How do I create a CSharpCompilation using Roslyn in Blazor WebAssembly? - .net-core

I am trying to write a Blazor WebAssembly (WASM) app that accepts some code (from some text input field) and compiles the code using Roslyn.
I'm using Roslyn's CSharpCompilation class to create the compilation. Its Create method takes four parameters, one of which is a list of MetadataReferences (aka assembly references). In other (non-blazor) type applications, like a C# console app, you could get these MetadataReferences based on Asssembly Location, like this:
var locatedAssemblies = AppDomain.CurrentDomain.GetAssemblies().Where(a => !string.IsNullOrEmpty(a.Location)).ToArray();
foreach (var assembly in locatedAssemblies)
{
MetadataReference reference = MetadataReference.CreateFromFile(assembly.Location);
}
This unfortunately no longer works in Blazor WASM, because the Locations of the assemblies are empty.
I had tried getting assemblies in different ways, like AppDomain.CurrentDomain.GetAssemblies() and Assembly.GetEntryAssembly().GetReferencedAssemblies(), but all had empty Locations. I also tried calling Assembly.Load(), but to no avail.
Does anyone know how to get MetadataReferences in Blazor WASM, or how I would otherwise create a compilation in Blazor WASM?
(I'm also aware of MetadataReference.CreateFromStream() that I'll probably need to use, but it still requires the assembly location).
Thanks in advance.

I also wanted to compile C# inside a Blazor WASM app and found your question without an answer. After some digging I was able to create a working demo (repo link below.) Basically get the bytes for each assembly with HttpClient and use MetadataReference.CreateFromImage(bytes).
Full basic example repo I created: https://github.com/LostBeard/BlazorWASMScriptLoader
ScriptLoaderService.cs source:
using Microsoft.AspNetCore.Components;
using Microsoft.CodeAnalysis;
using Microsoft.CodeAnalysis.CSharp;
using Microsoft.CodeAnalysis.Emit;
using Microsoft.CodeAnalysis.Text;
using System.Collections.Immutable;
using System.Reflection;
namespace BlazorWASMScriptLoader
{
// requires "Microsoft.CodeAnalysis.CSharp"
// can be added via nuget
public class ScriptLoaderService
{
HttpClient _httpClient = new HttpClient();
public ScriptLoaderService(NavigationManager navigationManager)
{
_httpClient.BaseAddress = new Uri(navigationManager.BaseUri);
}
async Task<MetadataReference?> GetAssemblyMetadataReference(Assembly assembly)
{
MetadataReference? ret = null;
var assmeblyName = assembly.GetName().Name;
var assemblyUrl = $"./_framework/{assmeblyName}.dll";
try
{
var tmp = await _httpClient.GetAsync(assemblyUrl);
if (tmp.IsSuccessStatusCode)
{
var bytes = await tmp.Content.ReadAsByteArrayAsync();
ret = MetadataReference.CreateFromImage(bytes);
}
}
catch { }
return ret;
}
public async Task<Assembly?> CompileToDLLAssembly(string sourceCode, string assemblyName = "")
{
if (string.IsNullOrEmpty(assemblyName)) assemblyName = Path.GetRandomFileName();
var codeString = SourceText.From(sourceCode);
var options = CSharpParseOptions.Default.WithLanguageVersion(LanguageVersion.CSharp11);
var parsedSyntaxTree = SyntaxFactory.ParseSyntaxTree(codeString, options);
var appAssemblies = Assembly.GetEntryAssembly()?.GetReferencedAssemblies().Select(o => Assembly.Load(o)).ToList();
appAssemblies.Add(typeof(object).Assembly);
var references = new List<MetadataReference>();
foreach (var assembly in appAssemblies)
{
var metadataReference = await GetAssemblyMetadataReference(assembly);
if (metadataReference == null)
{
// assembly may be located elsewhere ... handle if needed
continue;
}
var metadataReferene = metadataReference;
references.Add(metadataReferene);
}
CSharpCompilation compilation = CSharpCompilation.Create(
assemblyName,
syntaxTrees: new[] { parsedSyntaxTree },
references: references,
options: new CSharpCompilationOptions(
OutputKind.DynamicallyLinkedLibrary,
concurrentBuild: false,
optimizationLevel: OptimizationLevel.Debug
)
);
using (var ms = new MemoryStream())
{
EmitResult result = compilation.Emit(ms);
if (!result.Success)
{
IEnumerable<Diagnostic> failures = result.Diagnostics.Where(diagnostic =>
diagnostic.IsWarningAsError ||
diagnostic.Severity == DiagnosticSeverity.Error);
foreach (Diagnostic diagnostic in failures)
{
Console.Error.WriteLine("{0}: {1}", diagnostic.Id, diagnostic.GetMessage());
}
return null;
}
else
{
ms.Seek(0, SeekOrigin.Begin);
var assembly = Assembly.Load(ms.ToArray());
return assembly;
}
}
}
}
}

