feat: add FORCE_EXTRACT option and pre-extraction cleanup

Dan Notesteinrequested to merge
feature/extract-cache-improvements into develop

Summary

Adds improvements to extract-test-cache.sh based on patterns from hivemind's cache extraction logic:

  1. FORCE_EXTRACT option - Set to 1 to force extraction even if valid data exists (useful for debugging cache issues)
  2. Pre-extraction cleanup - Removes existing partial data before extraction to prevent permission issues

Changes

  • Add FORCE_EXTRACT environment variable (default: 0)
  • Skip marker file, PostgreSQL running, and valid data checks when FORCE_EXTRACT=1
  • Clean destination directory before extraction
  • Update documentation

Motivation

When debugging cache issues in CI, it's useful to force re-extraction even if local data looks valid. This was already supported in hivemind via FORCE_NFS_CACHE variable.

Merge request reports