-
-
[原创]使用自动化过腾讯滑块和阿里滑块
-
发表于: 2024-12-1 12:44
-
案例一：tx滑块（playwright）
1 | 目标网站：aHR0cHM6Ly93d3cudXJidGl4LmhrL2xvZ2lu |
1、创建Chromium实例（如果不设置为 False，默认是无头模式启动浏览器）
1 2 | async with async_playwright() as p: browser = await p.chromium.launch(headless=False, args=['--start-maximized']) |
2、最大化窗口
1 | context = await browser.new_context(viewport={"width": 1920, "height": 1080}, no_viewport=True) |
3、新建标签页
1 | page = await context.new_page() |
4、加载过检测js
1 | await page.add_init_script(js) #stealth.min.js |
5、监听response事件
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | async def on_response(response): if '/cap_union_new_getcapbysig' in response.url and response.status == 200: #对背景图以及滑块图进行拦截然后保存 if 'img_index=1' in response.url: with open("bg_picture.jpg", "wb") as f: f.write(requests.get(response.url).content) elif 'img_index=0' in response.url: with open("cut_picture.png", "wb") as f: f.write(requests.get(response.url).content) print("response.url:", response.url) if 'cap_union_new_verify' in response.url and response.status == 200: #滑块通过后获取参数 result = await response.text() print("response.url:", response.url,result) page.on('response',on_response) |
6、打开网页、触发滑块
1 2 3 4 5 | await page.goto('aHR0cHM6Ly93d3cudXJidGl4LmhrL2xvZ2lu')await page.wait_for_timeout(1500)await page.click('xpath=//*[@id="root"]/div/div[3]/div/div/div[5]/div/div') await page.wait_for_timeout(500)await page.click('xpath=//*[@id="root"]/div/div[3]/div/div/div[8]/div[2]/div') |
7、识别坐标
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | def get_gap_offset(): """ 识别坐标，滑块的图片需要切割 """ det = ddddocr.DdddOcr(det=False, ocr=False, show_ad=False) img = Image.open('cut_picture.png') region = img.crop((160, 508, 243, 595)) # region.save(f'cut_picture.png') with open('bg_picture.jpg', 'rb') as f: target_bytes = f.read() with open('cut_picture.png', 'rb') as f: background_bytes = f.read() res = det.slide_match(target_bytes, background_bytes, simple_target=True) print("识别到的坐标位置：", res) distance = int(res['target'][0]) return distance |
8、找到滑动起始点，并滑动
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 | async def move_down(page): #定位iframe new_frame = page.frame_locator('iframe[id="tcaptcha_iframe_dy"]') #定位起始点 move_tag = new_frame.locator('xpath=//*[@id="tcOperation"]/div[6]') #找到这个元素在当前页面的坐标 box = await move_tag.bounding_box() print("目前点击的位置",box) # 讲鼠标移动到到其实元素的中心 await page.mouse.move(box["x"] + box["width"] / 2, box["y"] + box["height"] / 2) # 按下鼠标 await page.mouse.down() #延时1.2s await page.wait_for_timeout(1200) # 这里获取到x坐标中心点位置 x = box["x"] + box["width"] / 2 #识别到坐标后与网页上的比例 distance = int(get_gap_offset()/1.97)-30 #轨迹 move_distance = get_track_list(distance) print("最终坐标：",distance,"轨迹：",move_distance) for i in move_distance: x += i await page.mouse.move(x, box["y"]) await page.mouse.up() |
9、关闭窗口
1 | await browser.close() |
至此tx滑块的分析就结束了
然后我还写了一版selenium的，相比于playwright就会麻烦一些
案例一：tx滑块（selenium）
1 2 3 4 5 6 7 8 9 10 11 12 13 | 对于我们日常使用而言两者主要区别在于：1、selenium只支持同步，playwright可以支持异步的2、操作iframe,selenium来回切换iframe非常麻烦，而playwright只需要定位元素即可2、在监听请求这一点上，playwright的page.on非常好用，而selenium一般是借助browsermobproxy通过代理的方式进行拦截 使用方式：(1)https://github.com/lightbody/browsermob-proxy/releases，下载并解压 (2)安装证书，参考链接https://4e6K9s2c8@1M7q4)9K6b7g2)9J5c8W2)9J5c8Y4N6%4N6#2)9J5k6h3u0A6L8r3W2T1K9h3I4A6i4K6u0W2j5$3!0E0/read/cv21263644/ (3)调用方式 server = Server('browsermob-proxy-2.1.4/bin/browsermob-proxy') server.start() proxy = server.create_proxy(params={'trustAllServers':'true'}) option = ChromeOptions() option.add_argument('--proxy-server={0}'.format(self.proxy.proxy)) driver = webdriver.Chrome(options=option) |
