We Keep Coding

How to create a TensorProto in c#?

This is a snipped of the c# client I created to query the tensorflow server I set up using this tutorial: https://tensorflow.github.io/serving/serving_inception.html
var channel = new Channel("TFServer:9000", ChannelCredentials.Insecure);
var request = new PredictRequest();
request.ModelSpec = new ModelSpec();
request.ModelSpec.Name = "inception";
var imgBuffer = File.ReadAllBytes(#"sample.jpg");
ByteString jpeg = ByteString.CopyFrom(imgBuffer, 0, imgBuffer.Length);
var jpgeproto = new TensorProto();
jpgeproto.StringVal.Add(jpeg);
jpgeproto.Dtype = DataType.DtStringRef;
request.Inputs.Add("images", jpgeproto); // new TensorProto{TensorContent = jpeg});
PredictionClient client = new PredictionClient(channel);
I found out that most classes needed to be generated from proto files using protoc
The only thing which I cant find is how to construct the TensorProto. The error I keep getting is : Additional information: Status(StatusCode=InvalidArgument, Detail="tensor parsing error: images")
There is a sample client (https://github.com/tensorflow/serving/blob/master/tensorflow_serving/example/inception_client.py) byt my Python skills are not sufficient to understand the last bit.

I also implemented that client in another language (Java).
Try to change
jpgeproto.Dtype = DataType.DtStringRef;
to
jpgeproto.Dtype = DataType.DtString;
You may also need to add a tensor shape with a dimension to your tensor proto. Here's my working solution in Java, should be similar in C#:
TensorShapeProto.Dim dim = TensorShapeProto.Dim.newBuilder().setSize(1).build();
TensorShapeProto shape = TensorShapeProto.newBuilder().addDim(dim).build();
TensorProto proto = TensorProto.newBuilder()
.addStringVal(ByteString.copyFrom(imageBytes))
.setTensorShape(shape)
.setDtype(DataType.DT_STRING)
.build();
ModelSpec spec = ModelSpec.newBuilder().setName("inception").build();
PredictRequest r = PredictRequest.newBuilder()
.setModelSpec(spec)
.putInputs("images", proto).build();
PredictResponse response = blockingStub.predict(r);

Get console output in Mercurial.Net

I am trying to do commit and push using Mercurial.Net library:
var repo = new Repository(repositoryPath);
var branchCommand = new BranchCommand { Name = branch };
repo.Branch(branchCommand);
var commitCommand = new CommitCommand { Message = commitMessage, OverrideAuthor = author };
repo.Commit(commitCommand);
var pushCommand = new PushCommand { AllowCreatingNewBranch = true, Force = true, };
repo.Push(pushCommand);
On repo.Push(pushCommand) it throws an exception Mercurial.MercurialExecutionException with message 'abort: Access is denied'.
The question is: Is there any way in Mercurial.Net to get the output of mercurial console?

The message you're receiving appears to be a message the remote is printing. It looks like you haven't set up auth properly — or you've authenticated correctly, but on the remote side you haven't got correct access rights.

Unable to create Azure WebSpace

I'm trying to create a new Website using the nuget package Microsoft.WindowsAzure.Management.WebSites (Version 3.0.0)
This tutorial has been helpful (even though its Java):
http://azure.microsoft.com/fi-fi/documentation/articles/java-create-azure-website-using-java-sdk/
except it suggests to use the WebSpaceNames.WestUSWebSpace constant.
var hostingPlanParams = new WebHostingPlanCreateParameters
{
Name = this.webhostingPlanName,
NumberOfWorkers = 1,
SKU = SkuOptions.Free,
WorkerSize = WorkerSizeOptions.Small
};
var result = new WebSiteManagementClient(this.Credentials)
.WebHostingPlans
.CreateAsync(WebSpaceNames.WestUSWebSpace, hostingPlanParams, CancellationToken.None)
.Result
This will result in an exception: NotFound: Cannot find WebSpace with name westuswebspace.
I actually want to create a custom WebSpace.
Except I can't find any method for it. See MSDN
So the only way I can make this work is using an existing WebSpace, that had created through the manage.windowsazure.com site. Which defeats the whole purpose of automating this.
The only Create[...] Method on IWebSpaceOperations is CreatePublishingUserAsync which I have tried running this as well but it results in an exception This operation is not supported for subscriptions that have co-admins. Which is pretty annoying in itself, doesn't make much sense to me, but is not really the core of my question.

