[problem record] - PDF generated by Google browser HTML

cause ：

Because the project needs to realize the silent printing effect of web pages , Then the silent printing effect cannot be achieved by directly using the browser printing function .

The preview interface will pop up when the browser prints ( Here's the picture ), Unable to achieve silent printing .

Solution ：

Google browser provides a way to html Print directly to pdf And save it as a file , And then pdf Print silently .

Before invoking Google commands , Need to get the current Google installation location ：

public static class ChromeFinder
{
#region Get application directory
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 Get the current program directory
private static string GetAppPath()
{
var appPath = AppDomain.CurrentDomain.BaseDirectory;
if (appPath.EndsWith(Path.DirectorySeparatorChar.ToString()))
return appPath;
return appPath + Path.DirectorySeparatorChar;
}
#endregion
#region lookup
/// <summary>
/// Try to find Google Apps
/// </summary>
/// <returns></returns>
public static string Find()
{
// about Windows, Let's first check the registry . This is the safest way , Non default installation locations are also considered . Please note that ,Chrome x64 At present （2019 year 2 month ） Also installed in the program file （x86） in , And use the same registry key ！
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;
}
}
// Collect common executable file names 
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");
}
// Check the running Directory 
var currentPath = GetAppPath();
foreach (var exeName in exeNames)
{
var path = Path.Combine(currentPath, exeName);
if (File.Exists(path)) return path;
}
// Find Google program files in the general software installation directory 
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、 Command mode ： 

Start the Google process by command , Incoming web address 、pdf Save location and other information , take html convert to pdf:

/// <summary>
/// function cmd command
/// </summary>
/// <param name="command"></param>
private void RunCMD(string command)
{
Process p = new Process();
p.StartInfo.FileName = "cmd.exe";
p.StartInfo.UseShellExecute = false; // Whether to use the operating system shell start-up 
p.StartInfo.RedirectStandardInput = true;// Accept input from the calling program 
p.StartInfo.RedirectStandardOutput = true;// Get the output information from the calling program 
p.StartInfo.RedirectStandardError = true;// Redirect standard error output 
p.StartInfo.CreateNoWindow = true;// Don't show program window 
p.Start();// Start the program
// towards cmd Window sends input information 
p.StandardInput.WriteLine(command + "&exit");
p.StandardInput.AutoFlush = true;
//p.StandardInput.WriteLine("exit");
// Write the command to be executed to the standard input . Use here & Is the symbol of a batch command , Whether the previous command is executed successfully or not (exit) command , If not implemented exit command , Call back ReadToEnd() The method will fake death
// Similar symbols and && and || The former means that the following command will not be executed until the previous command is executed successfully , The latter means that the following command will not be executed until the previous command fails
// obtain cmd The output of the window 
 p.StandardOutput.ReadToEnd();
p.WaitForExit();// Wait for the program to finish executing and exit the process 
 p.Close();
}
public void GetPdf(string url, List<string> args = null)
{
var chromeExePath = ChromeFinder.Find();
if (string.IsNullOrEmpty(chromeExePath))
{
MessageBox.Show(" Failed to get Google browser address ");
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");// Stealth mode 
args.Add("--headless");// No interface mode 
args.Add("--disable-gpu");// Ban gpu Speed up 
args.Add("--print-to-pdf-no-header");// Print generation pdf No header, no footers 
args.Add($"--print-to-pdf=\"{outpath}\" \"{url}\"");// Print generation pdf To the specified directory 
 }
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");
}

The main command parameters include ：

a)  --headless： No interface

b) --print-to-pdf-no-header ： Print generation pdf Do not include headers and footers

c) --print-to-pdf： Print the page as pdf, The parameter value is the output address

Existing problems ：

• • In this way, multiple Google processes will be generated ( As many as 5 individual ), And frequently create processes when performance is poor , Will lead to pdf slower
  • In some cases , The process created by Google ： Failed to exit completely , Cause subsequent generation pdf unexecuted .

The exception process parameters are similar ：--type=crashpad-handler "--user-data-dir=xxx" /prefetch:7 --monitor-self-annotation=ptype=crashpad-handler "--database=xx" "--metrics-dir=xx" --url=https://clients2.google.com/cr/report --annotation=channel= --annotation=plat=Win64 --annotation=prod=Chrome

that , There is no way to reuse Google processes , And can generate pdf Operation? ? Then you need to use the second way .

2、Chrome DevTools Protocol The way

The main steps of this method are ：

• Create a Google process without interface

#region Start the Google browser process
/// <summary>
/// Start the Google process , If it has been started, it will not be started
/// </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($" exceed {_conversionTimeout.Value}ms, Can't connect to Chrome development tool ");
}
_chromeWaitEvent.WaitOne();
_chromeProcess.ErrorDataReceived -= _chromeProcess_ErrorDataReceived;
_chromeProcess.Exited -= _chromeProcess_Exited;
}
/// <summary>
/// Exit event
/// </summary>
/// <param name="sender"></param>
/// <param name="e"></param>
private void _chromeProcess_Exited(object sender, EventArgs e)
{
try
{
if (_chromeProcess == null) return;
var exception = Marshal.GetExceptionForHR(_chromeProcess.ExitCode);
throw new Exception($"Chrome Unexpected exit , {exception}");
}
catch (Exception exception)
{
_chromeEventException = exception;
_chromeWaitEvent.Set();
}
}/// <summary>
/// When Chrome Raised when data is sent to the error output
/// </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

• Get the browser from the process output information ws Connection address , And create ws Connect ; Send... To the Google browser process ws news ： Open a TAB

WebSocket4Net.WebSocket _browserSocket = null;
/// <summary>
/// Create connection
/// </summary>
/// <param name="uri"></param>
private void ConnectToDevProtocol(Uri uri)
{
// establish socket Connect
// Browser connection ：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());
// Create page cards Socket Connect
// Page card connection ：ws://127.0.0.1:50160/devtools/browser/53add595-f351-4622-ab0a-5a4a100b3eae
var pageUrl = $"{uri.Scheme}://{uri.Host}:{uri.Port}/devtools/page/ Page card id";
}

• according to devtools The agreement is created to the current page card ws Connect

  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;
  }
  }

• Send a command to the page card , Jump to the need to generate pdf The page of

// Send refresh command 
JObject jObject = new JObject();
jObject["method"] = "Page.navigate"; // Method 
jObject["id"] = "2"; //id
jObject["params"] = new JObject(); // Parameters 
jObject["params"]["url"] = "http://www.baidu.com";
_pageSocket.Send(jObject.ToString());

• Finally, the page card sends a command to generate pdf

  // Send refresh command 
  jObject = new JObject();
  jObject["method"] = "Page.printToPDF"; // Method 
  jObject["id"] = "3"; //id
  jObject["params"] = new JObject(); // Parameter print parameter settings 
  jObject["params"]["landscape"] = false;
  jObject["params"]["displayHeaderFooter"] = false;
  jObject["params"]["printBackground"] = false;
  _pageSocket.Send(jObject.ToString());

Details of command support , Detailed view DevTools Content of agreement

Reference resources ：

DevTools agreement ： Chrome DevTools Protocol - Page domain

    Google parameter description ：List of Chromium Command Line Switches « Peter Beverloo