这里就不细致讲解了，主要代码如下
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 | class Tencent(): def __init__(self): server = Server('browsermob-proxy-2.1.4/bin/browsermob-proxy') server.start() self.proxy = server.create_proxy(params={'trustAllServers':'true'}) self.url = 'aHR0cHM6Ly93d3cudXJidGl4LmhrL2xvZ2lu' option = ChromeOptions() option.add_experimental_option('excludeSwitches', ['enable-automation']) option.add_experimental_option('useAutomationExtension', False) option.add_argument('--proxy-server={0}'.format(self.proxy.proxy)) self.proxy.new_har(options={'captureContent': True,'captureHeaders': True}) self.driver = webdriver.Chrome(options=option) self.driver.execute_cdp_cmd('Page.addScriptToEvaluateOnNewDocument', { 'source': 'Object.defineProperty(navigator,"webdriver",{get: () => undefined})' }) with open('stealth.min.js') as f: js = f.read() self.driver.execute_cdp_cmd("Page.addScriptToEvaluateOnNewDocument", {"source": js}) self.driver.maximize_window() self.det = ddddocr.DdddOcr(det=False, ocr=False, show_ad=False) self.headers = {'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/108.0.0.0 Safari/537.36'} def index(self): """ 主流程 """ self.driver.get(self.url) time.sleep(5) print("正在打开网页~~~") self.driver.find_element(by=By.XPATH, value=f'//*[@id="root"]/div/div[3]/div/div/div[5]/div/div').click() time.sleep(1) self.driver.find_element(by=By.XPATH, value=f'//*[@id="root"]/div/div[3]/div/div/div[8]/div[2]/div').click() time.sleep(5) self.driver.switch_to.frame('tcaptcha_iframe_dy') bg_style = self.driver.find_element('id','slideBg').get_attribute("style") cut_style = self.driver.find_element(by=By.XPATH, value=f'//*[@id="tcOperation"]/div[8]').get_attribute("style") bg_url = re.findall('url\("(.*?)"\)',str(bg_style))[0] cut_url = re.findall('url\("(.*?)"\)', str(cut_style))[0] print("获取到背景图片url:",bg_url) print("获取到滑块图片url:",cut_url) with open("bg_picture.jpg", "wb") as f: f.write(requests.get(bg_url).content) with open("cut_picture.png", "wb") as f: f.write(requests.get(cut_url).content) def get_gap_offset(self): """ 识别坐标 """ img = Image.open('cut_picture.png') region = img.crop((160, 508, 243, 595)) # region.save(f'cut_picture.png') with open('bg_picture.jpg', 'rb') as f: target_bytes = f.read() with open('cut_picture.png', 'rb') as f: background_bytes = f.read() res = self.det.slide_match(target_bytes, background_bytes, simple_target=True) print("识别到的坐标位置：",res) distance = int(res['target'][0]) return distance def get_track(self, offset): ''' 计算滑块的移动轨迹 ''' offset -= 30 # 滑块并不是从0开始移动，有一个初始值 a = offset / 4 track = [a, a, a, a] return track def shake_mouse(self): """ 模拟人手释放鼠标抖动 """ ActionChains(self.driver).move_by_offset(xoffset=-2, yoffset=0).perform() ActionChains(self.driver).move_by_offset(xoffset=2, yoffset=0).perform() def operate_slider(self, track): """ 拖动滑块 :param track: 运动轨迹 """ # 定位到拖动按钮 slider_bt = self.driver.find_element(by=By.XPATH,value ='//*[@id="tcOperation"]/div[6]') # 点击拖动按钮不放 ActionChains(self.driver).click_and_hold(slider_bt).perform() # 按正向轨迹移动 for i