cel_seo_click.py 4.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142
  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 celery.bin.base 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(
  19. '--jared', action='store_true',
  20. help='Enable our external monitoring utility, blahblah',
  21. ))
  22. @signals.user_preload_options.connect
  23. def handle_preload_options(options, **kwargs):
  24. if options['jared']:
  25. print(options)
  26. # enable_monitoring()
  27. @celeryd_init.connect
  28. def configure_workers(sender=None, **kwargs):
  29. global qname
  30. worker_name = sender.split("@")[-1]
  31. #rkey=app.request.delivery_info['routing_key']
  32. print("******")
  33. print(worker_name)
  34. qname=worker_name
  35. print("******")
  36. app.conf.beat_schedule = {
  37. # Executes every Monday morning at 7:30 a.m.
  38. 'add-every-monday-morning': {
  39. 'task': 'cel_seo_click.selenium_jared_click',
  40. # 'schedule': crontab(minute='*/1'),
  41. 'schedule': 100.0,
  42. 'args': (),
  43. 'options':{'queue': qname}
  44. },
  45. }
  46. #@app.on_after_configure.connect
  47. #def setup_periodic_tasks(sender, **kwargs):
  48. # print(app.request.delivery_info['routing_key'])
  49. # print(json_str)
  50. #sender task.request.hostnam
  51. # sender.add_periodic_task(100.0, selenium_jared_click.s(), name='add every 2 min')
  52. @app.task()
  53. def selenium_jared_click():
  54. db = dataset.connect('postgresql://postgres:eyJhbGciOiJI@172.105.241.163:5432/postgres')
  55. options = webdriver.ChromeOptions()
  56. options.add_argument("--no-sandbox")
  57. options.add_argument("--headless")
  58. options.add_argument("--incognito")
  59. #options.add_argument('--blink-settings=imagesEnabled=false')
  60. mobile_emulation = {
  61. "deviceMetrics": { "width": 360, "height": 640, "pixelRatio": 3.0 },
  62. "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" }
  63. try:
  64. driver = webdriver.Chrome(options=options)
  65. # driver = webdriver.Chrome(ChromeDriverManager().install(),options=options)
  66. ## Create Undetected Chromedriver with Options
  67. # driver = uc.Chrome(options=options)
  68. # options = uc.ChromeOptions()
  69. except:
  70. traceback.print_exc()
  71. #kw='真理大學國際生'
  72. #domain='cia.au.edu.tw'
  73. cursor=db.query("SELECT cust,plan,prefix,domain,kw,positive FROM public.seo_jobs where cust='真理' order by random() limit 1")
  74. kw=None
  75. domain=None
  76. for c in cursor:
  77. cust=c['cust']
  78. kw=c['kw']
  79. plan=c['plan']
  80. prefix=c['prefix']
  81. domain=eval(c['domain'])[0]
  82. positive=eval(c['positive'])
  83. break
  84. #kw='真理大學教堂'
  85. #domain='udn.com'
  86. print(kw)
  87. print(domain)
  88. #kw='真理大學校友'
  89. #kw='真理大學國際生'
  90. #kw='真理大學張聰聯'
  91. #domain='au.edu.tw'
  92. #domain='pronews.tw'
  93. driver.get('https://www.google.com?num=100')
  94. time.sleep(3)
  95. print(driver.current_url)
  96. elmt = driver.find_element(By.XPATH, "//textarea[@name='q']")
  97. time.sleep(1)
  98. elmt.send_keys(kw)
  99. elmt.send_keys(Keys.ENTER)
  100. time.sleep(4)
  101. elmts = driver.find_elements(By.XPATH, "//a[@jsname='UWckNb']")
  102. numresults=len(elmts)
  103. print('搜尋結果數量',numresults)
  104. if numresults<=0:
  105. print(elmts)
  106. for elmt in elmts:
  107. href=elmt.get_attribute('href')
  108. txt=elmt.text
  109. # print(txt)
  110. # print(href)
  111. if domain in href:
  112. webdriver.ActionChains(driver).move_to_element(elmt).perform()
  113. webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
  114. print(domain)
  115. print(href)
  116. driver.quit()
  117. return txt
  118. break
  119. return '{empty}'
  120. # time.sleep(5)