爬虫请求库——selenium
selenium模块
selenium 最初是一个自动化测试工具,而爬虫中使用它主要是为了解决 requests 无法直接执行 JavaScript 代码的问题。selenium 的缺点是效率会变得很慢。
selenium 本质是通过驱动浏览器，完全模拟浏览器的操作，比如跳转、输入、点击、下拉等，来拿到网页渲染之后的结果，可支持多种浏览器
from selenium import webdriver # 主流浏览器 browser = webdriver.Chrome() browser = webdriver.Firefox() browser = webdriver.PhantomJS() browser = webdriver.Safari() browser = webdriver.Edge() # 其他国产浏览器，例如360 chrome_options = webdriver.ChromeOptions() chrome_options.binary_location = r"C:\Users\kevin\AppData\Roaming\360se6\Application\360se.exe" # 这里是360安全浏览器的路径 chrome_options.add_argument(r'--lang=zh-CN') # 这里添加一些启动的参数 browser = webdriver.Chrome(chrome_options=chrome_options)
一、安装
安装：
pip3 install selenium
要自动启动浏览器需要安装相应的驱动，selenium3 默认支持的 webdriver 是 Firfox，而 Firefox 需要安装 geckodriver
driver 安装
selenium 需要调用 driver 来控制浏览器，所以要先安装 driver。不同的浏览器使用不同的 driver，并且需要根据浏览器的版本下载对应版本的 driver。
火狐浏览器安装 driver
下载链接：https://github.com/mozilla/geckodriver/releases
- Linux 下载后将 geckodriver 移动到 /usr/local/bin/ 目录下（或者软连接到 /usr/bin/chromdriver）
- Windows下载后将 geckodriver.exe 加入环境变量，或者将 geckodriver.exe 移动到脚本同级目录
谷歌浏览器安装 driver：
- windows：下载 chromdriver.exe 放到python安装路径的 scripts 目录中即可，注意最新版本是2.xx，并非2.9
- Linux：下载 chromdriver.zip，解压缩后将 chromdriver 文件复制到 /usr/bin 文件夹下（或者软连接到 /usr/bin/chromdriver）
国内镜像网站地址：http://npm.taobao.org/mirrors/chromedriver/ 或者 https://chromedriver.storage.googleapis.com/index.html
最新的版本去官网找:https://sites.google.com/a/chromium.org/chromedriver/downloads
其他浏览器安装 driver
其他国产浏览器（如360）的内核大部分都是 Chrome，所以只要下载对应版本的谷歌浏览器的 driver 即可。可能有极个别是 IE 内核，原理也是一样，下载 IE 的 driver，回头再聊。
from selenium import webdriver driver = webdriver.Chrome() driver.get('https://www.baidu.com') driver.page_source #能自动弹出浏览器并访问百度，就说明成功了
from selenium import webdriver brow=webdriver.Firefox() brow.get('https://www.baidu.com')
from selenium import webdriver chrome_options = webdriver.ChromeOptions() # 指定360安全浏览器所在路径 chrome_options.binary_location = r"C:\Users\kevin\AppData\Roaming\360se6\Application\360se.exe" chrome_options.add_argument(r'--lang=zh-CN') # 这里添加一些启动的参数 browser = webdriver.Chrome(chrome_options=chrome_options)
#安装：selenium+phantomjs pip3 install selenium 下载phantomjs，解压后把phantomjs.exe所在的bin目录放到环境变量 下载链接：http://phantomjs.org/download.html #验证安装 C:\Users\Administrator>phantomjs phantomjs> console.log('egon gaga') egon gaga undefined phantomjs> ^C C:\Users\Administrator>python3 Python 3.6.1 (v3.6.1:69c0db5, Mar 21 2017, 18:41:36) [MSC v.1900 64 bit (AMD64)] on win32 Type "help", "copyright", "credits" or "license" for more information. >>> from selenium import webdriver >>> driver=webdriver.PhantomJS() #无界面浏览器 >>> driver.get('https://www.baidu.com') >>> driver.page_source
二、等待元素被加载
selenium 是自动运行浏览器的一种行为，而浏览器的加载是需要时间的，我们在进行操作之前肯定必须要等到被操作的元素加载出来后才行，所以我们就需要用到等待的操作、
等待的方式分两种
#2、等待的方式分两种： 隐式等待：在browser.get（'xxx'）前就设置，针对所有元素有效 显式等待：在browser.get（'xxx'）之后设置，只针对某个元素有效
示例：
from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 browser=webdriver.Chrome()#启动浏览器 """ 方式一 隐式等待:在查找所有元素时，如果尚未被加载，则等10秒，在最前面设置 """ browser.implicitly_wait(10) browser.get('https://www.baidu.com') input_tag=browser.find_element_by_id('kw') input_tag.send_keys('美女') input_tag.send_keys(Keys.ENTER) """ 方式二 显式等待：显式地等待某个指定元素被加载 这两条效果与上面一条相同 """ #wait=WebDriverWait(browser,10) #wait.until(EC.presence_of_element_located((By.ID,'content_left'))) #查找内容 contents=browser.find_element_by_id('content_left') #没有进行等待操作就直接查找的话，找不到就会报错 print(contents) browser.close()