Related

Missing method calling puppeteer LaunchAsync

When I try to use puppeteersharp in AX I get a error
"Method not found: 'Void Microsoft.Extensions.Logging.LoggerExtensions.LogError(Microsoft.Extensions.Logging.ILogger, System.Exception, System.String, System.Object[])'."
I have a custom DLL added to my AX project where I create a PDF but I call this method from AX. The DLL alone is working fine since I tested it outside of my AX project yet when I call it from AX the error occours. I tried different verions but there is always a problem with Microsoft.Extensions.Logging DLL.
EDIT!
As I looked at code the method I seek is in Microsoft.Extensions.Logging.Abstractions. Since in AX i found that the DLL version is 1.1.2 and has no 4th method for LogError where you can add Exception value. Is there any safe way to update DLL in AX365 onprem?
//X++ code
class ConvertXMLtoPDF
{
public static void main(Args _args)
{
try
{
helper.PuppeteerPDF();
}
catch(Exception::CLRError)
{
System.Exception ex = CLRInterop::getLastException();
info(ex.ToString());
}
}}
C# code
using PuppeteerSharp;
using PuppeteerSharp.Media;
using Microsoft.Extensions.Logging;
public async void PuppeteerPDF()
{
string path = #"C:\Temp\test\tes2t.pdf";
string save = #"C:\Temp\test\test.html";
var options = new LaunchOptions
{
Headless = true,
ExecutablePath = "C:\\Program Files\\Google\\Chrome\\Application\\Chrome.exe"
};
try
{
using (var browser = await Puppeteer.LaunchAsync(options)) // <--- here the error occours
using (var page = await browser.NewPageAsync())
{
await page.GoToAsync(save);
var result = await page.GetContentAsync();
await page.PdfAsync(path, new PdfOptions
{
Format = PaperFormat.A4,
DisplayHeaderFooter = true,
Scale = (decimal)0.5,
PrintBackground = true,
});
await page.DisposeAsync();
await browser.DisposeAsync();
}
}
catch (Exception ex)
{
Console.WriteLine(ex);
logger.LogError(ex, "test");
logger.LogInformation(ex, "test");
}
}

