Skip to content

current-project-status.md

Latest commit

 

History

History
67 lines (49 loc) · 2.6 KB

current-project-status.md

File metadata and controls

67 lines (49 loc) · 2.6 KB
Raw

Current Project Status

Single reference for what is built, what is deferred, and what matters for portfolio vs production use.

Last aligned with the codebase: dashboard tabs, ranking API, runs/diagnostics/export, and compare diff are implemented.

Portfolio-ready (built)

  • Monorepo scaffolding, scripts, Docker Compose, PostgreSQL, CI
  • PropFlux ingestion (partial accept), dedup, raw/normalized/rejected persistence
  • Baseline (baseline_v1) and advanced (advanced_v2) scoring with explainability payloads
  • Dataset validation CLI and persisted validation summaries
  • Strategy profiles (scoring_profiles.yaml) and preset resolution
  • Ranking API: query, listing detail, profiles
  • Persisted ranking runs, run list/detail, CSV/JSON export (optional full listing detail)
  • Diagnostics summary API
  • Dataset sources API with status/search filters
  • CLI parity: rank-query, listing-detail, profiles-list, profile-show, evaluation/benchmark commands
  • Dashboard: Control Panel, Source Library, Runs (compare diff), Run detail, Diagnostics
  • Operator docs: README, demo.md, dashboard.md, configuration.md glossary
  • Backend test suite (~90 pytest cases); frontend lint/typecheck/build + unit tests

Deferred (documented, not required for portfolio)

  • LLM-derived scoring signals (gated in roadmap)
  • PDF export
  • Authentication / multi-tenant dashboard
  • Automated profile promotion in production
  • Cross-source entity resolution beyond dedup keys
  • Production deployment, observability, and SLO runbooks
  • Large real-dataset validation report pack (Week 4 hardening loop)

Known implementation notes

Topic State
profile_version Placeholder v1 in API; use profile_row_id + backup for reproducibility
Home route / Redirects to /dashboard/control (no marketing landing page)
Demo data scripts/demo-local.sh uses backend/tests/fixtures/propflux/valid_listings.json
Scoring tuning Ongoing manual iteration on weights/presets—not a code completion item

Verify after clone

./scripts/setup.sh
./scripts/lint.sh
./scripts/test.sh
./scripts/demo-local.sh    # optional: requires Postgres

See setup-checklist.md and demo.md.

Related documents