123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317 |
- from selenium import webdriver
- from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
- import time
- import os
- import urllib.parse
- from selenium.webdriver.support.ui import WebDriverWait
- from selenium.webdriver.common.by import By
- from selenium.webdriver.support import expected_conditions as EC
- import codecs
- import random
- from bs4 import BeautifulSoup
- import requests
- import time
- import rpyc
- import sys
- import docker
- headers = {
- "Authorization": "Bearer " + "t35vhZtWNgvDNWHc3DJh0OKll3mcB9GvC8K2EAkBug2",
- "Content-Type": "application/x-www-form-urlencoded"
- }
- def send_msg(kw):
- params = {"message": "處理關鍵字: "+kw}
- r = requests.post("https://notify-api.line.me/api/notify",headers=headers, params=params)
- def empty_query(q):
- global driver
- googleurl='https://www.google.com/search?q='+urllib.parse.quote(q)
- driver.get(googleurl)
- time.sleep(3)
- def process_query(q):
- global driver
- googleurl='https://www.google.com/search?q='+urllib.parse.quote(q)
- driver.get(googleurl)
- time.sleep(3)
- elmts=driver.find_elements_by_xpath("//div[@class='g']//div[@class='yuRUbf']//a")
- idx=1
- ranking=-1
- for elmt in elmts:
- href=elmt.get_attribute('href')
- txt=elmt.text
- if len(txt)>10:
- if 'hhh.com.tw' in href:
- webdriver.ActionChains(driver).move_to_element(elmt).perform()
- webdriver.ActionChains(driver).move_to_element(elmt).click().perform()
- break
- result=[]
- result.append({'company':'上築','designers':['魏子涵'],'fullname':'上築空間設計'})
- result.append({'company':'摩登雅舍','designers':['王思文','汪忠錠'],'fullname':'摩登雅舍室內裝修設計'})
- result.append({'company':'亞維','designers':['簡瑋琪'],'fullname':'亞維空間設計'})
- result.append({'company':'比沙列','designers':['張靜峰'],'fullname':'比沙列室內裝修設計'})
- result.append({'company':'京璽國際','designers':['周彥如'],'fullname':'京璽國際'})
- result.append({'company':'京築聯合','designers':['周彥如'],'fullname':'京築聯合室內裝修'})
- result.append({'company':'浩室設計','designers':['浩室設計'],'fullname':'浩室設計'})
- result.append({'company':'IS國際設計','designers':['陳嘉鴻'],'fullname':'IS國際設計'})
- result.append({'company':'豐聚設計','designers':['黃翊峰','李羽芝'],'fullname':'豐聚室內裝修設計'})
- result.append({'company':'優尼客','designers':['黃仲均'],'fullname':'優尼客空間設計'})
- result.append({'company':'富億設計','designers':['陳錦樹'],'fullname':'富億空間設計'})
- result.append({'company':'三宅一秀','designers':['郁琇琇','天使總監郁琇琇'],'fullname':'三宅一秀空間創藝有限公司'})
- result.append({'company':'元均','designers':['馬愷君','設計總監馬愷君'],'fullname':'元均制作空間設計'})
- result.append({'company':'格綸設計','designers':['馬愷君','設計總監馬愷君'],'fullname':'格綸設計工程'})
- result.append({'company':'雅典設計','designers':['雅典設計'],'fullname':'雅典設計工程'})
- result.append({'company':'雲方設計','designers':['潘仕敏'],'fullname':'雲方室內裝修設計'})
- result.append({'company':'采品設計','designers':['盧慧珊'],'fullname':'采品室內設計'})
- result.append({'company':'恆岳設計','designers':['蔡岳儒'],'fullname':'恆岳空間設計'})
- result.append({'company':'唐林設計','designers':['廖韋強'],'fullname':'唐林建築室內設計'})
- result.append({'company':'恆岳設計','designers':['蔡岳儒'],'fullname':'恆岳空間設計'})
- result.append({'company':'澤序設計','designers':['張于廷'],'fullname':'澤序空間設計'})
- result.append({'company':'允庭設計','designers':['張舜淵','林術榮','李晴沛'],'fullname':'允庭室內裝修設計'})
- result.append({'company':'禾亞國際','designers':['李柔瑩'],'fullname':'禾亞國際室內裝修'})
- result.append({'company':'禾築設計','designers':['譚淑靜'],'fullname':'禾築國際設計'})
- result.append({'company':'我思空間','designers':['陳佳佑'],'fullname':'我思空間設計'})
- result.append({'company':'沛沛設計','designers':['郭沛沛'],'fullname':'沛沛國際室內設計'})
- qlist=[]
- def gen_long():
- global qlist
- for r in result:
- ds=r['designers']
- fullname=r['fullname']
- abvname=r['company']
- qlist.append(fullname+"設計師")
- qlist.append(fullname+" 設計師")
- qlist.append(fullname+" 設計 師")
- qlist.append(abvname+"設計師")
- qlist.append(abvname+"費用")
- qlist.append(abvname+"評價")
- qlist.append(abvname+" 設計 師")
- print(qlist)
- gen_long()
- options = webdriver.ChromeOptions()
- options2 = webdriver.ChromeOptions()
- options2.add_argument("--disable-infobars")
- options2.add_argument("--enable-file-cookies")
- driver = webdriver.Remote('http://127.0.0.1:4444/wd/hub', desired_capabilities=options2.to_capabilities())
- driver.close()
- driver.session_id='c2f0c1ae40546cb586e06f0deb18d08c'
- driver.set_window_size(1400,1000)
- driver.get('https://www.google.com')
- executor_url = driver.command_executor._url
- session_id = driver.session_id
- print (session_id)
- print (executor_url)
- time.sleep(3)
- fw=codecs.open('c:/tmp/gg.html','w','utf-8')
- fw.write(driver.page_source)
- fw.close()
- time.sleep(9999)
- while True:
- q=random.choice(qlist)
- prob=random.randint(0,2)
- if prob <=1:
-
- print(q)
- process_query(q)
- send_msg(q)
- else:
- empty_query(q)
-
- intsleep=random.randint(5,12)
- time.sleep(intsleep)
|