Labor Repository Extraction Control Workflow

Labor Repository Extraction Control Workflow A workflow diagram generated by Archify. 01 / Operator Console 02 / Discovery & Download 03 / Extraction Workers EX / QA & Recovery 05 / History & Outputs Control Collect Interpret + verify Run controls Parallel worker pool Quality gates Durable outputs Start Run · Operator Console › Run controls › Control Start Run Control Plane · Operator Console › Run controls › Control Control Plane Paused · Operator Console › Run controls › Collect Paused Cancelled · Operator Console › Run controls › Collect Cancelled Report Export · Operator Console › Interpret + verify Report Export Page Discovery · Discovery & Download › Collect Page Discovery File Collector · Discovery & Download › Collect File Collector Parser Workers · Extraction Workers › Parallel worker pool › Collect Parser Workers Spanish OCR Worker · Extraction Workers › Parallel worker pool › Interpret + verify Spanish OCR Worker Field Normalizer · Extraction Workers › Parallel worker pool › Interpret + verify Field Normalizer Self-Heal Queue · QA & Recovery › Quality gates › Collect Self-Heal Queue Human Review · QA & Recovery › Quality gates › Interpret + verify Human Review QA Gates · QA & Recovery › Quality gates › Interpret + verify QA Gates Versioned Store · History & Outputs › Durable outputs › Interpret + verify Versioned Store cancel pause page + file URLs new or changed files text + tables scan detected candidate record OCR text + score resume versioned records pass failed gate low confidence retry stage approved / corrected start Legend User UI Agent logic Policy Tool action Context / trace Cloud service

QA Contract

  • • Every record keeps source page, file URL, hash, timestamp, and extraction method
  • • Schema, provenance, duplicate, and confidence checks gate publication

Operator Contract

  • • Pause and cancel stop at durable checkpoints
  • • Retry can target one item or one failed stage; review can approve or correct uncertain fields