cel_single.py 4.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146
  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. import random
  16. from click import Option
  17. from selenium import webdriver
  18. import time
  19. app = Celery('tasks', backend ='redis://172.104.92.245', broker='redis://172.104.92.245')
  20. qname=None
  21. custname=None
  22. app.user_options['preload'].add(Option(('-Z', '--cust'),
  23. default='default',
  24. help='Configuration template to use.'))
  25. @signals.user_preload_options.connect
  26. def on_preload_parsed(options, **kwargs):
  27. global custname
  28. # if options.get_key('template')
  29. print(options['cust'])
  30. custname=options['cust']
  31. # print(options)
  32. @celeryd_init.connect
  33. def configure_workers(sender=None, **kwargs):
  34. global qname
  35. worker_name = sender.split("@")[-1]
  36. print("******")
  37. print(worker_name)
  38. qname=worker_name
  39. print("******")
  40. app.send_task(
  41. 'cel_seo_click.selenium_jared_click',
  42. args=(),
  43. queue=qname)
  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. global custname
  53. db = dataset.connect('postgresql://postgres:eyJhbGciOiJI@172.105.241.163:5432/postgres')
  54. options = webdriver.ChromeOptions()
  55. options.add_argument("--no-sandbox")
  56. options.add_argument("--headless")
  57. options.add_argument("--incognito")
  58. #options.add_argument('--blink-settings=imagesEnabled=false')
  59. PROXY=None
  60. if PROXY is not None:
  61. options.add_argument('--proxy-server=http://%s' % PROXY)
  62. mobile_emulation = {
  63. "deviceMetrics": { "width": 360, "height": 640, "pixelRatio": 3.0 },
  64. "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" }
  65. try:
  66. driver = webdriver.Chrome(options=options)
  67. # driver = webdriver.Chrome(ChromeDriverManager().install(),options=options)
  68. ## Create Undetected Chromedriver with Options
  69. # driver = uc.Chrome(options=options)
  70. # options = uc.ChromeOptions()
  71. except:
  72. traceback.print_exc()
  73. #kw='真理大學國際生'
  74. #domain='cia.au.edu.tw'
  75. if custname is None:
  76. custname='真理'
  77. cursor=db.query("SELECT cust,plan,prefix,domain,kw,positive FROM public.seo_jobs where cust='"+custname+"' order by random() limit 1")
  78. kw=None
  79. domain=None
  80. for c in cursor:
  81. cust=c['cust']
  82. kw=c['kw']
  83. plan=c['plan']
  84. prefix=c['prefix']
  85. domain=eval(c['domain'])[0]
  86. positive=eval(c['positive'])
  87. break
  88. #kw='真理大學教堂'
  89. #domain='udn.com'
  90. print(kw)
  91. print(domain)
  92. #kw='真理大學校友'
  93. #kw='真理大學國際生'
  94. #kw='真理大學張聰聯'
  95. #domain='au.edu.tw'
  96. #domain='pronews.tw'
  97. driver.get('https://www.google.com?num=100')
  98. time.sleep(3)
  99. print(driver.current_url)
  100. elmt = driver.find_element(By.XPATH, "//textarea[@name='q']")
  101. time.sleep(1)
  102. elmt.send_keys(kw)
  103. elmt.send_keys(Keys.ENTER)
  104. time.sleep(4)
  105. elmts = driver.find_elements(By.XPATH, "//a[@jsname='UWckNb']")
  106. numresults=len(elmts)
  107. print('搜尋結果數量',numresults)
  108. if numresults<=0:
  109. print(elmts)
  110. for elmt in elmts:
  111. href=elmt.get_attribute('href')
  112. txt=elmt.text
  113. # print(txt)
  114. # print(href)
  115. if domain in href:
  116. webdriver.ActionChains(driver).move_to_element(elmt).perform()
  117. webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
  118. print(domain)
  119. print(href)
  120. driver.quit()
  121. return txt
  122. break
  123. return '{empty}'
  124. # time.sleep(5)