三、选择器
选择器就是用来做对象爬取的内容进行定位用的，选择器有很多，每种的功能都不同
基本选择器
#官网链接：http://selenium-python.readthedocs.io/locating-elements.html from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 import time driver=webdriver.Chrome() driver.get('https://www.baidu.com') wait=WebDriverWait(driver,10) try: #===============所有方法=================== # 1、find_element_by_id # 2、find_element_by_link_text # 3、find_element_by_partial_link_text #模糊匹配 # 4、find_element_by_tag_name # 5、find_element_by_class_name # 6、find_element_by_name # 7、find_element_by_css_selector # 8、find_element_by_xpath #后续详解 # 强调： # 1、上述均可以改写成find_element(By.ID,'kw')的形式 # 2、find_elements_by_xxx的形式是查找到多个元素，结果为列表 #===============示范用法=================== # 1、find_element_by_id print(driver.find_element_by_id('kw')) # 2、find_element_by_link_text # login=driver.find_element_by_link_text('登录') # login.click() # 3、find_element_by_partial_link_text login=driver.find_elements_by_partial_link_text('录')[0] login.click() # 4、find_element_by_tag_name print(driver.find_element_by_tag_name('a')) # 5、find_element_by_class_name button=wait.until(EC.element_to_be_clickable((By.CLASS_NAME,'tang-pass-footerBarULogin'))) button.click() # 6、find_element_by_name input_user=wait.until(EC.presence_of_element_located((By.NAME,'userName'))) input_pwd=wait.until(EC.presence_of_element_located((By.NAME,'password'))) commit=wait.until(EC.element_to_be_clickable((By.ID,'TANGRAM__PSP_10__submit'))) input_user.send_keys('18611453110') input_pwd.send_keys('lhf@094573') commit.click() # 7、find_element_by_css_selector driver.find_element_by_css_selector('#kw') # 8、find_element_by_xpath time.sleep(5) finally: driver.close()
xpath
#官网链接：http://selenium-python.readthedocs.io/locating-elements.html from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 import time driver=webdriver.PhantomJS() driver.get('https://doc.scrapy.org/en/latest/_static/selectors-sample1.html') # wait=WebDriverWait(driver,3) driver.implicitly_wait(3) #使用隐式等待 try: # find_element_by_xpath #//与/的不同 # driver.find_element_by_xpath('//body/a') # 开头的//代表从整篇文档中寻找,body之后的/代表body的儿子，这一行找不到就会报错了 driver.find_element_by_xpath('//body//a') # 开头的//代表从整篇文档中寻找,body之后的//代表body的子子孙孙 driver.find_element_by_css_selector('body a') #取第n个 res1=driver.find_elements_by_xpath('//body//a[1]') #取第一个a标签，这里的索引是从1开始 print(res1[0].text) #按照属性查找,下述三者查找效果一样 res1=driver.find_element_by_xpath('//a[5]') res2=driver.find_element_by_xpath('//a[@href="image5.html"]') res3=driver.find_element_by_xpath('//a[contains(@href,"image5")]') #模糊查找 print('==>', res1.text) print('==>',res2.text) print('==>',res3.text) #其他 res1=driver.find_element_by_xpath('/html/body/div/a') print(res1.text) res2=driver.find_element_by_xpath('//a[img/@src="image3_thumb.jpg"]') #找到子标签img的src属性为image3_thumb.jpg的a标签 print(res2.tag_name,res2.text) res3 = driver.find_element_by_xpath("//input[@name='continue'][@type='button']") #查看属性name为continue且属性type为button的input标签 res4 = driver.find_element_by_xpath("//*[@name='continue'][@type='button']") #查看属性name为continue且属性type为button的所有标签 time.sleep(5) finally: driver.close()
获取标签属性
