Posts tagged with c

Comma-separated parameter values in WebAPI

The model binding mechanism in ASP.NET is pretty slick - it’s clever and highly extensible and built on TypeDescriptor system for all sorts of re-use that lets you get out of having to write boilerplate code to map between CLR objects and their web representations.

One surprising thing however is that out of the box neither WebAPI or MVC support comma-separated parameter values when bound to an array, e.g.

public class MyController : Controller {
    public string Page([FromUri]int[] ids) {
        return String.Join(" ; ", ids);
    }
}

Will only return 1 ; 2 ; 3 when supplied with /my/page?ids=1&ids=2&ids=3 and if you instead give it /my/page?ids=1,2,3 it will fail.

The reason for this was likely because there isn’t a standard for this at all and that the former - supported - scenario maps to what forms do when they post multiple value selections such as that in a select list box. The latter however is much more readable and is expected by some client frameworks and supported by some other web frameworks such as the Java Spring MVC framework.

Of course that extensible system lets us easily extend this behavior so that we can support both transparently - and interestingly enough - even mix-and-match on the same URL. So for example;

/my/page?ids=1,2&ids=3 will now return 1 ; 2 ; 3 in our example.

Although this supports both types if you are currently using commas in your number format this would break your app. e.g. ?ids=1,200&ids=3,500 would have been correctly received as 1200, 500 but now would be incorrectly received as 1, 200, 3, 500

CommaSeparatedArrayModelBinder class

The source is available in the DamienGKit project but also here.

Out of the box it supports integer types and Guid’s although you could extend it to floats and decimals – again just be careful with that formatting!

public class CommaSeparatedArrayModelBinder : IModelBinder {
    private static readonly Type[] supportedElementTypes = {
        typeof(int), typeof(long), typeof(short), typeof(byte),
        typeof(uint), typeof(ulong), typeof(ushort), typeof(Guid)
    };

    public bool BindModel(HttpActionContext actionContext, ModelBindingContext bindingContext) {
        if (!IsSupportedModelType(bindingContext.ModelType)) return false;
        var valueProviderResult = bindingContext.ValueProvider.GetValue(bindingContext.ModelName);
        var stringArray = valueProviderResult?.AttemptedValue
            ?.Split(new[] { ',' }, StringSplitOptions.RemoveEmptyEntries);
        if (stringArray == null) return false;
        var elementType = bindingContext.ModelType.GetElementType();
        if (elementType == null) return false;

        bindingContext.Model = CopyAndConvertArray(stringArray, elementType);
        return true;
    }

    private static Array CopyAndConvertArray(IReadOnlyList<string> sourceArray, Type elementType) {
        var targetArray = Array.CreateInstance(elementType, sourceArray.Count);
        if (sourceArray.Count > 0) {
            var converter = TypeDescriptor.GetConverter(elementType);
            for (var i = 0; i < sourceArray.Count; i++)
                targetArray.SetValue(converter.ConvertFromString(sourceArray[i]), i);
        }
        return targetArray;
    }

    internal static bool IsSupportedModelType(Type modelType) {
        return modelType.IsArray && modelType.GetArrayRank() == 1
                && modelType.HasElementType
                && supportedElementTypes.Contains(modelType.GetElementType());
    }

}

public class CommaSeparatedArrayModelBinderProvider : ModelBinderProvider {
    public override IModelBinder GetBinder(HttpConfiguration configuration, Type modelType) {
        return CommaSeparatedArrayModelBinder.IsSupportedModelType(modelType)
            ? new CommaSeparatedArrayModelBinder() : null;
    }
}

To register

It’s necessary to register ModelBinderProviders with your ASP.NET application at start-up - usually in the WebApiConfig.cs file.

public static class WebApiConfig {
    public static void Register(HttpConfiguration config) {
        // All your usual configuration up here
        config.Services.Insert(typeof(ModelBinderProvider), 0, new CommaSeparatedArrayModelBinderProvider());
    }
}

[)amien

Model binding form posts to immutable objects

I’ve been working on porting over my blog to a static site generator and fired up an Azure Function to handle the form-comment to PR process to enable user comments to still be part of the site without using a 3rd party commenting system - more on that in a future post - and found the ASP.NET model binding for form posts distinctly lacking.

It’s been great getting back into .NET and brushing up some skills making the code clear, short and reusable. What I wanted was a super-clear action on my controller that tried to collect, validate and sanitize the data then if all was well create the pull request or report errors.