How to Read barcode image in Xamarin forms

I am trying to read the text from a QRcode image on my mobile app. I am using Xamarin.Forms with ZXing NuGet package.
I have been able to get the file using Xamarin.Essentials FilePicker. But I don't know how to actually read the barcode. I have looked at some stackoverflow solutions and they all seem to be Xamarin.Android based (using BinaryBitmap objects). I need a solution that can work for iOS and UWP as well. Here is what I have so far:
string file = "";
var filePickerOptions = new PickOptions
{
PickerTitle = "Select Barcode Image",
FileTypes = FilePickerFileType.Images
};
var result = await FilePicker.PickAsync(filePickerOptions);
if (result != null)
{
file = result.FullPath;
var res = Decode(file, BarcodeFormat.QR_CODE);
Console.WriteLine(res.Text);
}
public Result Decode(string file, BarcodeFormat? format = null, KeyValuePair<DecodeHintType, object>[] aditionalHints = null)
{
var r = GetReader(format, aditionalHints);
/* I need some function here that will allow me to get the BinaryBitmap from the image file path or something along those lines.*/
var image = GetBinaryBitmap(file);
var result = r.decode(image);
return result;
}
MultiFormatReader GetReader(BarcodeFormat? format, KeyValuePair<DecodeHintType, object>[] aditionalHints)
{
var reader = new MultiFormatReader();
var hints = new Dictionary<DecodeHintType, object>();
if (format.HasValue)
{
hints.Add(DecodeHintType.POSSIBLE_FORMATS, new List<BarcodeFormat>() { format.Value });
}
if (aditionalHints != null)
{
foreach (var ah in aditionalHints)
{
hints.Add(ah.Key, ah.Value);
}
}
reader.Hints = hints;
return reader;
}
https://github.com/Redth/ZXing.Net.Mobile/issues/981. This thread solved it for me. Credit to #jason for this response.

Result of the "R Script" without ColumnNames

I'm going crazy!
I'm using Azure Machine Learning and R Script. I deploy it as Web Service. I use sample code based on HttpClient.
using (var client = new HttpClient())
{
var scoreRequest = new
{
Inputs = new Dictionary<string, StringTable>() {
{
"input1",
new StringTable()
{
ColumnNames = new string[] {
"experts_estimates",
"experts_share_of_unique_information",
"avg_correlation",
"point_a",
"point_b",
"is_export_mode"
},
Values = new string[,] {
{
expertsEstimatesStr,
expertsShareOfUniqueInformationStr,
avgCorrelationStr,
pointAStr,
pointBStr,
isExportModeStr
},
}
}
},
},
GlobalParameters = new Dictionary<string, string>()
{
}
};
client.DefaultRequestHeaders.Authorization = new AuthenticationHeaderValue("Bearer", apiKey);
client.BaseAddress = new Uri(apiUrl);
// WARNING: The 'await' statement below can result in a deadlock
// if you are calling this code from the UI thread of an ASP.Net application.
// One way to address this would be to call ConfigureAwait(false)
// so that the execution does not attempt to resume on the original context.
// For instance, replace code such as:
// result = await DoSomeTask()
// with the following:
// result = await DoSomeTask().ConfigureAwait(false)
HttpResponseMessage response = await client.PostAsJsonAsync("", scoreRequest);
if (response.IsSuccessStatusCode)
{
string result = await response.Content.ReadAsStringAsync();
return result;
}
else
{
// Print the headers - they include the requert ID and the timestamp,
// which are useful for debugging the failure
var headers = response.Headers.ToString();
string responseContent = await response.Content.ReadAsStringAsync();
throw new Exception(responseContent, new Exception(headers));
}
}
and when I run code from Visual Studio I get:
but when I run code from Azure App Service I get:
Any ideas?
One solution is adding "edit metadata" module inside the model and rename the output columns. It'll be easy than using the code to name the columns.

