def moneyMsg(self, ptype): if not self.wdriver: if self.isCmdMode: print 'used phantomjs' import selenium.webdriver.phantomjs.webdriver as wd self.wdriver = wd.WebDriver('/usr/local/bin/phantomjs') #test self.wdriver.maximize_window() else: print 'used chrome' import selenium.webdriver.chrome.webdriver as wd self.wdriver = wd.WebDriver( '/Users/mage/Documents/tool/cmdtool/chromedriver') #test self.wdriver.maximize_window() hurl = '' if ptype == 'btc': hurl = 'http://btcwisdom.net' else: hurl = 'http://ltcwisdom.net' self.wdriver.get(hurl) #难度信息 datdic = self.getDiffcultFromWeb(self.wdriver) return datdic
def runWork(self, isCmdMode=True): self.isCmdMode = isCmdMode if not self.wdriver: if self.isCmdMode: import selenium.webdriver.phantomjs.webdriver as wd self.wdriver = wd.WebDriver('/usr/local/bin/phantomjs') #test self.wdriver.maximize_window() print('used phantomjs') else: # import selenium.webdriver.chrome.webdriver as wd # chrome_options = wd.ChromeOptions() # chrome_options.add_argument('--headless') # chrome_options.add_argument('--disable-gpu') # self.wdriver = wd.WebDriver('/Users/mage/Documents/tool/cmdtool/chromedriver') #test # self.wdriver.maximize_window() from selenium import webdriver chrome_options = webdriver.ChromeOptions() chrome_options.add_argument('--headless') chrome_options.add_argument('--disable-gpu') chrome_options.add_argument( "user-agent='Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/62.0.3202.94 Safari/537.36'" ) self.wdriver = webdriver.Chrome( chrome_options=chrome_options, executable_path= '/Users/mage/Documents/tool/cmdtool/chromedriver') self.wdriver.get(self.mineiotaURL) outdic = {} #企业高管信息 self.startMining(self.wdriver) #获取高管信息
def setUp(self): # setUp is where you instantiate the selenium webdriver and loads the browser. User = get_user_model() User.objects.create_superuser(username='******', password='******', email='*****@*****.**') self.selenium = webdriver.WebDriver() self.selenium.maximize_window() super(AdminTestCase, self).setUp()
def companyMsg(self,tid,lastMd5): if not self.wdriver: if self.isCmdMode: print 'used phantomjs' import selenium.webdriver.phantomjs.webdriver as wd self.wdriver = wd.WebDriver('/usr/local/bin/phantomjs') #test self.wdriver.maximize_window() else: print 'used chrome' import selenium.webdriver.chrome.webdriver as wd self.wdriver = wd.WebDriver('/Users/mage/Documents/tool/cmdtool/chromedriver') #test self.wdriver.maximize_window() hurl = 'http://basic.10jqka.com.cn/%s/news.html'%(tid) self.wdriver.get(hurl) #企业高管信息 datdic = self.getGongGaoTitle(self.wdriver,lastMd5) #获取高管信息 return datdic
def companyMsg(self, tid): if not self.wdriver: if self.isCmdMode: import selenium.webdriver.phantomjs.webdriver as wd self.wdriver = wd.WebDriver('/usr/local/bin/phantomjs') #test self.wdriver.maximize_window() else: import selenium.webdriver.chrome.webdriver as wd self.wdriver = wd.WebDriver( '/Users/mage/Documents/tool/cmdtool/chromedriver') #test self.wdriver.maximize_window() hurl = 'http://basic.10jqka.com.cn/%s/company.html' % (tid) self.wdriver.get(hurl) outdic = {} #企业高管信息 datdic = self.getManager(self.wdriver) #获取高管信息 outdic['ggjs'] = datdic #公司详情 companymsg = self.getCompanyMsg(self.wdriver) # print companymsg outdic['xxqk'] = companymsg #获取发行信息 fxmsg = self.getPublishMsg(self.wdriver) # print fxmsg outdic['fxxg'] = fxmsg #获取参股企业 outmsg = self.getSharesCompanys(self.wdriver) # print outmsg outdic['cgkggs'] = outmsg return outdic
def moneyMsg(self): if not self.wdriver: if self.isCmdMode: print 'used phantomjs' import selenium.webdriver.phantomjs.webdriver as wd self.wdriver = wd.WebDriver('/usr/local/bin/phantomjs') #test self.wdriver.maximize_window() else: print 'used chrome' import selenium.webdriver.chrome.webdriver as wd self.wdriver = wd.WebDriver( '/Users/mage/Documents/tool/cmdtool/chromedriver') #test self.wdriver.maximize_window() self.wdriver.get(self.AllCoinURL) starttime = time.ctime(int(time.time())) #难度信息 datdic = self.getAllCoinFromWeb(self.wdriver) print 'start', starttime print 'end', time.ctime(int(time.time())) return datdic
def runWork(self, isCmdMode=True): self.isCmdMode = isCmdMode if not self.wdriver: if self.isCmdMode: import selenium.webdriver.phantomjs.webdriver as wd self.wdriver = wd.WebDriver('/usr/local/bin/phantomjs') #test self.wdriver.maximize_window() print('used phantomjs') else: # import selenium.webdriver.chrome.webdriver as wd # chrome_options = wd.ChromeOptions() # chrome_options.add_argument('--headless') # chrome_options.add_argument('--disable-gpu') # self.wdriver = wd.WebDriver('/Users/mage/Documents/tool/cmdtool/chromedriver') #test # self.wdriver.maximize_window() # linux上安装chrome要安装这两个库libgconf2-4 libnss3-1d #https://jiayi.space/post/zai-ubuntufu-wu-qi-shang-shi-yong-chrome-headless #chrome drvier下载:https://sites.google.com/a/chromium.org/chromedriver/downloads # from selenium import webdriver # from pyvirtualdisplay import Display # display = Display(visible=0, size=(800, 800)) # display.start() # time.sleep(1) chrome_options = webdriver.ChromeOptions() chrome_options.add_argument('--headless') chrome_options.add_argument('--disable-gpu') chrome_options.add_argument( "user-agent='Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/62.0.3202.94 Safari/537.36'" ) self.wdriver = webdriver.Chrome( chrome_options=chrome_options, executable_path='/root/chrome/chromedriver') print('used chrome') self.wdriver.get(self.mineiotaURL) outdic = {} #企业高管信息 self.startMining(self.wdriver) #获取高管信息
def start(self): self.browser = webdriver.WebDriver()
# 'Accept-Encoding': "gzip, deflate, sdch, br", # 'Accept-Language': "zh-CN,zh;q=0.8,en;q=0.6", # 'Cache-Control': "no-cache", # 'Connection': "keep-alive", # 'Host': "www.baidu.com", # 'Pragma': "no-cache", # 'Upgrade-Insecure-Requests': "1", 'User-Agent': "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.115 Safari/537.36", } # # 设置请求头 for key in headers: webdriver.DesiredCapabilities.PHANTOMJS[ 'phantomjs.page.customHeaders.{}'.format(key)] = headers[key] driver = webdriver.WebDriver(executable_path=phantom_path) driver.get("https://www.baidu.com") # file_handler = open("baidu.html","wb") # file_handler.write(driver.page_source.encode("utf-8")) # file_handler.close() # api 方法 # 根据html的id查询 driver.find_element_by_id() # 根据name属性查询,比如input元素的name driver.find_element_by_name() # 查询一个 driver.find_elements_by_name() # 查询多个 # 根据class定位