Better way to download files directly using Amazon S3 API - SDK on ASP.MVC - asp.net-mvc

Question

Amazon provides a vast documentation, but there are so many docs that I'm lost, so here is my current service for upload/download files. Upload works as expected but on the download its where I have to download the files to a physical path and later serve the download to the user, I don't have much experience working with streams. Here is the FileManagerService class that connects to Amazon API.
using Amazon.S3;
using Amazon.S3.Model;
public class FileManagerService
{
public FileManagerService()
{
string serverPath = HttpContext.Current.Server.MapPath("~/");
string uploadPath = Path.Combine(serverPath, "FileUploads");
Directory.CreateDirectory(uploadPath);
UploadDirectory = uploadPath;
}
private string UploadDirectory { get; set; }
private docucloudEntities db = new docucloudEntities();
private IAmazonS3 S3Client = new AmazonS3Client();
private string S3Bucket = "bucketname";
public async Task<string> DownloadFile(string AmazonFileKey, string FileName)
{
var fileRequest = new GetObjectRequest
{
BucketName = S3Bucket,
Key = AmazonFileKey
};
var localRoute = Path.Combine(UploadDirectory, FileName);
using (var fileObject = await S3Client.GetObjectAsync(fileRequest))
{
if (fileObject.HttpStatusCode == HttpStatusCode.OK)
{
fileObject.WriteResponseStreamToFile(localRoute);
}
}
return localRoute;
}
}
This method returns the string, it's not complete yet with try catch blocks, but it currently works. Here is my controller method that download the file to the client:
public class FileManagerController : Controller
{
private FileManagerService FileService = new FileManagerService();
public async Task<ActionResult> DownloadFileAmazon(long FileId)
{
if (db.Archivos.Any(i => i.ArchivoID == FileId))
{
var archivo = db.Archivos.Single(i => i.ArchivoID == FileId);
var rutaarchivo = await FileService.DownloadFile(archivo.Ruta, archivo.Nombre);
if (System.IO.File.Exists(rutaarchivo))
{
var fileBytes = System.IO.File.ReadAllBytes(rutaarchivo);
var response = new FileContentResult(fileBytes, "application/octet-stream");
response.FileDownloadName = archivo.Nombre;
System.IO.File.Delete(rutaarchivo);
return response;
}else
{
return HttpNotFound();
}
}else
{
return HttpNotFound();
}
}
}
So here on the controller I read the file bytes and serve the download, after deleting the file, but this could lead to a slower perfomance, its there a way of achieving direct download.

Answer 1

As far as I can tell there is no reason to dispose GetObjectResponse (return type of GetObjectAsync) even if the docs says so. GetObjectResponse is not implementing IDisposable but is inheriting StreamResponse that is. However, as far as I can tell it's only disposing the ResponseStream. So you could return the stream from GetObjectResponse (fileObject.ResponseStream) together with the ContentTypefrom the headers (fileObject.Headers.ContentType) that you then can return as a file in your controller:
[HttpGet]
[Route("blob/{filename}")]
public async Task<IActionResult> GetFile(string filename)
{
try
{
var file = await _fileStorageService.GetBlobAsync(filename);
return File(file.Stream, file.ContentType);
}
catch (Exception ex)
{
// Handle exceptions
}
}
FileResult will dispose the stream after it has written the file so there the stream will finally get disposed.