Store cache contents into an S3 bucket

[csala]

• SDGym version: 0.3.0

Description

Similarly to what is described on #80, it should be possible to store the cache contents into an S3 bucket.

The behavior would be similar to the results path, where one can specify the s3:// prefix in the cache_dir path specification to trigger the S3 storage.

Python

sdgym.run(..., cache_dir='s3://my-bucket/path/to/my/cache/dir')

CLI

sdgym run ... -c s3://my-bucket/path/to/my/cache/dir

The collect command introduced in the PR #78 should also be adapted to read the cache contents from S3 and store the resulting CSV file to S3.