【问题记录】- 谷歌浏览器 Html生成PDF

2021 年 8 月 8 日
筆記
Chrome, PDF, 问题记录

起因：

　由于项目需要实现将网页静默打印效果，那么直接使用浏览器打印功能无法达到静默打印效果。

　浏览器打印都会弹出预览界面(如下图)，无法达到静默打印。

解决方案：

　谷歌浏览器提供了将html直接打印成pdf并保存成文件方法，然后再将pdf进行静默打印。

　在调用谷歌命令前，需要获取当前谷歌安装位置：

public static class ChromeFinder
{
    #region 获取应用程序目录
    private static void GetApplicationDirectories(ICollection<string> directories)
    {
        if (RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
        {
            const string subDirectory = "Google\\Chrome\\Application";
            directories.Add(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.ProgramFiles), subDirectory));
            directories.Add(Path.Combine(Environment.GetFolderPath(Environment.SpecialFolder.ProgramFilesX86), subDirectory));
        }
        else if (RuntimeInformation.IsOSPlatform(OSPlatform.Linux))
        {
            directories.Add("/usr/local/sbin");
            directories.Add("/usr/local/bin");
            directories.Add("/usr/sbin");
            directories.Add("/usr/bin");
            directories.Add("/sbin");
            directories.Add("/bin");
            directories.Add("/opt/google/chrome");
        }
        else if (RuntimeInformation.IsOSPlatform(OSPlatform.OSX))
            throw new Exception("Finding Chrome on MacOS is currently not supported, please contact the programmer.");
    }
    #endregion
    #region 获取当前程序目录
    private static string GetAppPath()
    {
        var appPath = AppDomain.CurrentDomain.BaseDirectory;
        if (appPath.EndsWith(Path.DirectorySeparatorChar.ToString()))
            return appPath;
        return appPath + Path.DirectorySeparatorChar;
    }
    #endregion
    #region 查找
    /// <summary>
    /// 尝试查找谷歌程序
    /// </summary>
    /// <returns></returns>
    public static string Find()
    {
        // 对于Windows，我们首先检查注册表。这是最安全的方法，也考虑了非默认安装位置。请注意，Chrome x64当前（2019年2月）也安装在程序文件（x86）中，并使用相同的注册表项！
        if (RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
        {
            var key = Registry.GetValue(@"HKEY_LOCAL_MACHINE\SOFTWARE\WOW6432Node\Microsoft\Windows\CurrentVersion\Uninstall\Google Chrome","InstallLocation", string.Empty);
            if (key != null)
            {
                var path = Path.Combine(key.ToString(), "chrome.exe");
                if (File.Exists(path)) return path;
            }
        }
        // 收集常用的可执行文件名
        var exeNames = new List<string>();

        if (RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
            exeNames.Add("chrome.exe");
        else if (RuntimeInformation.IsOSPlatform(OSPlatform.Linux))
        {
            exeNames.Add("google-chrome");
            exeNames.Add("chrome");
            exeNames.Add("chromium");
            exeNames.Add("chromium-browser");
        }
        else if (RuntimeInformation.IsOSPlatform(OSPlatform.OSX))
        {
            exeNames.Add("Google Chrome.app/Contents/MacOS/Google Chrome");
            exeNames.Add("Chromium.app/Contents/MacOS/Chromium");
        }
        //检查运行目录
        var currentPath = GetAppPath();
        foreach (var exeName in exeNames)
        {
            var path = Path.Combine(currentPath, exeName);
            if (File.Exists(path)) return path;
        }
        //在通用软件安装目录中查找谷歌程序文件
        var directories = new List<string>();
        GetApplicationDirectories(directories);
        foreach (var exeName in exeNames)
        {
            foreach (var directory in directories)
            {
                var path = Path.Combine(directory, exeName);
                if (File.Exists(path)) return path;
            }
        }
        return null;
    }
    #endregion
}

　1、命令方式：

　　通过命令方式启动谷歌进程，传入网页地址、pdf保存位置等信息，将html转换成pdf:

