Python利用Selenium实现全网页截图

import wx
import os
import time
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.chrome.service import Service
from webdriver_manager.chrome import ChromeDriverManager
import base64
from urllib.parse import urlparse
import threading
 
class ScreenshotApp(wx.Frame):
    def __init__(self, parent, title):
        super(ScreenshotApp, self).__init__(parent, title=title, size=(600, 300))
        self.InitUI()
        self.Centre()
        self.Show()
 
    def InitUI(self):
        panel = wx.Panel(self)
        vbox = wx.BoxSizer(wx.VERTICAL)
 
        # URL输入区域
        url_box = wx.BoxSizer(wx.HORIZONTAL)
        url_label = wx.StaticText(panel, label="网页URL:")
        self.url_text = wx.TextCtrl(panel, size=(400, -1))
        url_box.Add(url_label, flag=wx.RIGHT | wx.ALIGN_CENTER_VERTICAL, border=8)
        url_box.Add(self.url_text, proportion=1)
 
        # 保存路径区域
        path_box = wx.BoxSizer(wx.HORIZONTAL)
        path_label = wx.StaticText(panel, label="保存路径:")
        self.path_text = wx.TextCtrl(panel, size=(300, -1))
        browse_button = wx.Button(panel, label="浏览...")
        path_box.Add(path_label, flag=wx.RIGHT | wx.ALIGN_CENTER_VERTICAL, border=8)
        path_box.Add(self.path_text, proportion=1, flag=wx.RIGHT, border=5)
        path_box.Add(browse_button)
 
        # 截图按钮
        screenshot_button = wx.Button(panel, label="截取网页")
 
        # 进度条
        self.progress_bar = wx.Gauge(panel, range=100, size=(400, 20))
        self.progress_bar.SetValue(0)
 
        # 状态显示区域
        self.status_text = wx.StaticText(panel, label="")
 
        # 添加到垂直布局
        vbox.Add((-1, 20))
        vbox.Add(url_box, flag=wx.EXPAND | wx.LEFT | wx.RIGHT, border=10)
        vbox.Add((-1, 20))
        vbox.Add(path_box, flag=wx.EXPAND | wx.LEFT | wx.RIGHT, border=10)
        vbox.Add((-1, 30))
        vbox.Add(screenshot_button, flag=wx.ALIGN_CENTER)
        vbox.Add((-1, 20))
        vbox.Add(self.progress_bar, flag=wx.EXPAND | wx.LEFT | wx.RIGHT, border=10)
        vbox.Add((-1, 10))
        vbox.Add(self.status_text, flag=wx.ALIGN_CENTER)
 
        # 绑定事件
        browse_button.Bind(wx.EVT_BUTTON, self.OnBrowse)
        screenshot_button.Bind(wx.EVT_BUTTON, self.OnScreenshot)
 
        # 设置默认保存路径为用户的图片文件夹
        default_path = os.path.join(os.path.expanduser("~"), "Pictures")
        self.path_text.SetValue(default_path)
 
        panel.SetSizer(vbox)
 
    def OnBrowse(self, event):
        dialog = wx.DirDialog(self, "选择保存截图的文件夹", style=wx.DD_DEFAULT_STYLE)
        if dialog.ShowModal() == wx.ID_OK:
            self.path_text.SetValue(dialog.GetPath())
        dialog.Destroy()
 
    def OnScreenshot(self, event):
        url = self.url_text.GetValue().strip()
        save_path = self.path_text.GetValue().strip()
 
        # 验证URL
        if not url:
            wx.MessageBox("请输入有效的URL", "错误", wx.OK | wx.ICON_ERROR)
            return
 
        # 如果URL没有http前缀，添加https://
        if not url.startswith(('http://', 'https://')):
            url = 'https://' + url
            self.url_text.SetValue(url)
 
        # 验证保存路径
        if not os.path.exists(save_path):
            try:
                os.makedirs(save_path)
            except Exception as e:
                wx.MessageBox(f"创建保存路径失败: {str(e)}", "错误", wx.OK | wx.ICON_ERROR)
                return
 
        self.status_text.SetLabel("正在加载网页并滚动，请稍候...")
        self.progress_bar.SetValue(0)
        self.Layout()
 
        # 使用线程避免界面冻结
        thread = threading.Thread(target=self.take_screenshot, args=(url, save_path))
        thread.daemon = True
        thread.start()
 
    def scroll_page(self, driver):
        """滚动页面以触发所有懒加载内容，显示滚动效果"""
        total_height = driver.execute_script("return document.body.scrollHeight")
        viewport_height = driver.execute_script("return window.innerHeight")
        steps = max(1, total_height // viewport_height)  # 计算滚动步数
        step_height = total_height / steps
 
        for i in range(steps + 1):
            scroll_position = int(i * step_height)
            driver.execute_script(f"window.scrollTo({{top: {scroll_position}, behavior: 'smooth'}});")
            wx.CallAfter(self.progress_bar.SetValue, int((i + 1) / (steps + 1) * 100))
            time.sleep(0.5)  # 等待平滑滚动和内容加载
        time.sleep(1)  # 确保所有内容加载完成
 
    def take_screenshot(self, url, save_path):
        try:
            # 设置Chrome选项，移除无头模式以显示浏览器
            chrome_options = Options()
            chrome_options.add_argument("--disable-gpu")
            chrome_options.add_argument("--disable-infobars")
            chrome_options.add_argument("--disable-extensions")
            chrome_options.add_argument("--window-size=1920,1080")  # 设置初始窗口大小
 
            # 启动浏览器
            service = Service(ChromeDriverManager().install())
            driver = webdriver.Chrome(service=service, options=chrome_options)
 
            # 访问网页
            driver.get(url)
            time.sleep(3)  # 等待页面初始加载
 
            # 滚动页面以触发懒加载内容
            self.scroll_page(driver)
 
            # 获取域名和时间戳
            domain = urlparse(url).netloc or "webpage"
            timestamp = time.strftime("%Y%m%d_%H%M%S")
 
            # 使用CDP方法获取完整页面截图
            result = driver.execute_cdp_cmd('Page.captureScreenshot', 
                                            {'format': 'png', 'captureBeyondViewport': True})
            image_data = base64.b64decode(result['data'])
            save_file = os.path.join(save_path, f"{domain}_{timestamp}.png")
            with open(save_file, 'wb') as f:
                f.write(image_data)
 
            driver.quit()
            wx.CallAfter(self.screenshot_complete, save_file)
 
        except Exception as e:
            if 'driver' in locals():
                driver.quit()
            wx.CallAfter(self.screenshot_error, str(e))
 
    def screenshot_complete(self, filepath):
        self.status_text.SetLabel(f"截图已保存至: {filepath}")
        self.progress_bar.SetValue(100)
        wx.MessageBox(f"截图已保存至:\n{filepath}", "成功", wx.OK | wx.ICON_INFORMATION)
 
    def screenshot_error(self, error_msg):
        self.status_text.SetLabel(f"截图失败: {error_msg}")
        self.progress_bar.SetValue(0)
        wx.MessageBox(f"截图失败: {error_msg}", "错误", wx.OK | wx.ICON_ERROR)
 
if __name__ == '__main__':
    app = wx.App()
    ScreenshotApp(None, title='网页截图工具')
    app.MainLoop()