Generating PDFs using Phantom JS on .NET applications

I have been looking into phantomJS and looks like it could be a great tool to use generating PDFs. I wonder if anyone have successfully used it for their .NET applications.
My specific question is: how would you use modules like rasterize.js on the server, receive requests and send back generated pdfs as a response.
My general question is: is there any best practice for using phantomJS with .NET Applications. What would be the best way to achieve it?
I am fairly new in .NET World and I would appreciate the more detailed answers. Thanks everyone. :)
I don't know about best practices, but, I'm using phantomJS with no problems with the following code.
public ActionResult DownloadStatement(int id)
{
string serverPath = HttpContext.Server.MapPath("~/Phantomjs/");
string filename = DateTime.Now.ToString("ddMMyyyy_hhmmss") + ".pdf";
new Thread(new ParameterizedThreadStart(x =>
{
ExecuteCommand("cd " + serverPath + #" & phantomjs rasterize.js http://localhost:8080/filetopdf/" + id.ToString() + " " + filename + #" ""A4""");
})).Start();
var filePath = Path.Combine(HttpContext.Server.MapPath("~/Phantomjs/"), filename);
var stream = new MemoryStream();
byte[] bytes = DoWhile(filePath);
return File(bytes, "application/pdf", filename);
}
private void ExecuteCommand(string Command)
{
try
{
ProcessStartInfo ProcessInfo;
Process Process;
ProcessInfo = new ProcessStartInfo("cmd.exe", "/K " + Command);
ProcessInfo.CreateNoWindow = true;
ProcessInfo.UseShellExecute = false;
Process = Process.Start(ProcessInfo);
}
catch { }
}
public ViewResult FileToPDF(int id)
{
var viewModel = file.Get(id);
return View(viewModel);
}
private byte[] DoWhile(string filePath)
{
byte[] bytes = new byte[0];
bool fail = true;
while (fail)
{
try
{
using (FileStream file = new FileStream(filePath, FileMode.Open, FileAccess.Read))
{
bytes = new byte[file.Length];
file.Read(bytes, 0, (int)file.Length);
}
fail = false;
}
catch
{
Thread.Sleep(1000);
}
}
System.IO.File.Delete(filePath);
return bytes;
}
Here is the action flow:
The user clicks on a link to DownloadStatement Action. Inside there, a new Thread is created to call the ExecuteCommand method.
The ExecuteCommand method is responsible to call phantomJS. The string passed as an argument do the following.
Go to the location where the phantomJS app is and, after that, call rasterize.js with an URL, the filename to be created and a print format. (More about rasterize here).
In my case, what I really want to print is the content delivered by the action filetoupload. It's a simple action that returns a simple view. PhantomJS will call the URL passed as parameter and do all the magic.
While phantomJS is still creating the file, (I guess) I can not return the request made by the client. And that is why I used the DoWhile method. It will hold the request until the file is created by phantomJS and loaded by the app to the request.
If you're open to using NReco.PhantomJS, which provides a .NET wrapper for PhantomJS, you can do this very succinctly.
public async Task<ActionResult> DownloadPdf() {
var phantomJS = new PhantomJS();
try {
var temp = Path.Combine(Path.GetTempPath(),
Path.ChangeExtension(Path.GetRandomFileName(), "pdf")); //must end in .pdf
try {
await phantomJS.RunAsync(HttpContext.Server.MapPath("~/Scripts/rasterize.js"),
new[] { "https://www.google.com", temp });
return File(System.IO.File.ReadAllBytes(temp), "application/pdf");
}
finally {
System.IO.File.Delete(temp);
}
}
finally {
phantomJS.Abort();
}
}
Here's some very basic code to generate a PDF using Phantom.JS but you can find more information here: https://buttercms.com/blog/generating-pdfs-with-node
var webPage = require('webpage');
var page = webPage.create();
page.viewportSize = { width: 1920, height: 1080 };
page.open("http://www.google.com", function start(status) {
page.render('google_home.pdf, {format: 'pdf', quality: '100'});
phantom.exit();
});

