opencv+playwright滑动验证码的实现

from playwright.sync_api import sync_playwright
from PIL import Image
import numpy as np
import cv2 as cv
import requests
import random
import re
 
 
class QQZonSlide:
    def __init__(self):
        self.login_url = "https://i.qq.com/"
        self.username = "你的账号"
        self.password = "你的密码"
        self.page = None
 
    def start(self):
        with sync_playwright() as p:
            self.init_page(p)
            self.login()
 
            while True:
                self.get_slide_bg_img()
                start_x = self.get_slide_block_img_and_start_x()
                distance1, distance2 = self.get_slide_distance(start_x)
                slide_result = self.move_to_notch(distance1, distance2)
                if not slide_result:
                    self.refresh_captcha()
                else:
                    break
 
    def init_page(self, p):
        """初始化浏览器，获取page对象"""
        browser = p.chromium.launch(headless=False)
        self.page = browser.new_page()
 
    def login(self):
        """通过账号密码登录"""
        print("开始登录")
 
        # 访问页面
        self.page.goto(self.login_url)
 
        # 定位到登录框元素并点击密码登录
        login_frame = self.page.frame_locator("#login_frame")
        login_frame.get_by_role("link", name="密码登录").click()
 
        # 清空账号框然后输入账号
        login_frame.locator("#u").clear()
        login_frame.locator("#u").fill(self.username)
 
        # 清空密码框然后输入密码
        login_frame.locator("#p").clear()
        login_frame.locator("#p").fill(self.password)
 
        # 点击登录按钮
        self.page.wait_for_timeout(1000)
        login_frame.locator("#login_button").click()
 
    def get_slide_bg_img(self):
        """截取滑动验证码背景图片"""
        self.page.wait_for_timeout(2000)
        print("正在获取滑动验证码背景图片")
 
        # 获取滑动验证码所在的iframe
        captcha_iframe = self.page.frame_locator("#login_frame").frame_locator("#tcaptcha_iframe_dy")
 
        # 获取滑动验证码的背景图
        slide_bg_style = captcha_iframe.locator("#slideBg").get_attribute("style")
        slide_bg_url = re.search(r'url\("(.+)"\)', slide_bg_style).groups()[0]
        r = requests.get(slide_bg_url)
        with open("./slide_bg.png", "wb") as f:
            f.write(r.content)
 
        # 调整图片大小，根据style内容将宽度调整为280，高度等比例调整
        img = Image.open("./slide_bg.png")
        ratio = img.width / 280
        img = img.resize(size=(280, int(img.height/ratio)))
        img.save("./slide_bg.png")
 
    def get_slide_block_img_and_start_x(self):
        """获取滑块图片以及初始x坐标"""
        print("正在获取滑块图片")
 
        # 首先保存整个登录背景截图
        self.page.screenshot(path="bg.png")
 
        # 获取滑动验证码所在的iframe
        captcha_iframe = self.page.frame_locator("#login_frame").frame_locator("#tcaptcha_iframe_dy")
 
        # 获取滑块图片
        # .tc-fg-item对应的有三个元素，一个是目标滑块，一个是滑轨，还有一个是滑轨上的按钮
        for i in range(3):
            slide_block_ele = captcha_iframe.locator(".tc-fg-item").nth(i)
            slide_block_style = slide_block_ele.get_attribute("style")
 
            # 滑轨按钮元素的style值中不包含url字符串
            if "url" not in slide_block_style:
                continue
 
            # 从元素的style值中分析得出只有目标滑块的top值小于150
            top_value = re.search(r'top: (.+)px;', slide_block_style).groups()[0]
            if float(top_value) > 150:
                continue
 
            # 获取x坐标
            slide_block_x = float(re.search(r'left: (.+)px; top: ', slide_block_style).groups()[0])
 
            # 通过滑块位置，从背景图中截取滑块图片
            slide_block_rect = slide_block_ele.bounding_box()
            bg = Image.open("./bg.png")
            offset = slide_block_rect["width"] // 4  # 从背景图上截取会混入滑块周围的一些像素点，所以加一个偏移值，截取到滑块内部的图片。
            slide_block_img = bg.crop((slide_block_rect["x"] + offset, slide_block_rect["y"] + offset,
                                       slide_block_rect["x"] + slide_block_rect["width"] - offset,
                                       slide_block_rect["y"] + slide_block_rect["height"] - offset))
            slide_block_img.save("slide_block.png")
 