/// <summary>
/// 运行cmd命令
/// </summary>
/// <param name="command"></param>
private void RunCMD(string command)
{
    Process p = new Process();
    p.StartInfo.FileName = "cmd.exe";
    p.StartInfo.UseShellExecute = false;    //是否使用操作系统shell启动
    p.StartInfo.RedirectStandardInput = true;//接受来自调用程序的输入信息
    p.StartInfo.RedirectStandardOutput = true;//由调用程序获取输出信息
    p.StartInfo.RedirectStandardError = true;//重定向标准错误输出
    p.StartInfo.CreateNoWindow = true;//不显示程序窗口
    p.Start();//启动程序
    //向cmd窗口发送输入信息
    p.StandardInput.WriteLine(command + "&exit");
    p.StandardInput.AutoFlush = true;
    //p.StandardInput.WriteLine("exit");
    //向标准输入写入要执行的命令。这里使用&是批处理命令的符号，表示前面一个命令不管是否执行成功都执行后面(exit)命令，如果不执行exit命令，后面调用ReadToEnd()方法会假死
    //同类的符号还有&&和||前者表示必须前一个命令执行成功才会执行后面的命令，后者表示必须前一个命令执行失败才会执行后面的命令
    //获取cmd窗口的输出信息
    p.StandardOutput.ReadToEnd();
    p.WaitForExit();//等待程序执行完退出进程
    p.Close();
}

public void GetPdf(string url, List<string> args = null)
{
    var chromeExePath = ChromeFinder.Find();
    if (string.IsNullOrEmpty(chromeExePath))
    {
        MessageBox.Show("获取谷歌浏览器地址失败");
        return;
    }
    var outpath = Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "tmppdf");
    if (!Directory.Exists(outpath))
    {
        Directory.CreateDirectory(outpath);
    }
    outpath = Path.Combine(outpath, DateTime.Now.Ticks + ".pdf");
    if (args == null)
    {
        args = new List<string>();
        args.Add("--start-in-incognito");//隐身模式
        args.Add("--headless");//无界面模式
        args.Add("--disable-gpu");//禁用gpu加速
        args.Add("--print-to-pdf-no-header");//打印生成pdf无页眉页脚
        args.Add($"--print-to-pdf=\"{outpath}\" \"{url}\"");//打印生成pdf到指定目录
    }
    string command = $"\"{chromeExePath}\"";
    if (args != null && args.Count > 0)
    {
        foreach (var item in args)
        {
            command += $" {item} ";
        }
    }
    Stopwatch sw = new Stopwatch();
    sw.Start();
    RunCMD(command);
    sw.Stop();
    MessageBox.Show(sw.ElapsedMilliseconds + "ms");
}

　　其中最主要的命令参数包含：

　　a) –headless：无界面

　　b) –print-to-pdf-no-header ：打印生成pdf不包含页眉页脚

　　c) –print-to-pdf：将页面打印成pdf，参数值为输出地址

　　存在问题：

- 通过该方式会生成多个谷歌进程(多达5个)，并且频繁的创建进程在性能较差时，会导致生成pdf较慢
- 在某些情况下，谷歌创建的进程：未能完全退出，导致后续生成pdf未执行。

　　　　　　异常进程参数类似：–type=crashpad-handler “–user-data-dir=xxx” /prefetch:7 –monitor-self-annotation=ptype=crashpad-handler “–database=xx” “–metrics-dir=xx” –url=//clients2.google.com/cr/report –annotation=channel= –annotation=plat=Win64 –annotation=prod=Chrome

　　那么，有没有方式能达到重用谷歌进程，并且能生成pdf操作呢? 那就需要使用第二种方式。

　2、Chrome DevTools Protocol 方式

　　该方式主要步骤：

创建一个无界面谷歌进程

#region 启动谷歌浏览器进程
/// <summary>
/// 启动谷歌进程，如已启动则不启动
/// </summary>
/// <exception cref="ChromeException"></exception>
private void StartChromeHeadless()
{
    if (IsChromeRunning)
    {
        return;
    }

