Backfill reconstructs authoring metadata for process definitions that already exist. It does not change the source process.
python .pf/runtime/bin/pf.py process-authoring-import --project-root . --process task-batch-execution --applyThe command writes:
.pf/authoring/backfill/processes/<process-id>/answers.yaml.pf/authoring/backfill/processes/<process-id>/draft.process.yaml.pf/authoring/backfill/processes/<process-id>/source.process.yaml.pf/authoring/backfill/processes/<process-id>/semantic-map.yaml.pf/authoring/backfill/processes/<process-id>/unsupported-fields.yaml.pf/authoring/backfill/processes/<process-id>/import-report.md
Unsupported fields are not automatically failures. They mean the current authoring answers model does not have a direct field for that source data. Review them before release and decide whether to add authoring coverage, keep a warning, or narrow the process source.
Run a parity check after import:
python .pf/runtime/bin/pf.py process-parity-check --project-root . --process task-batch-executionReports are written under .pf/reviews/parity/ and .pf/artifacts/parity/.