from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 browser=webdriver.Chrome() browser.get('https://www.amazon.cn/') wait=WebDriverWait(browser,10) wait.until(EC.presence_of_element_located((By.ID,'cc-lm-tcgShowImgContainer'))) tag=browser.find_element(By.CSS_SELECTOR,'#cc-lm-tcgShowImgContainer img') #获取标签属性， print(tag.get_attribute('src')) #获取标签ID，位置，名称，大小（了解） print(tag.id) print(tag.location) print(tag.tag_name) print(tag.size) browser.close()
四、元素交互操作
交互操作就是指模拟人在网页中进行输入或点击鼠标的操作，主要针对的是 input 框和链接。交互操作中可以自己植入 js 代码
模拟天猫搜索：
from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 import time driver=webdriver.Chrome() driver.get('https://www.tmall.com/') driver.implicitly_wait(3) try: input_tag=driver.find_element_by_id('mq')#搜索框 input_tag.send_keys('NIKE')#写入搜索条件 input_tag.send_keys(Keys.ENTER)#回车 time.sleep(3) input_tag=driver.find_element_by_id('mq') input_tag.clear()#清空搜索框 input_tag.send_keys('科比战靴') input_tag.send_keys(Keys.ENTER) time.sleep(5) finally: driver.close()
模拟滑动验证码操作（网页很low逼）：
from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 import time driver=webdriver.Chrome() driver.get('http://www.runoob.com/try/try.php?filename=jqueryui-api-droppable') #driver.get('http://www.baidu.com') driver.implicitly_wait(3) try: driver.switch_to.frame('iframeResult')#已被弃用的html标签，作用是html下套html driver.switch_to.parent_frame()#返回父级html标签 source=driver.find_element_by_id('draggable')#源滑块 target=driver.find_element_by_id('droppable')#目标 print(source,target) #方式一： # actions=ActionChains(driver) # actions.drag_and_drop(source,target) # actions.perform() #方式二： distance=target.location['x']-source.location['x']#距离 ActionChains(driver).click_and_hold(source).perform()#模拟按住鼠标不放手 print(distance) s=0 while s < distance: print(s) ActionChains(driver).move_by_offset(xoffset=2,yoffset=0).perform()#每次移动2像素 s+=2 ActionChains(driver).release().perform()#释放鼠标 driver.execute_script('alert("xxxxxxxxx")')#可自写js代码，会执行 time.sleep(6) finally: driver.close()
补充：frame的切换
#frame相当于一个单独的网页，在父frame里是无法直接查看到子frame的元素的，必须switch_to_frame切到该frame下，才能进一步查找 from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 try: browser=webdriver.Chrome() browser.get('http://www.runoob.com/try/try.php?filename=jqueryui-api-droppable') browser.switch_to.frame('iframeResult') #切换到id为iframeResult的frame tag1=browser.find_element_by_id('droppable') print(tag1) # tag2=browser.find_element_by_id('textareaCode') #报错，在子frame里无法查看到父frame的元素 browser.switch_to.parent_frame() #切回父frame,就可以查找到了 tag2=browser.find_element_by_id('textareaCode') print(tag2) finally: browser.close()
模拟浏览器的前进后退
#模拟浏览器的前进后退 import time from selenium import webdriver browser=webdriver.Chrome() browser.get('https://www.baidu.com') browser.get('https://www.taobao.com') browser.get('http://www.sina.com.cn/') browser.back()#后腿 time.sleep(10) browser.forward()#前进 browser.close()
#cookies from selenium import webdriver browser=webdriver.Chrome() browser.get('https://www.zhihu.com/explore') print(browser.get_cookies()) browser.add_cookie({'k1':'xxx','k2':'yyy'}) print(browser.get_cookies()) # browser.delete_all_cookies()
选项卡管理
