🤖 Resource Automation

🤖 Resource Automation

This project uses automation to keep Pashto, Pukhto, and Pushto AI resources current without lowering quality.

🎯 What This Automation Does

• Finds new Pashto-relevant resources from trusted public endpoints.
• Keeps one canonical machine-readable catalog.
• Promotes only candidates that pass validation + deduplication.
• Generates two search payloads:
  • technical resources (non-paper)
  • papers/documentation resources

📁 Core Files

• Canonical verified catalog: ../resources/catalog/resources.json
• Candidate feed: ../resources/catalog/pending_candidates.json
• Catalog schema: ../resources/schema/resource.schema.json
• Technical search export: search/resources.json
• Papers search export: papers/resources.json

🧰 Scripts (Manual Use)

• Validate catalog: python scripts/validate_resource_catalog.py
• Generate README + search outputs: python scripts/generate_resource_views.py
• Review existing resources: python scripts/review_existing_resources.py
• Discover new candidates: python scripts/sync_resources.py --limit 20
• Promote valid candidates: python scripts/promote_candidates.py
• Full cycle wrapper: python scripts/run_resource_cycle.py --limit 25

⚙️ GitHub Actions (Automatic Use)

• CI workflow: .github/workflows/ci.yml
  • validates catalog
  • checks generated files
  • checks markdown links
  • runs tests
• Daily resource workflow: .github/workflows/resource_sync.yml
  • audits existing entries
  • discovers candidates
  • auto-promotes valid entries
  • regenerates views/search payloads
  • opens a PR for review

🔄 Promotion Flow (Simple)

1. Review current resources for stale links.
2. Sync candidates into pending_candidates.json.
3. Promote only valid, non-duplicate, reachable candidates.
4. Regenerate outputs and validate.
5. Review and merge.

📚 Related Guide

• Full runbook: resource_cycle_runbook.md