[BUG] VEP and SnpEff cache cannot be defined in nextflow.config

[lassefolkersen]

Description of the bug

For Sarek 3.0, dev-branch, wWhen including the following in a config file

params{
   vep_cache = 's3://somewhere'
}

The expected behaviour is to not pull the VEP-container, but instead use a cache. However it does pull the (quite big) container. So the bug may show as a CannotPullContainerError: write /var/lib/docker/tmp/GetImageBlob370078793: no space left on device error, especially when running with awsbatch. The choice is made on this line, where params.vep_cache is always empty. This can be bypassed by instead providing --vep_cache as command line switch, which is why the bug is marked as low-priority and postponed for backlog.

Additional context

In this slack thread, the sarek3.0 channel of nf-core slack. Possibly related to this issue.

[maxulysse]

Since we can't see the Slack thread as it's a private one, I'll recapitulate the problem, and hopefully give some solutions for now:
We can't access from the modules.configs params defined in a custom config files, due to nextflow-io/nextflow#2662
Solution is either to specify from command line, edit the nextflow.config, or in an institutional profile.
I'm closing the issue as won't fix for now, but it's because it's closed that it's forgotten ;-)