#选项卡管理：切换选项卡，有js的方式windows.open,有windows快捷键：ctrl+t等，最通用的就是js的方式 import time from selenium import webdriver browser=webdriver.Chrome() browser.get('https://www.baidu.com') browser.execute_script('window.open()') print(browser.window_handles) #获取所有的选项卡 browser.switch_to_window(browser.window_handles[1]) browser.get('https://www.taobao.com') time.sleep(10) browser.switch_to_window(browser.window_handles[0]) browser.get('https://www.sina.com.cn') browser.close()
异常处理
from selenium import webdriver from selenium.common.exceptions import TimeoutException,NoSuchElementException,NoSuchFrameException try: browser=webdriver.Chrome() browser.get('http://www.runoob.com/try/try.php?filename=jqueryui-api-droppable') browser.switch_to.frame('iframssseResult') except TimeoutException as e: print(e) except NoSuchFrameException as e: print(e) finally: browser.close()
小项目扩展
#注意：网站都策略都是在不断变化的，精髓在于学习流程。下述代码生效与2017-11-7，不能保证永久有效 from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By from selenium.webdriver.common.keys import Keys from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait browser=webdriver.Chrome() try: browser.get('http://mail.163.com/') wait=WebDriverWait(browser,5) frame=wait.until(EC.presence_of_element_located((By.ID,'x-URS-iframe'))) browser.switch_to.frame(frame) wait.until(EC.presence_of_element_located((By.CSS_SELECTOR,'.m-container'))) inp_user=browser.find_element_by_name('email') inp_pwd=browser.find_element_by_name('password') button=browser.find_element_by_id('dologin') inp_user.send_keys('18611453110') inp_pwd.send_keys('xxxx') button.click() #如果遇到验证码，可以把下面一小段打开注释 # import time # time.sleep(10) # button = browser.find_element_by_id('dologin') # button.click() wait.until(EC.presence_of_element_located((By.ID,'dvNavTop'))) write_msg=browser.find_elements_by_css_selector('#dvNavTop li')[1] #获取第二个li标签就是“写信”了 write_msg.click() wait.until(EC.presence_of_element_located((By.CLASS_NAME,'tH0'))) recv_man=browser.find_element_by_class_name('nui-editableAddr-ipt') title=browser.find_element_by_css_selector('.dG0 .nui-ipt-input') recv_man.send_keys('378533872@qq.com') title.send_keys('圣旨') print(title.tag_name) frame=wait.until(EC.presence_of_element_located((By.CLASS_NAME,'APP-editor-iframe'))) browser.switch_to.frame(frame) body=browser.find_element(By.CSS_SELECTOR,'body') body.send_keys('egon很帅，可以加工资了') browser.switch_to.parent_frame() #切回他爹 send_button=browser.find_element_by_class_name('nui-toolbar-item') send_button.click() #可以睡时间久一点别让浏览器关掉，看看发送成功没有 import time time.sleep(10000) except Exception as e: print(e) finally: browser.close()
from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 import time def get_goods(driver): try: goods=driver.find_elements_by_class_name('gl-item') for good in goods: detail_url=good.find_element_by_tag_name('a').get_attribute('href') p_name=good.find_element_by_css_selector('.p-name em').text.replace('\n','') price=good.find_element_by_css_selector('.p-price i').text p_commit=good.find_element_by_css_selector('.p-commit a').text msg = ''' 商品 : %s 链接 : %s 价钱 ：%s 评论 ：%s ''' % (p_name,detail_url,price,p_commit) print(msg,end='\n\n') button=driver.find_element_by_partial_link_text('下一页') button.click() time.sleep(1) get_goods(driver) except Exception: pass def spider(url,keyword): driver = webdriver.Chrome() driver.get(url) driver.implicitly_wait(3) # 使用隐式等待 try: input_tag=driver.find_element_by_id('key') input_tag.send_keys(keyword) input_tag.send_keys(Keys.ENTER) get_goods(driver) finally: driver.close() if __name__ == '__main__': spider('https://www.jd.com/',keyword='科比战靴')
