Default Snakemake options and profiles

[victorlin]

Context

Default Snakemake options used by our pathogen repos are a mix of options passed by Nextstrain CLI and repo-specific default profile config.

List of actively used default options

Sourced from GitHub query.

The following are redundant when running the workflow via Nextstrain CLI:

• cores: all
• printshellcmds: true
• rerun-incomplete: true/ri: true (only for nextstrain run)

The following have no effect:

• reason: True (removed in v8)

Others:

• keep-going: true

Proposal

Remove the repo-specific default profiles for ease of maintenance. I think we'd just need to add --rerun-incomplete as a default for nextstrain build for parity with nextstrain run. --keep-going is context-dependent and shouldn't be a default.

The downside is less predefined defaults when using snakemake, but there's already precedent for that with the existing defaults set by Nextstrain CLI.

[joverlee521]

IIRC, the Snakemake default profile config was originally added by @corneliusroemer because he uses snakemake to run workflows instead of nextstrain build. Not sure if that's still the case since most of the workflows are now in GitHub Actions?

[victorlin]

I think it's fine to have default profile configs if they're useful to people working on the repos. The main point is that it may be confusing for users of nextstrain build to need --rerun-incomplete only in some repos. Maybe we leave the default profile configs as-is and just add --rerun-incomplete to nextstrain build?

[joverlee521]

Yeah, I think it's reasonable for add --rerun-incomplete to the flags for nextstrain build.