cel_seo_click.py 4.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147
  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. custname=None
  19. app.user_options['preload'].add(Option(('-Z', '--cust'),
  20. default='default',
  21. help='Configuration template to use.'))
  22. @signals.user_preload_options.connect
  23. def on_preload_parsed(options, **kwargs):
  24. global custname
  25. # if options.get_key('template')
  26. print(options['cust'])
  27. custname=options['cust']
  28. # print(options)
  29. @celeryd_init.connect
  30. def configure_workers(sender=None, **kwargs):
  31. global qname
  32. worker_name = sender.split("@")[-1]
  33. #rkey=app.request.delivery_info['routing_key']
  34. print("******")
  35. print(worker_name)
  36. qname=worker_name
  37. print("******")
  38. app.conf.beat_schedule = {
  39. # Executes every Monday morning at 7:30 a.m.
  40. 'add-every-monday-morning': {
  41. 'task': 'cel_seo_click.selenium_jared_click',
  42. # 'schedule': crontab(minute='*/1'),
  43. 'schedule': 100.0,
  44. 'args': (),
  45. 'options':{'queue': qname}
  46. },
  47. }
  48. #@app.on_after_configure.connect
  49. #def setup_periodic_tasks(sender, **kwargs):
  50. # print(app.request.delivery_info['routing_key'])
  51. # print(json_str)
  52. #sender task.request.hostnam
  53. # sender.add_periodic_task(100.0, selenium_jared_click.s(), name='add every 2 min')
  54. @app.task()
  55. def selenium_jared_click():
  56. global custname
  57. db = dataset.connect('postgresql://postgres:eyJhbGciOiJI@172.105.241.163:5432/postgres')
  58. options = webdriver.ChromeOptions()
  59. options.add_argument("--no-sandbox")
  60. options.add_argument("--headless")
  61. options.add_argument("--incognito")
  62. #options.add_argument('--blink-settings=imagesEnabled=false')
  63. mobile_emulation = {
  64. "deviceMetrics": { "width": 360, "height": 640, "pixelRatio": 3.0 },
  65. "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" }
  66. try:
  67. driver = webdriver.Chrome(options=options)
  68. # driver = webdriver.Chrome(ChromeDriverManager().install(),options=options)
  69. ## Create Undetected Chromedriver with Options
  70. # driver = uc.Chrome(options=options)
  71. # options = uc.ChromeOptions()
  72. except:
  73. traceback.print_exc()
  74. #kw='真理大學國際生'
  75. #domain='cia.au.edu.tw'
  76. if custname is None:
  77. custname='真理'
  78. cursor=db.query("SELECT cust,plan,prefix,domain,kw,positive FROM public.seo_jobs where cust='"+custname+"' order by random() limit 1")
  79. kw=None
  80. domain=None
  81. for c in cursor:
  82. cust=c['cust']
  83. kw=c['kw']
  84. plan=c['plan']
  85. prefix=c['prefix']
  86. domain=eval(c['domain'])[0]
  87. positive=eval(c['positive'])
  88. break
  89. #kw='真理大學教堂'
  90. #domain='udn.com'
  91. print(kw)
  92. print(domain)
  93. #kw='真理大學校友'
  94. #kw='真理大學國際生'
  95. #kw='真理大學張聰聯'
  96. #domain='au.edu.tw'
  97. #domain='pronews.tw'
  98. driver.get('https://www.google.com?num=100')
  99. time.sleep(3)
  100. print(driver.current_url)
  101. elmt = driver.find_element(By.XPATH, "//textarea[@name='q']")
  102. time.sleep(1)
  103. elmt.send_keys(kw)
  104. elmt.send_keys(Keys.ENTER)
  105. time.sleep(4)
  106. elmts = driver.find_elements(By.XPATH, "//a[@jsname='UWckNb']")
  107. numresults=len(elmts)
  108. print('搜尋結果數量',numresults)
  109. if numresults<=0:
  110. print(elmts)
  111. for elmt in elmts:
  112. href=elmt.get_attribute('href')
  113. txt=elmt.text
  114. # print(txt)
  115. # print(href)
  116. if domain in href:
  117. webdriver.ActionChains(driver).move_to_element(elmt).perform()
  118. webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
  119. print(domain)
  120. print(href)
  121. driver.quit()
  122. return txt
  123. break
  124. return '{empty}'
  125. # time.sleep(5)