How to unit test code that uses HostingEnvironment.MapPath

I have some code that uses HostingEnvironment.MapPath which I would like to unit test.
How can I setup HostingEnvironment so that it returns a path and not null in my unit test (mstest) project?
Why would you have a code that depends on HostingEnvironment.MapPath in an ASP.NET MVC application where you have access to objects like HttpServerUtilityBase which allow you to achieve this and which can be easily mocked and unit tested?
Let's take an example: a controller action which uses the abstract Server class that we want to unit test:
public class HomeController : Controller
{
public ActionResult Index()
{
var file = Server.MapPath("~/App_Data/foo.txt");
return View((object)file);
}
}
Now, there are many ways to unit test this controller action. Personally I like using the MVcContrib.TestHelper.
But let's see how we can do this using a mocking framework out-of-the-box. I use Rhino Mocks for this example:
[TestMethod]
public void Index_Action_Should_Calculate_And_Pass_The_Physical_Path_Of_Foo_As_View_Model()
{
// arrange
var sut = new HomeController();
var server = MockRepository.GeneratePartialMock<HttpServerUtilityBase>();
var context = MockRepository.GeneratePartialMock<HttpContextBase>();
context.Expect(x => x.Server).Return(server);
var expected = #"c:\work\App_Data\foo.txt";
server.Expect(x => x.MapPath("~/App_Data/foo.txt")).Return(expected);
var requestContext = new RequestContext(context, new RouteData());
sut.ControllerContext = new ControllerContext(requestContext, sut);
// act
var actual = sut.Index();
// assert
var viewResult = actual as ViewResult;
Assert.AreEqual(viewResult.Model, expected);
}
Well I was writing a test today for code that I don't control and they used
private static String GetApplicationPath()
{
return HostingEnvironment.ApplicationVirtualPath.TrimEnd('/');
}
so here is a C# reflection hack to set that value
var path = "/aaaa/bb";
HostingEnvironment hostingEnvironment;
if (HostingEnvironment.IsHosted.isFalse())
new HostingEnvironment();
hostingEnvironment = (HostingEnvironment)typeof(HostingEnvironment).fieldValue("_theHostingEnvironment");
var virtualPath = "System.Web".assembly()
.type("VirtualPath").ctor();
virtualPath.field("_virtualPath", path);
//return virtualPath.prop("VirtualPathString");
//return virtualPath.prop("VirtualPathStringNoTrailingSlash");
hostingEnvironment.field("_appVirtualPath", virtualPath);
//hostingEnvironment.field("_appVirtualPath") == virtualPath;
return HostingEnvironment.ApplicationVirtualPath == path;
//using System.Web.Hosting
It will depend on what mocking or isolation framework you are using. You might want to look into either a) creating a wrapper type around the static property that can be mocked, or b) using a framework which can mock static properties - e.g. Moles or Typemock Isolator
As i faced same issue i changed my code bit.
From
strhtmlTemplate = File.ReadAllText(System.Web.Hosting.HostingEnvironment.MapPath(Lgetfilepath.CVal));
To
strhtmlTemplate = File.ReadAllText(HttpContextFactory.Current.Server.MapPath(Lgetfilepath.CVal));
For Unit test
public HttpContextBase mockHttpContextBase()
{
var moqContext = new Mock<HttpContextBase>();
var moqRequest = new Mock<HttpRequestBase>();
var moqServer = new Mock<HttpServerUtilityBase>();
var moqPath = new Mock<ConfigurationBase>();
moqContext.Setup(x => x.Request).Returns(moqRequest.Object);
moqContext.Setup(x => x.Server.MapPath(#"~\Data\xxxxxxx")).Returns(Environment.CurrentDirectory+#"\xxxxxx");
setupApplication(moqContext);
return moqContext.Object;
}
Now we while Writing TestClass you need to refer above method to mock. Hope it will helpful for your TestCases.
MockDataUT mockData = new MockDataUT();
var mockRequestContext = new HttpRequestContext();
HttpContextFactory.SetCurrentContext(mockData.mockHttpContextBase());
Just use this code..
Make a new folder name Reference in root directory and added your file inside this folder.
Use this
public static XElement GetFile()
{
HttpContext.Current = new HttpContext(new HttpRequest("", "http://www.google.com", ""), new HttpResponse(new StringWriter()));
var doc = new XmlDocument();
var file = HttpContext.Current.Server.MapPath("\\") + "abc.xml";
doc.Load(file);
var e = XElement.Load(new XmlNodeReader(doc));
return e;
}

Resources