pyawsbuckets 1.2.0

Initialise

Initiate the access object with your credentials:

from pyawsbuckets import AwsInterface
aws_interface = AwsInterface(amazon_access_key, amazon_secret_key)

PUT object

Put an object into an existing bucket at S3 (repeat: the bucket must ALREADY exist):

aws_interface.put(
    'https',
    'bucket999',
    'somefile.pdf',
    content)

Put an object into an existing bucket at S3, activating server-side encryption for that object:

aws_interface.put(
    'https',
    'bucket999',
    'somefile.pdf',
    content,
    server_side_encryption=True)

The put method also accepts an optional content_type keyword argument which should be a standard internet media type (e.g. “image/jpeg”)

DELETE object

Delete an object from S3:

aws_interface.delete('bucket999', 'somefile.pdf')

Expiring access URL

Get a signed URL which gives access to a private object, but only for (e.g.) 15 minutes. This is useful if you wish for users to be able to download private objects directly, but only for a brief window of time (to stop, for example, link sharing):

expiring_url = aws_interface.sign_object_request('https', 'bucket999', 'somefile.pdf', 15)

Using HTTPS

You can specify ‘https’ as the protocol for putting/getting files and for signing.

However you should bear in mind that given the way wildcard certificates work, the certificate presented by the S3 service will not match the requested hostname if your bucket name contains dots (periods).

This is because a certificate for “*.s3.amazon.com” will work nicely for hostnames like “foo.s3.amazon.com” and “foo-example-com.s3.example.com” but will NOT work for hostnames like “foo.bar.s3.amazon.com”. This is a feature of wildcard certificates and their jurisdiction, not of S3 or this library.

Because the bucket name gets translated into a hostname, this matters. It matters because your request will fail.

In short: avoid dots (periods) in bucket names if you plan to use https at all.