Ideally it would look like this;

[FunctionName("PostComment")]
public static async Task<HttpResponseMessage> Run([HttpTrigger(AuthorizationLevel.Anonymous, "post")] HttpRequestMessage request) {
    var form = await request.Content.ReadAsFormDataAsync();
    if (TryCreateComment(form, out Comment comment, out var errors))
        await CreateCommentAsPullRequest(comment);
    return request.CreateResponse(errors.Any()
      ? HttpStatusCode.BadRequest : HttpStatusCode.OK, String.Join("\n", errors));
}

To do that however we need a function capable of creating the Comment class from the form post. Sure you can manually do it field by field but that’s not very reusable, repetitive and of course no fun. The Comment class is also - like all good little objects - immutable.

Creating a function to do this is simple with a little bit of reflection;

private static object ConvertParameter(string parameter, Type targetType) {
    return String.IsNullOrWhiteSpace(parameter)
           ? null : TypeDescriptor.GetConverter(targetType).ConvertFrom(parameter);
}

private static bool TryCreateCommentFromForm(NameValueCollection form, out Comment comment, out List<string> errors) {
    var constructor = typeof(Comment).GetConstructors()[0];
    var values = constructor.GetParameters()
                            .ToDictionary(p => p.Name, p => ConvertParameter(form[p.Name], p.ParameterType)
                                      ?? (p.HasDefaultValue ? p.DefaultValue : new MissingRequiredValue()));
    errors = values.Where(p => p.Value is MissingRequiredValue)
                   .Select(p => $"Form value missing for '{p.Key}'").ToList();
    comment = errors.Any() ? null : (Comment)constructor.Invoke(values.Values.ToArray());
    return !errors.Any();
}

What this does is grab the constructor for the Comment and try to find keys in the form that match the parameter name. Any that are missing are reported as errors unless they have a default value in which case that is used. MissingRequiredValue is just an empty object to act as a sentinel. The use of TypeDescriptor.GetConverter means it should be quite happy handling ints, decimals, urls etc.

The constructor for Comment specifies which fields are required and the parameter names must match the form field names by convention. Any value that is optional has a default value that the constructor will happily fill in a sensible default for.

public Comment(string post_id, string message, string author, string email,
    DateTime? date = null, Uri url = null, int? id = null, string gravatar = null) {
    this.post_id = pathValidChars.Replace(post_id, "-");
    this.message = message;
    this.author = author;
    this.email = email;
    this.date = date ?? DateTime.UtcNow;
    this.url = url;
    this.id = id ?? new { this.post_id, this.author, this.message, this.date }.GetHashCode();
    this.gravatar = gravatar ?? EncodeGravatar(email);
}

I’ll post more of the form commenting system source soon once it’s a bit better tested and I look into anti-spam integration. Ideally I’ll also provide an AWS Lambda variant of the code so you can choose (or load balance) comment posting and almost certainly get what you need on the free tier. For now the Jekyll rendering templates and WordPress exporter are available.

[)amien

Differences between Azure Functions v1 and v2 in C#

I’ve been messing around in the .NET ecosystem again and have jumped back in with Azure Functions (similar to AWS Lambda) to get my blog onto 99% static hosting. I immediately ran into the API changes between v1 and v2 (currently in beta).

These changes are because v1 was based around .NET 4.6 using WebAPI 2 while the v2 is based on ASP.NET Core which uses MVC 6. There are some guides around to converting but none in the pure context of Azure Functions.

I’ll illustrate with a PageViewCount sample that uses Table Storage to retrieve and update a simple page count.

v1 (.NET 4.61 / WebAPI 2)

[FunctionName("PageView")]
public static async Task<HttpResponseMessage> Run(
    [HttpTrigger(AuthorizationLevel.Anonymous, "get")]HttpRequestMessage req, TraceWriter log) {
    var page = req.MessageUri.ParseQueryString()["page"];
    if (String.IsNullOrEmpty(page))
        return req.CreateErrorResponse(HttpStatusCode.BadRequest, "'page' parameter missing.");

    var table = Helpers.GetTableReference("PageViewCounts");
    var pageView = await table.RetrieveAsync<PageViewCount>("damieng.com", page)
        ?? new PageViewCount(page) { ViewCount = 0 };
    var operation = pageView.ViewCount == 0
        ? TableOperation.Insert(pageView)
        : TableOperation.Replace(pageView);
    pageView.ViewCount++;
    await table.ExecuteAsync(operation);

    return req.CreateResponse(HttpStatusCode.OK, new { viewCount = pageView.ViewCount });
}

v2 (ASP.NET Core / MVC 6)