#首先要安装Pillow pip3 install pillow #Pillow:基于PIL，处理python 3.x的图形图像库.因为PIL只能处理到python 2.x，而这个模块能处理Python3.x，目前用它做图形的很多. ###########思路整理########## from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 from PIL import Image #pip3 install pillow import time def get_snap(driver): driver.save_screenshot('snap.png')#截图 snap_obj=Image.open('snap.png')#保存 return snap_obj def get_image(driver): img=driver.find_element_by_class_name('geetest_canvas_img') time.sleep(2) #等待图片加载完毕 size=img.size location=img.location #获取图片位置 left=location['x'] top=location['y'] right=left+size['width'] bottom=top+size['height'] snap_obj=get_snap(driver) image_obj=snap_obj.crop((left,top,right,bottom))#截图操作 # image_obj.show() return image_obj def get_distance(image1,image2): start_x=58#滑块最左侧 threhold=60#去除伪影响 # print(image1.size) # print(image2.size) for x in range(start_x,image1.size[0]): for y in range(image1.size[1]): rgb1=image1.load()[x,y] rgb2=image2.load()[x,y] res1=abs(rgb1[0]-rgb2[0]) res2=abs(rgb1[1]-rgb2[1]) res3=abs(rgb1[2]-rgb2[2]) if not (res1 < threhold and res2 < threhold and res3 < threhold): return x-7#误差范围 def get_tracks(distance): distance+=20#故意划过头20像素 #v=v0+a*t #s=v*t+0.5*a*(t**2) v0=0 s=0 t=0.2 mid=distance*3/5 forward_tracks=[] while s < distance: if s < mid: a=2 else: a=-3 v=v0 track=v*t+0.5*a*(t**2) track=round(track)#取整数 v0=v+a*t s+=track forward_tracks.append(track) back_tracks=[-1,-1,-1,-2,-2,-2,-3,-3,-2,-2,-1] #20 return {"forward_tracks":forward_tracks,'back_tracks':back_tracks} try: driver = webdriver.Chrome()#谷歌浏览器 driver.get('https://passport.cnblogs.com/user/signin')#博客园 driver.implicitly_wait(10)#隐形等待10秒 #1、输入账号、密码，然后点击登陆 input_user=driver.find_element_by_id('input1') input_pwd=driver.find_element_by_id('input2') login_button=driver.find_element_by_id('signin') input_user.send_keys('wall-a')#输入账号 input_pwd.send_keys('lg19950726..')#输入密码 login_button.click()#点击登录按钮 #2、点击验证人机按钮，弹出没有缺口的图 button=driver.find_element_by_class_name('geetest_radar_tip_content') button.click() #3、针对没有缺口的图片进行截图 image1=get_image(driver) #4、点击滑动按钮，弹出有缺口的图 slider_button=driver.find_element_by_class_name('geetest_slider_button') slider_button.click() #5、针对有缺口的图片进行截图 image2=get_image(driver) #6、对比两张图片，找出缺口，即滑动的位移 distance=get_distance(image1,image2) # print(distance) #7、按照人的行为行为习惯，把总位移切成一段段小的位移 traks_dic=get_tracks(distance) #8、按照位移移动 slider_button=driver.find_element_by_class_name('geetest_slider_button') ActionChains(driver).click_and_hold(slider_button).perform()#按住不放手 #先向前移动 forward_tracks=traks_dic["forward_tracks"] back_tracks=traks_dic["back_tracks"] for forward_track in forward_tracks: ActionChains(driver).move_by_offset(xoffset=forward_track,yoffset=0).perform() #短暂停顿，发现傻逼，移过了 time.sleep(0.2) # 先向后移动 for back_track in back_tracks: ActionChains(driver).move_by_offset(xoffset=back_track,yoffset=0).perform() # 抖一抖 ActionChains(driver).move_by_offset(xoffset=-4,yoffset=0).perform() ActionChains(driver).move_by_offset(xoffset=3,yoffset=0).perform() time.sleep(0.1) ActionChains(driver).move_by_offset(xoffset=-2,yoffset=0).perform() ActionChains(driver).move_by_offset(xoffset=3,yoffset=0).perform() time.sleep(0.3) ActionChains(driver).release().perform()#松开鼠标 