Only retrieve files from a single S3 bucket
URLs used to be of the form:
Now they are of the form:
The S3 bucket must be specified ahead of time via the S3_BUCKET
This allows our code to be considerably simpler.