Querystring with ampersand in Swashbuckle xml comments
For now, we're using a workaround based on EspressoBean's answer but adapted for the ASP.NET Core Swashbuckle library.
In your remarks or summary comments use XML-escaped syntax:
/// <summary>
/// Gets a TodoItem.
/// </summary>
/// <remarks>
/// Sample request:
///
/// GET /Todo?iscomplete=true&owner=mike
/// </remarks>
In Startup.cs (the ConfigureServices method) add your custom XmlCommentsEscapeFilter:
services.AddSwaggerGen(c =>
{
...
c.OperationFilter<XmlCommentsEscapeFilter>();
});
Add a class called XmlCommentsEscapeFilter.cs:
using Swashbuckle.AspNetCore.Swagger;
using Swashbuckle.AspNetCore.SwaggerGen;
namespace **MyNamespaceHere**
{
/// <summary>
/// Replace & with ampersand character in XML comments
/// </summary>
internal class XmlCommentsEscapeFilter : IOperationFilter
{
public void Apply(Operation operation, OperationFilterContext context)
{
operation.Description = operation.Description?.Replace("&", "&");
operation.Summary = operation.Summary?.Replace("&", "&");
}
}
}
For future reference, here's a link to the github issue (still open as of Aug-19-2019): https://github.com/domaindrivendev/Swashbuckle.AspNetCore/issues/1151
I had this same issue and I ended up adding some logic in my document processor as a workaround. I kept the &
so that I could search and replace it.
Note: I'm using NSwag which references Swashbuckle libraries, but should be same or close to same code.
In my code comment (note the <remarks>
section where I use &
) :
/// <summary>
/// Get items in cart
/// </summary>
/// <remarks>
/// api/cart?page=1&size=3
/// </remarks>
In my Startup.cs (ConfigureServices) I add the use of a Document Processor :
// sets swagger spec object properties
services.AddOpenApiDocument(s => s.DocumentProcessors.Add(new SwaggerDocumentProcessor()));
In my Document Processor :
public class SwaggerDocumentProcessor : IDocumentProcessor
{
public Task ProcessAsync(DocumentProcessorContext context)
{
context.Document.Info.Title = "My API Title";
context.Document.Info.Version = "v1.4";
foreach (var path in context.Document.Paths)
{
foreach (var item in path.Value.Values)
{
item.Description = item.Description.Replace("&", "&");
}
}
context.Document.Info.Description = "Description with markdown";
context.Document.Info.ExtensionData = new ConcurrentDictionary<string, object>();
context.Document.Info.ExtensionData.Add("x-logo", new
{
url =
"https://www.logos.com/mylogo.jpg",
altText = "Logo",
href = "https://website.com/"
});
return Task.CompletedTask;
}
}
In the Document Processor above, note these lines of code:
foreach (var path in context.Document.Paths)
{
foreach (var item in path.Value.Values)
{
item.Description = item.Description.Replace("&", "&");
}
}
Basically what it's doing is that within the Document.Paths
(the url GET, POST, DELETE, etc examples) of the API spec document, it searches and replaces all the &
instances with just &
.