time.sleep(10) finally: driver.close() ###############优化后的代码（将功能封装成函数调用）####### from selenium import webdriver from selenium.webdriver import ActionChains from selenium.webdriver.common.by import By #按照什么方式查找，By.ID,By.CSS_SELECTOR from selenium.webdriver.common.keys import Keys #键盘按键操作 from selenium.webdriver.support import expected_conditions as EC from selenium.webdriver.support.wait import WebDriverWait #等待页面加载某些元素 from PIL import Image #pip3 install pillow import time def get_snap(driver): driver.save_screenshot('snap.png')#截图 snap_obj=Image.open('snap.png')#保存 return snap_obj def get_image(driver): img=driver.find_element_by_class_name('geetest_canvas_img') time.sleep(2) #等待图片加载完毕 size=img.size location=img.location #获取图片位置 left=location['x'] top=location['y'] right=left+size['width'] bottom=top+size['height'] snap_obj=get_snap(driver) image_obj=snap_obj.crop((left,top,right,bottom))#截图操作 # image_obj.show() return image_obj def get_distance(image1,image2): start_x=58#滑块最左侧 threhold=60#去除伪影响 # print(image1.size) # print(image2.size) for x in range(start_x,image1.size[0]): for y in range(image1.size[1]): rgb1=image1.load()[x,y] rgb2=image2.load()[x,y] res1=abs(rgb1[0]-rgb2[0]) res2=abs(rgb1[1]-rgb2[1]) res3=abs(rgb1[2]-rgb2[2]) if not (res1 < threhold and res2 < threhold and res3 < threhold): return x-7#误差范围 def get_tracks(distance): distance+=20#故意划过头20像素 #v=v0+a*t #s=v*t+0.5*a*(t**2) v0=0 s=0 t=0.2 mid=distance*3/5 forward_tracks=[] while s < distance: if s < mid: a=2 else: a=-3 v=v0 track=v*t+0.5*a*(t**2) track=round(track)#取整数 v0=v+a*t s+=track forward_tracks.append(track) back_tracks=[-1,-1,-1,-2,-2,-2,-3,-3,-2,-2,-1] #20 return {"forward_tracks":forward_tracks,'back_tracks':back_tracks} def crack(driver):#封装滑动的函数 # 2、点击验证人机按钮，弹出没有缺口的图 button = driver.find_element_by_class_name('geetest_radar_tip_content') button.click() # 3、针对没有缺口的图片进行截图 image1 = get_image(driver) # 4、点击滑动按钮，弹出有缺口的图 slider_button = driver.find_element_by_class_name('geetest_slider_button') slider_button.click() # 5、针对有缺口的图片进行截图 image2 = get_image(driver) # 6、对比两张图片，找出缺口，即滑动的位移 distance = get_distance(image1, image2) # print(distance) # 7、按照人的行为行为习惯，把总位移切成一段段小的位移 traks_dic = get_tracks(distance) # 8、按照位移移动 slider_button = driver.find_element_by_class_name('geetest_slider_button') ActionChains(driver).click_and_hold(slider_button).perform() # 按住不放手 # 先向前移动 forward_tracks = traks_dic["forward_tracks"] back_tracks = traks_dic["back_tracks"] for forward_track in forward_tracks: ActionChains(driver).move_by_offset(xoffset=forward_track, yoffset=0).perform() # 短暂停顿，发现傻逼，移过了 time.sleep(0.2) # 先向后移动 for back_track in back_tracks: ActionChains(driver).move_by_offset(xoffset=back_track, yoffset=0).perform() # 抖一抖 ActionChains(driver).move_by_offset(xoffset=-4, yoffset=0).perform() ActionChains(driver).move_by_offset(xoffset=3, yoffset=0).perform() time.sleep(0.1) ActionChains(driver).move_by_offset(xoffset=-2, yoffset=0).perform() ActionChains(driver).move_by_offset(xoffset=3, yoffset=0).perform() time.sleep(0.3) ActionChains(driver).release().perform() # 松开鼠标 def login_cnblogs(username,pwd): driver = webdriver.Chrome() # 谷歌浏览器driver = webdriver.Chrome()#谷歌浏览器 