PLEB Deep Dive

This landing page is a quick operational overview.

Use it to navigate the deep-dive subchapters by concern, instead of scanning one long monolithic page.

Quick overview

PLEB is the run-coordination layer for:

  • ingesting and structuring timing data,

  • applying deterministic dataset-fix policies,

  • running tempo2 and PQC stages,

  • applying post-QC actions,

  • optionally running whitenoise EFAC/EQUAD/ECORR estimation,

  • optionally comparing against public release par values,

  • generating run and review output files with reproducibility metadata.

A practical production pattern is:

  1. ingest into a controlled branch,

  2. run fix normalization and variant generation,

  3. run detection,

  4. apply comments/actions in a separate pass,

  5. run whitening and public-compare on cleaned outputs,

  6. publish compact reports for reviewer triage.

How to read the deep dive

  • If you are new: start with Overview and Ownership + Stage-by-Stage Flow.

  • If you are configuring runs: read Operational Config Groups and Configuration Guide.

  • If you are debugging production failures: go directly to Failure Cookbook and Recipes.

Deep-dive chapters