[FunctionName("PageView")]
public static async Task<IActionResult> Run(
    [HttpTrigger(AuthorizationLevel.Anonymous, "get")]HttpRequest req, TraceWriter log) {
    var page = req.Query["page"];
    if (String.IsNullOrEmpty(page))
       return new BadRequestObjectResult("'page' parameter missing.");

    var table = Helpers.GetTableReference("PageViewCounts");
    var pageView = await table.RetrieveAsync<PageViewCount>("damieng.com", page)
        ?? new PageViewCount(page) { ViewCount = 0 };
    var operation = pageView.ViewCount == 0
        ? TableOperation.Insert(pageView)
        : TableOperation.Replace(pageView);
    pageView.ViewCount++;
    await table.ExecuteAsync(operation);

    return new OkObjectResult(new { viewCount = pageView.ViewCount });
}

Differences

The main differences are that:

  1. Return types are IActionResult/ObjectResult objects rather than extension methods against HttpRequestMessage (easier to mock/create custom ones)
  2. Input is the HttpRequest object rather than HttpResponseMessage (easier to get query parameters)

If you get the error ‘Can not create abstract class’ when executing your function then you are trying to use the wrong tech for that environment.

Helpers

Both classes above use a small helper class to take care of Table Storage which doesn’t have the nicest to use API. A wrapper much like a data context that ensures the right types go to the right table names might be an even better options.

static class Helpers {
    public static CloudStorageAccount GetCloudStorageAccount() {
        var connection = ConfigurationManager.AppSettings["DamienGTableStorage"];
        return connection == null ? CloudStorageAccount.DevelopmentStorageAccount : CloudStorageAccount.Parse(connection);
    }

    public static CloudTable GetTableReference(string name) {
        return GetCloudStorageAccount().CreateCloudTableClient().GetTableReference(name);
    }

    public static async Task<T> RetrieveAsync<T>(this CloudTable cloudTable, string partitionKey, string rowKey)
        where T:TableEntity {
        var tableResult = await cloudTable.ExecuteAsync(TableOperation.Retrieve<T>(partitionKey, rowKey));
        return (T)tableResult.Result;
    }
}

To compile

If you want to compile this or maybe you were just looking for code to do a simple page counter here’s the missing TableEntity class;

public class PageViewCount : TableEntity
{
    public PageViewCount(string pageName)
    {
        PartitionKey = "damieng.com";
        RowKey = pageName;
    }

    public PageViewCount() { }
    public int ViewCount { get; set; }
}

[)amien

Sequence averages in Scala

I’ve been learning Scala and decided to put together a C# to Scala cheat sheet. All is going pretty well but then I got stuck on the equivalent of Average.

Enumerable.Average in .NET calculates a mean average from your sequence by summing up all the values and counting them in a single pass then returning the sum divided by the count in a floating point format (or decimal).

The problem

Given that Scala has nothing built-in there are more than a few suggestions online that boil down to:

val average = seq.sum / seq.length

This has a few problems:

  1. Visiting a sequence twice can be inefficient
  2. Sum can overflow as it is the same type as the sequence
  3. Applied to an integer without casting it returns an integer average

A solution

Scala provides a useful high-order function called foldLeft. Its job is to take an initial state and a function then keep applying the function with each value to the state. So one more efficient solution to the problem is:

val average = seq.foldLeft((0.0, 1)) ((acc, i) => ((acc._1 + (i - acc._1) / acc._2), acc._2 + 1))._1

How does this work?

What we do here is calculate an average as we go, adding the new weighted average each time.

It achieves this by setting up a tuple to contain our initial state with (0.0, 1). This specifies our starting average of 0.0 and our starting position of 1.

The next part specifies the function that takes that state as acc (for accumulator) and the next value in the sequence as i and calculates our rolling average for each value and increases the position as it goes along.

Finally at the end of our call we specify ._1 which tells the compiler we want the first value from the tuple – the average – as we no longer care about the position.

If you wanted to make this function more reusable you could do this:

def average(s: Seq[Int]): Double = s.foldLeft((0.0, 1)) ((acc, i) => ((acc._1 + (i - acc._1) / acc._2), acc._2 + 1))._1

Be aware you might need multiple overloads for each numeric sequence type you want to be able to average given the lack of a common numeric trait that allows for the subtraction and division.

Precision and rounding

There is some slight variance in results between this approach and the total / count due to rounding precision. If you wanted to preserve that you could always add and then divide at the end still in a single pass much like .NET does but with Scala’s foldLeft rather than a foreach.

def average(s: Seq[Int]): Double = { val t = s.foldLeft((0.0, 0)) ((acc, i) => (acc._1 + i, acc._2 + 1)); t._1 / t._2 }

[)amien