try: driver.get('https://passport.cnblogs.com/user/signin')#博客园 driver.implicitly_wait(10)#隐形等待10秒 #1、输入账号、密码，然后点击登陆 input_user=driver.find_element_by_id('input1') input_pwd=driver.find_element_by_id('input2') login_button=driver.find_element_by_id('signin') input_user.send_keys(username)#输入账号 input_pwd.send_keys(pwd)#输入密码 login_button.click()#点击登录按钮 # 调用 封装滑动的函数 crack(driver) time.sleep(10) finally: driver.close() if __name__ == '__main__': login_cnblogs(username='Angelababy',pwd='sonoface')
#-*- coding:utf-8 -*- from selenium import webdriver from selenium.webdriver.support.wait import WebDriverWait from selenium.webdriver.common.by import By from selenium.webdriver.support import expected_conditions as EC import time import datetime import random from threading import Thread class Driver: def __init__(self, _webdriver, brow_name): self.brow = _webdriver self.name = brow_name class OP: def __init__(self, drivers ,start_page=1, end_page=11): self.start_page = start_page self.end_page = end_page self.drivers = [] for driver in drivers: self.drivers.append(driver) def my_print(self, *args): print(datetime.datetime.now(), args) def my_sleep(self): time.sleep(random.randint(1, 5)) def main(self, _driver): driver = _driver.brow() # 弹出浏览器 for i in range(self.start_page, self.end_page): if _driver.name == 'Safari': for j in range(0, 22 if i == 1 else 20): # driver.implicitly_wait(10) url = 'https://www.cnblogs.com/zhuminghui/default.html?page=' + str(i) self.my_print(url) driver.get(url) self.inner(driver, i, j, _driver.name) self.my_sleep() else: url = 'https://www.cnblogs.com/zhuminghui/default.html?page=' + str(i) self.my_print(url) driver.get(url) for j in range(0,22 if i == 1 else 20): self.inner(driver, i, j, _driver.name) # wait = WebDriverWait(driver, 5) # wait.until(EC.presence_of_element_located((By.CSS_SELECTOR, '#homepage1_HomePageDays_ctl00_ImageLink'))) self.my_sleep() # readmore_list = driver.find_elements_by_class_name('c_b_p_desc_readmore') driver.close() def inner(self, driver, i, j, name): try: postTitle2 = driver.find_elements_by_class_name('postTitle2') if len(postTitle2) < j + 1: self.my_print(name, j + 1, 'end') return if 'adowsocks' in postTitle2[j].text or '****' in postTitle2[j].text: self.my_print('---<<< %s >>> page %s ，line %s <%s> ---' % (name, i, j + 1, postTitle2[j].text)) return postTitle2[j].click() self.my_sleep() wait = WebDriverWait(driver, 5) wait.until(EC.presence_of_element_located((By.CSS_SELECTOR, '#blog_post_info_block'))) driver.execute_script("arguments[0].scrollIntoView();", driver.find_element_by_id('blog_post_info_block')) title = driver.find_element_by_class_name('postTitle2') self.my_print('<<< %s >>>page %s ，line %s <%s>' % (name, i, j + 1, title.text)) self.my_sleep() driver.back() return True except Exception as e: self.my_print(name, 'Error：', e) driver.back() self.my_sleep() def start(self): ts = [] for driver in self.drivers: ts.append(Thread(target=self.main, args=(driver,))) for i in ts: i.start() for i in ts: i.join() if __name__ == '__main__': chrome = Driver(webdriver.Chrome, 'Chrome') firefox = Driver(webdriver.Firefox, 'Firefox') safari = Driver(webdriver.Safari, 'Safari') obj = OP(drivers=[chrome, firefox], start_page=1, end_page=13) obj.start()