    var workingDirectory = Path.GetDirectoryName(_chromeExeFileName);
    _chromeProcess = new Process();
    var processStartInfo = new ProcessStartInfo
    {
        FileName = _chromeExeFileName,
        Arguments = string.Join(" ", DefaultChromeArguments),
        CreateNoWindow = true,
    };
    _chromeProcess.ErrorDataReceived += _chromeProcess_ErrorDataReceived;
    _chromeProcess.EnableRaisingEvents = true;
    processStartInfo.UseShellExecute = false;
    processStartInfo.RedirectStandardError = true;
    _chromeProcess.StartInfo = processStartInfo;
    _chromeProcess.Exited += _chromeProcess_Exited;
    try
    {
        _chromeProcess.Start();
    }
    catch (Exception exception)
    {
        throw;
    }
    _chromeWaitEvent = new ManualResetEvent(false);
    _chromeProcess.BeginErrorReadLine();
    if (_conversionTimeout.HasValue)
    {
        if (!_chromeWaitEvent.WaitOne(_conversionTimeout.Value))
            throw new Exception($"超过{_conversionTimeout.Value}ms，无法连接到Chrome开发工具");
    }
    _chromeWaitEvent.WaitOne();
    _chromeProcess.ErrorDataReceived -= _chromeProcess_ErrorDataReceived;
    _chromeProcess.Exited -= _chromeProcess_Exited;
}
/// <summary>
/// 退出事件
/// </summary>
/// <param name="sender"></param>
/// <param name="e"></param>
private void _chromeProcess_Exited(object sender, EventArgs e)
{
    try
    {
        if (_chromeProcess == null) return;
        var exception = Marshal.GetExceptionForHR(_chromeProcess.ExitCode);
        throw new Exception($"Chrome意外退出, {exception}");
    }
    catch (Exception exception)
    {
        _chromeEventException = exception;
        _chromeWaitEvent.Set();
    }
}/// <summary>
/// 当Chrome将数据发送到错误输出时引发
/// </summary>
/// <param name="sender"></param>
/// <param name="args"></param>
private void _chromeProcess_ErrorDataReceived(object sender, DataReceivedEventArgs args)
{
    try
    {
        if (args.Data == null || string.IsNullOrEmpty(args.Data) || args.Data.StartsWith("[")) return;
        if (!args.Data.StartsWith("DevTools listening on")) return;
        // DevTools listening on ws://127.0.0.1:50160/devtools/browser/53add595-f351-4622-ab0a-5a4a100b3eae
        var uri = new Uri(args.Data.Replace("DevTools listening on ", string.Empty));
        ConnectToDevProtocol(uri);
        _chromeProcess.ErrorDataReceived -= _chromeProcess_ErrorDataReceived;
        _chromeWaitEvent.Set();
    }
    catch (Exception exception)
    {
        _chromeEventException = exception;
        _chromeWaitEvent.Set();
    }
}
#endregion

从进程输出信息中获取浏览器ws连接地址，并创建ws连接；向谷歌浏览器进程发送ws消息：打开一个选项卡

WebSocket4Net.WebSocket _browserSocket = null;
/// <summary>
/// 创建连接
/// </summary>
/// <param name="uri"></param>
private void ConnectToDevProtocol(Uri uri)
{
    //创建socket连接
    //浏览器连接：ws://127.0.0.1:50160/devtools/browser/53add595-f351-4622-ab0a-5a4a100b3eae
    _browserSocket = new WebSocket4Net.WebSocket(uri.ToString());
    _browserSocket.MessageReceived += WebSocket_MessageReceived;
    JObject jObject = new JObject();
　　 jObject["id"] = 1;
　　 jObject["method"] = "Target.createTarget";
    jObject["params"] = new JObject();
    jObject["params"]["url"] = "about:blank";
    _browserSocket.Send(jObject.ToString());
    //创建页卡Socket连接
    //页卡连接：ws://127.0.0.1:50160/devtools/browser/53add595-f351-4622-ab0a-5a4a100b3eae
    var pageUrl = $"{uri.Scheme}://{uri.Host}:{uri.Port}/devtools/page/页卡id";
}

根据devtools协议向当前页卡创建ws连接

WebSocket4Net.WebSocket _pageSocket = null;
private void WebSocket_MessageReceived(object sender, WebSocket4Net.MessageReceivedEventArgs e)
{
    string msg = e.Message;
    var pars = JObject.Parse(msg);
    string id = pars["id"].ToString();
    switch (id)
    {
        case "1":
            var pageUrl = $"{_browserUrl.Scheme}://{_browserUrl.Host}:{_browserUrl.Port}/devtools/page/{pars["result"]["targetId"].ToString()}";
            _pageSocket = new WebSocket4Net.WebSocket(pageUrl);
            _pageSocket.MessageReceived += _pageSocket_MessageReceived;
            _pageSocket.Open();
            break;
    }
}

向页卡发送命令，跳转到需要生成pdf的页面

//发送刷新命令
JObject jObject = new JObject();
jObject["method"] = "Page.navigate"; //方法
jObject["id"] = "2"; //id
jObject["params"] = new JObject(); //参数
jObject["params"]["url"] = "//www.baidu.com";
_pageSocket.Send(jObject.ToString());

最后项该页卡发送命令生成pdf　　

//发送刷新命令
jObject = new JObject();
jObject["method"] = "Page.printToPDF"; //方法
jObject["id"] = "3"; //id
jObject["params"] = new JObject(); //参数打印参数设置
jObject["params"]["landscape"] = false;
jObject["params"]["displayHeaderFooter"] = false;
jObject["params"]["printBackground"] = false;
_pageSocket.Send(jObject.ToString());

　　命令支持的详细内容，详细查看DevTools协议内容

参考：

　DevTools协议： Chrome DevTools Protocol – Page domain

谷歌参数说明：List of Chromium Command Line Switches « Peter Beverloo

Tags: Chrome PDF 问题记录