CODEWITHSUNDEEP
c#restencodingaws-lambda
Robin Day
Robin Day

Reputation: 102518

Encoding issue with file on AWS Lambda/API Gateway

I'm producing a REST API that does some file conversion / processing.

My Visual Studio 2015 and building on an AWS Serverless Core - ASP.Net Core Web API template.

I'm running some initial test methods and have encountered what appears to be an encoding issue.

My controller has the following. It simply pulls the posted file into a byte array via a memory stream and then passes it back. (The final application will process the byte array)

    [HttpPost]
    public IActionResult Post(IFormFile file)
    {
        var inputStream = new MemoryStream();
        file.CopyTo(inputStream);

        var fileBytes = inputStream.ToArray();

        var outputStream = new MemoryStream(fileBytes);

        return File(outputStream, "application/octet-stream");
    }

I then have a test application I'm using to pass a file to this controller and save the return.

        using (var client = new HttpClient())
        {
            using (var content = new MultipartFormDataContent())
            {
                var inputFileStream = new FileStream(Server.MapPath("~/App_Data/InputFile.pdf"), FileMode.Open, FileAccess.Read);
                var inputFileBytes = new Byte[inputFileStream.Length];
                inputFileStream.Read(inputFileBytes, 0, inputFileBytes.Length);
                inputFileStream.Close();

                content.Add(new ByteArrayContent(inputFileBytes), "file", "InputFile.pdf");

                var requestUri = "http://localhost:5000/api/controller";
                //var requestUri = "https://xxxxxxxxxx.execute-api.eu-west-1.amazonaws.com/Prod/api/controller";

                var result = client.PostAsync(requestUri, content).Result;
                var resultStream = result.Content.ReadAsStreamAsync().Result;

                var memoryStream = new MemoryStream();
                resultStream.CopyTo(memoryStream);

                var outputFileBytes = memoryStream.ToArray();
                FileStream outputFileStream = new FileStream(Server.MapPath("~/App_Data/OutputFile.pdf"), FileMode.Create, FileAccess.ReadWrite);
                outputFileStream.Write(outputFileBytes, 0, outputFileBytes.Length);
                outputFileStream.Close();
            }
        }

When I run using the localhost application, the duplicate file is saved back. However, when I publish the API to AWS, the file returned is exactly double in size to the original, certainly indicated an encoding issue.

If I pass an ANSI text file with the contents TEST then the saved file contains VEVTVA==

Can someone point me at where I should be setting any encoding settings and any suggested settings to ensure that the output stream from my HttpClient is the same as my input?

Upvotes: 3

Views: 1368

Answers (1)

Mikes3ds
Mikes3ds

Reputation: 600

So I noticed this as well with a AWS ASP.Net Core Web API. I changed the MIME type from application/octet-stream to application/text and that seemed to fix it on AWS.

    [HttpPost]
    public IActionResult Post(IFormFile file)
    {
        var inputStream = new MemoryStream();
        file.CopyTo(inputStream);

        var fileBytes = inputStream.ToArray();

        var outputStream = new MemoryStream(fileBytes);

        return File(outputStream, "application/text");
    }

Upvotes: 1

Related Questions