            return slide_block_x + slide_block_rect["width"] // 4
 
    @staticmethod
    def set_contrast_brightness(frame, contrast_value, brightness_value):
        if not contrast_value:
            contrast_value = 0.0
 
        if not brightness_value:
            brightness_value = 0
 
        blank = np.zeros(frame.shape, frame.dtype)
        frame = cv.addWeighted(frame, contrast_value, blank, 1 - contrast_value, brightness_value)
 
        return frame
 
    def get_slide_distance(self, start_x):
        """获取滑动距离"""
        print("正在获取滑动距离")
        # 通过opencv比较图片，获取缺口位置
        slide_bg_img = cv.imread("./slide_bg.png")
        slide_block_img = cv.imread("./slide_block.png")
        slide_block_img = self.set_contrast_brightness(slide_block_img, 0.4, 0)
        result = cv.matchTemplate(slide_block_img, slide_bg_img, cv.TM_CCOEFF_NORMED)
        minVal, maxVal, minLoc, maxLoc = cv.minMaxLoc(result)
 
        # 缺口的x坐标
        notch_x1 = minLoc[0]
        notch_x2 = maxLoc[0]
 
        # 距离
        distance1 = notch_x1 - start_x
        distance2 = notch_x2 - start_x
        return distance1, distance2
 
    @staticmethod
    def get_tracks(distance):
        """获取移动轨迹"""
        tracks = []                             # 移动轨迹
        current = 0                             # 当前位移
        mid = distance * 4 / 5                  # 减速阈值
        t = 0.2                                 # 计算间隔
        v = 0                                   # 初始速度
 
        while current < distance:
            if current < mid:
                a = random.randint(3, 5)        # 加速度为正5
            else:
                a = random.randint(-5, -3)      # 加速度为负3
 
            v0 = v                              # 初速度 v0
            v = v0 + a * t                      # 当前速度
            move = v0 * t + 1 / 2 * a * t * t   # 移动距离
            current += move
            tracks.append(round(current))
 
        return tracks
 
    def move_to_notch(self, distance1, distance2):
        """移动滑轨按钮到缺口处"""
        # 获取滑动验证码所在的iframe
        captcha_iframe = self.page.frame_locator("#login_frame").frame_locator("#tcaptcha_iframe_dy")
 
        for i in range(2):
            # 获取按钮位置，将鼠标移到上方并按下
            slider_btn_rect = captcha_iframe.get_by_alt_text("slider").bounding_box()
            self.page.mouse.move(slider_btn_rect['x'], slider_btn_rect['y'])
            self.page.mouse.down()
 
            distance = [distance1, distance2][i]
            if distance <= 0:   # 距离不可能小于等于0
                continue
 
            print(f"正在进行第{i+1}次滑动")
            tracks = self.get_tracks(distance)
            for x in tracks:
                self.page.mouse.move(slider_btn_rect['x']+x, random.randint(-5, 5)+slider_btn_rect['y'])
            self.page.mouse.move(slider_btn_rect['x'] + tracks[-1] + 5, random.randint(-5, 5) + slider_btn_rect['y'])
            self.page.mouse.move(slider_btn_rect['x'] + tracks[-1] - 5, random.randint(-5, 5) + slider_btn_rect['y'])
            self.page.mouse.up()
 
            # 滑动结束后等待一段时间
            self.page.wait_for_timeout(2000)
 
            # 寻找按钮是否还存在，不存在的话表明已通过滑动验证码，存在的话尝试下一个距离
            try:
                captcha_iframe.get_by_alt_text("slider").wait_for(timeout=2000)
            except Exception as e:
                print("已通过滑动验证码")
                return True
            else:
                print(f"第{i+1}次滑动失败")
 
        return False
 
    def refresh_captcha(self):
        """刷新验证码"""
        # 获取滑动验证码所在的iframe
        print("刷新验证码")
        captcha_iframe = self.page.frame_locator("#login_frame").frame_locator("#tcaptcha_iframe_dy")
        captcha_iframe.locator("#e_reload").click()
        self.page.wait_for_timeout(2000)
 
 
if __name__ == "__main__":
    slide = QQZonSlide()
    slide.start()