shoudn't S3FileHandle's

read()

return an

S3FileHandle

to be consistent with the

write()

behavior? I want to construct an asset which just reads from a known location on S3, but having to construct the S3FileHandle manually feels awkward, since I have to specify the bucket and prefix again -- which are already configured in my s3_file_manager. https://github.com/dagster-io/dagster/blob/master/python_modules/libraries/dagster-aws/dagster_aws/s3/file_manager.py#L96