in track: ActionChains(self.driver).move_by_offset(xoffset=i, yoffset=0).perform() time.sleep(random.random() / 100) # 每移动一次随机停顿0-1/100秒之间骗过了极验，通过率很高 time.sleep(random.random()) # 按逆向轨迹移动 back_tracks = [-1, -0.5, -1] for i in back_tracks: time.sleep(random.random() / 100) ActionChains(self.driver).move_by_offset(xoffset=i, yoffset=0).perform() # 模拟人手抖动 self.shake_mouse() time.sleep(random.random()) # 松开滑块按钮 ActionChains(self.driver).release().perform() time.sleep(2) def login(self): ''' 实现主要的登陆逻辑 ''' self.index() distance = self.get_gap_offset() distance = int(distance/1.97) track = self.get_track(distance) self.operate_slider(track) result = self.proxy.har for entry in result['log']['entries']: if entry['request']['url'] == 'ef9K9s2c8@1M7s2y4Q4x3@1q4Q4x3V1k6Q4x3V1k6@1i4K6u0W2j5$3q4H3N6r3y4Z5j5g2)9J5k6i4q4I4i4K6u0W2j5$3!0E0i4K6u0r3j5$3q4H3i4K6g2X3N6h3&6A6L8$3&6Q4y4h3k6F1k6i4N6Q4y4h3k6$3k6i4u0A6k6Y4W2Q4x3U0M7`.: print(entry['request']['url'],entry['response']['content']) print(entry['response']['content']['text']) |
案例三：ali滑块(playwright)
1 | 目标网站：aHR0cHM6Ly9wYXNzcG9ydC5kYW1haS5jbi9sb2dpbg== |
1、前面的初始化流程
1 2 3 4 5 6 7 8 9 10 | async with async_playwright() as p: browser = await p.chromium.launch(headless=False, args=['--start-maximized']) context = await browser.new_context(viewport={"width": 1920, "height": 1080}, no_viewport=True) context.set_default_timeout(8000) page = await context.new_page() await page.add_init_script(js) print("打开网页~~~") await page.goto('aHR0cHM6Ly9wYXNzcG9ydC5kYW1haS5jbi9sb2dpbg==') await page.wait_for_timeout(1000) page.on('response', on_response) |
2、输入账号密码
1 2 3 4 5 6 7 8 9 | #这里需要注意这个iframe,前面的iframe和后面出滑块之后的iframe属于包含关系new_frame = page.frame_locator('iframe[id="alibaba-login-box"]')await page.wait_for_timeout(1000)await new_frame.locator('#fm-login-id').fill("正确的手机号码")await page.wait_for_timeout(1000)await new_frame.locator('#fm-login-password').fill("错误的密码") await page.wait_for_timeout(1000)await new_frame.get_by_role("button", name="登录").click()await page.wait_for_timeout(1000) |
3、强制弹出滑块，并判断
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | #这里为了让它出滑块要先输出错误的密码，然后一直click，直到出滑块为止while True: try: new_frame2 = new_frame.frame_locator('iframe[id="baxia-dialog-content"]') move_tag = new_frame2.locator('xpath=//*[@id="nc_1_n1z"]') number = await move_tag.count() if number>=1: box = await move_tag.bounding_box() print("目前点击的位置", box) break else: print(f"没出滑块，重新点击") await page.wait_for_timeout(1000) await new_frame.get_by_role("button", name="登录").click() except: await new_frame.get_by_role("button", name="登录").click() |
4、定位以及滑动
1 2 3 4 5 6 7 8 9 10 11 12 | async def move_down(page,box): await page.mouse.move(box["x"] + box["width"] / 2, box["y"] + box["height"] / 2) await page.mouse.down() # 按下鼠标 await page.wait_for_timeout(1200) x = box["x"] + box["width"] / 2 # 这里获取到x坐标中心点位置 move_distance = get_track_list(265) print("轨迹：",move_distance) for i in move_distance: x += i await page.mouse.move(x, box["y"]) await page.mouse.up() await page.wait_for_timeout(500) |
至此ali滑块的分析就结束了
结言
其实爬虫圈一直有个不好的习惯，就是协议>自动化，但是在面对企业与客户的需求时，我们所需要的是完整的解决方案，所以不管协议与自动化，只要解决问题就是最好的方式，不存在谁大于谁，另外我也运营了一个星球，后面所有的公众号文章的示例代码都会放到星球里面去，同时也会在星球解答大家的问题。
[培训]Windows内核深度攻防：从Hook技术到Rootkit实战！
最后于 2024-12-1 12:46
被Harden13编辑
，原因：
|
|
|---|---|
|
|
|
|
|
用协议过工作量会大很多很多 
|
|
|
|
