[Fixed]-Django as S3 proxy

10👍

Rather than proxying, why not simply redirect?

Use the django view at www.myproject.com/downloads/1 to serve a HTTP Redirect to the S3 storage URL – it’s possible to generate time-limited authenticated URLs e.g. see here https://docs.aws.amazon.com/AmazonS3/latest/dev/ShareObjectPreSignedURL.html

Then the client downloads the file directly from S3, but the contents is still secure and access has to come through your Django application

20👍

Yes, you can use Django as a proxy to stream your S3 files to the user.

  1. Generate a standard S3 url for your file.
  2. Use the requests library to fetch your file in streaming mode.
  3. Use Django to serve up the raw stream to the user.
import requests, boto3
from django.views.generic import View
from django.http.response import HttpResponse

class MediaDownloadView(View):

    def get(self, request, *args, **kwargs):
        s3 = boto3.resource(
            service_name='s3', aws_access_key_id='XXX_AWS_ACCESS_KEY_ID',
            aws_secret_access_key='XXX_AWS_ACCESS_KEY',
        )
        url = s3.meta.client.generate_presigned_url(
            ClientMethod="get_object", ExpiresIn=3600,
            Params={
                "Bucket": 'XXX_BUCKET_NAME',
                "Key": 'XXX_OBJECT_KEY',
            },
        )
        r = requests.get(url=url, stream=True)
        r.raise_for_status()
        response = HttpResponse(r.raw, content_type='audio/mp3')
        response['Content-Disposition'] = 'inline; filename=music.mp3'
        return response

Leave a comment