configure_workers(sender=None, **kwargs):
  27. global qname
  28. worker_name = sender.split("@")[-1]
  29. #rkey=app.request.delivery_info['routing_key']
  30. print("******")
  31. print(worker_name)
  32. qname=worker_name
  33. print("******")
  34. app.conf.beat_schedule = {
  35. # Executes every Monday morning at 7:30 a.m.
  36. 'add-every-monday-morning': {
  37. 'task': 'cel_seo_click.selenium_jared_click',
  38. # 'schedule': crontab(minute='*/1'),
  39. 'schedule': 100.0,
  40. 'args': (),
  41. 'options':{'queue': qname}
  42. },
  43. }
  44. #@app.on_after_configure.connect
  45. #def setup_periodic_tasks(sender, **kwargs):
  46. # print(app.request.delivery_info['routing_key'])
  47. # print(json_str)
  48. #sender task.request.hostnam
  49. # sender.add_periodic_task(100.0, selenium_jared_click.s(), name='add every 2 min')
  50. @app.task()
  51. def selenium_jared_click():
  52. db = dataset.connect('postgresql://postgres:eyJhbGciOiJI@172.105.241.163:5432/postgres')
  53. options = webdriver.ChromeOptions()
  54. options.add_argument("--no-sandbox")
  55. options.add_argument("--headless")
  56. options.add_argument("--incognito")
  57. #options.add_argument('--blink-settings=imagesEnabled=false')
  58. mobile_emulation = {
  59. "deviceMetrics": { "width": 360, "height": 640, "pixelRatio": 3.0 },
  60. "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" }
  61. try:
  62. driver = webdriver.Chrome(options=options)
  63. # driver = webdriver.Chrome(ChromeDriverManager().install(),options=options)
  64. ## Create Undetected Chromedriver with Options
  65. # driver = uc.Chrome(options=options)
  66. # options = uc.ChromeOptions()
  67. except:
  68. traceback.print_exc()
  69. #kw='真理大學國際生'
  70. #domain='cia.au.edu.tw'
  71. cursor=db.query("SELECT cust,plan,prefix,domain,kw,positive FROM public.seo_jobs where cust='真理' order by random() limit 1")
  72. kw=None
  73. domain=None
  74. for c in cursor:
  75. cust=c['cust']
  76. kw=c['kw']
  77. plan=c['plan']
  78. prefix=c['prefix']
  79. domain=eval(c['domain'])[0]
  80. positive=eval(c['positive'])
  81. break
  82. #kw='真理大學教堂'
  83. #domain='udn.com'
  84. print(kw)
  85. print(domain)
  86. #kw='真理大學校友'
  87. #kw='真理大學國際生'
  88. #kw='真理大學張聰聯'
  89. #domain='au.edu.tw'
  90. #domain='pronews.tw'
  91. driver.get('https://www.google.com?num=100')
  92. time.sleep(3)
  93. print(driver.current_url)
  94. elmt = driver.find_element(By.XPATH, "//textarea[@name='q']")
  95. time.sleep(1)
  96. elmt.send_keys(kw)
  97. elmt.send_keys(Keys.ENTER)
  98. time.sleep(4)
  99. elmts = driver.find_elements(By.XPATH, "//a[@jsname='UWckNb']")
  100. numresults=len(elmts)
  101. print('搜尋結果數量',numresults)
  102. if numresults<=0:
  103. print(elmts)
  104. for elmt in elmts:
  105. href=elmt.get_attribute('href')
  106. txt=elmt.text
  107. # print(txt)
  108. # print(href)
  109. if domain in href:
  110. webdriver.ActionChains(driver).move_to_element(elmt).perform()
  111. webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
  112. print(domain)
  113. print(href)
  114. driver.quit()
  115. return txt
  116. break
  117. return '{empty}'
  118. # time.sleep(5)