Snakemake Data Access

[acrinklaw]

I'm trying to understand how to convert my Snakemake to be compatible with AGC and the documentation is not quite clear on how exactly things happen behind the scenes. If I understand correctly, input files passed in to InputFileURLs will get copied into a provisioned EFS (in this case, my input files are on an S3 bucket that AGC is given access to). But where exactly do they end up on the EFS? My Snakemake originally relies on globbing files from a directory. Unfortunately the example for Snakemake workflows doesn't show any data transfer and I couldn't find any that people have developed.

Thanks

[acrinklaw]

For anyone who comes across this, it seems that the data gets stored under the EFS volume with whatever keyword you provide in inputs.json.

For example, if I outlined "input_files": "s3://some-bucket/*.bam", these would be copied from the S3 bucket and placed on the EFS volume under /mnt/efs/input_files/