Skip to content

Profile Audit Guide

Guide Maps

graph LR
  family["Reproducible Research"]
  program["Deep Dive Snakemake"]
  guide["Capstone docs"]
  section["PROFILE_AUDIT_GUIDE"]
  page["Profile Audit Guide"]
  proof["Proof route"]

  family --> program --> guide --> section --> page
  page -.checks against.-> proof
flowchart LR
  profiles["profiles/local, ci, slurm"] --> audit["make profile-audit"]
  audit --> bundle["profile audit bundle"]
  bundle --> compare["Compare dry-runs and profile files"]
  compare --> decision["Separate policy from workflow meaning"]

This guide explains how to review execution policy honestly. The key question is not whether the profiles differ. The key question is whether they differ only in policy, not in workflow meaning.

Use CONFIG_CONTRACT_GUIDE.md first when the confusion is about config truth in general, not profile comparison specifically.

Primary Review Route

  1. Run make profile-audit.
  2. Run make profile-summary when you want the shortest comparison of shared and differing settings.
  3. Read route.txt, review-questions.txt, and profile-summary.json.
  4. Compare profiles/local/config.yaml, profiles/ci/config.yaml, and profiles/slurm/config.yaml.
  5. Compare local-dryrun.txt, ci-dryrun.txt, and slurm-dryrun.txt.

Back to top

What The Audit Should Prove

  • executor and resource policy are visible as configuration
  • the workflow plan remains semantically stable across operating contexts
  • profile differences do not become hidden analytical changes

Back to top

Review Questions

  • Which profile differences are expected because of execution context?
  • Which profile difference would worry you because it smells like semantic drift?
  • Which profile should a maintainer inspect first before approving a migration to new infrastructure?

Back to top