I resolved this by using the prerelease package: PM> Install-Package Microsoft.Azure.Management.WebSites -Pre
Which works perfectly well. Except that it's only a pre-release of cause
// Create Web Hosting Plan
var hostingPlanParams = new WebHostingPlanCreateOrUpdateParameters
{
WebHostingPlan = new WebHostingPlan()
{
Name = "WebHostingPlanName",
Location = "Australia Southeast",
Properties = new WebHostingPlanProperties
{
NumberOfWorkers = 1,
Sku = SkuOptions.Standard,
WorkerSize = WorkerSizeOptions.Small
}
},
};
var result = this.ManagementContext.WebSiteManagementClient.WebHostingPlans.CreateOrUpdateAsync(
"ResourceGroupName",
"WebHostingPlanName",
CancellationToken.None).Result;
// Create Website
var websiteParams = new WebSiteCreateOrUpdateParameters
{
WebSite = new WebSiteBase
{
Location = "Australia Southeast",
Name = "WebSiteName",
Properties = new WebSiteBaseProperties
{
ServerFarm = "WebHostingPlanName"
}
}
};
var siteResult = this.ManagementContext.WebSiteManagementClient.WebSites.CreateOrUpdateAsync(
"ResourceGroupName",
"WebSiteName",
null,
websiteParams,
CancellationToken.None).Result;
If you want to use deployment slots you have to take this under consideration:
https://github.com/Azure/azure-sdk-for-net/issues/1088

How do I get a previous version of a file with libgit2sharp

I'm trying to use libgit2sharp to get a previous version of a file. I would prefer the working directory to remain as is, at the very least restored to previous condition.
My initial approach was to try to stash, checkout path on the file I want, save that to a string variable, then stash pop. Is there a way to stash pop? I can't find it easily. Here's the code I have so far:
using (var repo = new Repository(DirectoryPath, null))
{
var currentCommit = repo.Head.Tip.Sha;
var commit = repo.Commits.Where(c => c.Sha == commitHash).FirstOrDefault();
if (commit == null)
return null;
var sn = "Stash Name";
var now = new DateTimeOffset(DateTime.Now);
var diffCount = repo.Diff.Compare().Count();
if(diffCount > 0)
repo.Stashes.Add(new Signature(sn, "x#y.com", now), options: StashModifiers.Default);
repo.CheckoutPaths(commit.Sha, new List<string>{ path }, CheckoutModifiers.None, null, null);
var fileText = File.ReadAllText(path);
repo.CheckoutPaths(currentCommit, new List<string>{path}, CheckoutModifiers.None, null, null);
if(diffCount > 0)
; // stash Pop?
}
If there's an easier approach than using Stash, that would work great also.

Is there a way to stash pop? I can't find it easily
Unfortunately, Stash pop requires merging which isn't available yet in libgit2.
I'm trying to use libgit2sharp to get a previous version of a file. I would prefer the working directory to remain as is
You may achieve such result by opening two instances of the same repository, each of them pointing to different working directories. The Repository constructor accepts a RepositoryOptions parameter which should allow you to do just that.
The following piece of code demonstrates this feature. This creates an additional instance (otherRepo) that you can use to retrieve a different version of the file currently checked out in your main working directory.
string repoPath = "path/to/your/repo";
// Create a temp folder for a second working directory
string tempWorkDir = Path.Combine(Path.GetTempPath(), "tmp_wd");
Directory.CreateDirectory(newWorkdir);
// Also create a new index to not alter the main repository
string tempIndex = Path.Combine(Path.GetTempPath(), "tmp_idx");
var opts = new RepositoryOptions
{
WorkingDirectoryPath = tempWorkDir,
IndexPath = tempIndex
};
using (var mainRepo = new Repository(repoPath))
using (var otherRepo = new Repository(mainRepo.Info.Path, opts))
{
string path = "file.txt";
// Do your stuff with mainrepo
mainRepo.CheckoutPaths("HEAD", new[] { path });
var currentVersion = File.ReadAllText(Path.Combine(mainRepo.Info.WorkingDirectory, path));
// Use otherRepo to temporarily checkout previous versions of files
// Thank to the passed in RepositoryOptions, this checkout will not
// alter the workdir nor the index of the main repository.
otherRepo.CheckoutPaths("HEAD~2", new [] { path });
var olderVersion = File.ReadAllText(Path.Combine(otherRepo.Info.WorkingDirectory, path));
}
You can get a better grasp of this RepositoryOptions type by taking a look at the tests in RepositoryOptionFixture that exercise it.

