another bugfix

This commit is contained in:
2026-03-12 20:29:29 +11:00
parent d1237eed44
commit 20316cee91
8 changed files with 293 additions and 23 deletions

View File

@@ -18,7 +18,7 @@ Priority key: `P0` = critical/blocking, `P1` = important, `P2` = later optimizat
- [x] [P1] Extract reusable dataset-builder logic from training script into a maintainable module/workflow.
- [x] [P1] Add lag/rolling features (means, stddev, deltas) for core sensor inputs.
- [x] [P1] Encode wind direction properly (cyclical encoding).
- [ ] [P2] Add calendar features (hour-of-day, day-of-week, seasonality proxies).
- [x] [P2] Add calendar features (hour-of-day, day-of-week, seasonality proxies). (`feature-set=extended_calendar`)
- [x] [P1] Join aligned forecast features from `forecast_openmeteo_hourly` (precip prob, cloud cover, wind, pressure).
- [x] [P1] Persist versioned dataset snapshots for reproducibility.
@@ -34,7 +34,7 @@ Priority key: `P0` = critical/blocking, `P1` = important, `P2` = later optimizat
## 5) Evaluation and Reporting
- [x] [P0] Report ROC-AUC, PR-AUC, confusion matrix, precision, recall, and Brier score.
- [x] [P1] Compare against naive baselines (persistence and simple forecast-threshold rules).
- [ ] [P2] Slice performance by periods/weather regimes (day/night, rainy weeks, etc.).
- [x] [P2] Slice performance by periods/weather regimes (day/night, rainy weeks, etc.). (`sliced_performance_test`)
- [x] [P1] Produce a short model card (data window, features, metrics, known limitations). (`--model-card-out`)
## 6) Packaging and Deployment
@@ -42,7 +42,7 @@ Priority key: `P0` = critical/blocking, `P1` = important, `P2` = later optimizat
- [x] [P0] Implement inference path with feature parity between training and serving.
- [x] [P0] Add prediction storage table for predicted probabilities and realized outcomes.
- [x] [P1] Expose predictions via API and optionally surface in web dashboard.
- [ ] [P2] Add scheduled retraining with rollback to last-known-good model.
- [x] [P2] Add scheduled retraining with rollback to last-known-good model. (`run_rain_ml_worker.py` candidate promote + `RAIN_MODEL_BACKUP_PATH`)
## 7) Monitoring and Operations
- [x] [P1] Track feature drift and prediction drift over time. (view: `rain_feature_drift_daily`, `rain_prediction_drift_daily`)