How to access WinRM in C#

I'd like to create a small application that can collect system information (Win32_blablabla) using WinRM as opposed to WMI. How can i do that from C#?
The main goal is to use WS-Man (WinRm) as opposed to DCOM (WMI).

I guess the easiest way would be to use WSMAN automation. Reference wsmauto.dll from windwos\system32 in your project:
then, code below should work for you. API description is here: msdn: WinRM C++ API
IWSMan wsman = new WSManClass();
IWSManConnectionOptions options = (IWSManConnectionOptions)wsman.CreateConnectionOptions();
if (options != null)
{
try
{
// options.UserName = ???;
// options.Password = ???;
IWSManSession session = (IWSManSession)wsman.CreateSession("http://<your_server_name>/wsman", 0, options);
if (session != null)
{
try
{
// retrieve the Win32_Service xml representation
var reply = session.Get("http://schemas.microsoft.com/wbem/wsman/1/wmi/root/cimv2/Win32_Service?Name=winmgmt", 0);
// parse xml and dump service name and description
var doc = new XmlDocument();
doc.LoadXml(reply);
foreach (var elementName in new string[] { "p:Caption", "p:Description" })
{
var node = doc.GetElementsByTagName(elementName)[0];
if (node != null) Console.WriteLine(node.InnerText);
}
}
finally
{
Marshal.ReleaseComObject(session);
}
}
}
finally
{
Marshal.ReleaseComObject(options);
}
}
hope this helps, regards

I've got an article that describes an easy way to run Powershell through WinRM from .NET at http://getthinktank.com/2015/06/22/naos-winrm-windows-remote-management-through-net/.
The code is in a single file if you want to just copy it and it's also a NuGet package that includes the reference to System.Management.Automation.
It auto manages trusted hosts, can run script blocks, and also send files (which isn't really supported but I created a work around). The returns are always the raw objects from Powershell.
// this is the entrypoint to interact with the system (interfaced for testing).
var machineManager = new MachineManager(
"10.0.0.1",
"Administrator",
MachineManager.ConvertStringToSecureString("xxx"),
true);
// will perform a user initiated reboot.
machineManager.Reboot();
// can run random script blocks WITH parameters.
var fileObjects = machineManager.RunScript(
"{ param($path) ls $path }",
new[] { #"C:\PathToList" });
// can transfer files to the remote server (over WinRM's protocol!).
var localFilePath = #"D:\Temp\BigFileLocal.nupkg";
var fileBytes = File.ReadAllBytes(localFilePath);
var remoteFilePath = #"D:\Temp\BigFileRemote.nupkg";
machineManager.SendFile(remoteFilePath, fileBytes);
Hope this helps, I've been using this for a while with my automated deployments. Please leave comments if you find issues.

I would like to note that this shows an interop error by default in Visual Studio 2010.
c.f. http://blogs.msdn.com/b/mshneer/archive/2009/12/07/interop-type-xxx-cannot-be-embedded-use-the-applicable-interface-instead.aspx
There appear to be two ways to solve this. This first is documented in the article listed above and appears to be the correct way to handle the problem. The pertinent changes for this example is:
WSMan wsManObject = new WSMan();
This is in lieu of IWSMan wsman = new WSManClass(); which will throw the error.
The second resolution is to go to the VS2010—>Solution Explorer—>Solution—>Project—>References and select WSManAutomation. Right click or hit Alt-Enter to access the properties. Change the value of the "Embed Interop Types" property of the wsmauto reference.