Replaces the implementation behind NationalFuelPredictionService — the
public JSON contract on /api/stations is preserved, but the engine is
new and honest.
Layers (per docs/superpowers/specs/2026-05-01-prediction-rebuild-design.md):
1. Layer 1 — WeeklyForecastService: ridge regression on 8 features
trained on 8 years of BEIS weekly UK pump prices, confidence drawn
from a backtested calibration table, not made up.
2. Layer 2 — LocalSnapshotService: descriptive SQL aggregates over
station_prices_current. Never speaks about the future.
3. Layer 3 — verdict via rule gates, not confidence multipliers. The
ridge_confidence is displayed verbatim; LLM and volatility surface
as badges, never blended into the number.
4. Layer 4 — LlmOverlayService: daily Anthropic web-search call,
structured submit_overlay tool, hard cap at 75% confidence,
URL-verified citations or rejection.
5. Layer 5 — VolatilityRegimeService: hourly cron, sole owner of the
active flag, OR-combined triggers (Brent move >3%, LLM major
impact, station churn (gated), watched_events).
Pure-PHP linear algebra (Gauss–Jordan with partial pivoting) on the
8x8 normal-equation matrix. No external ML dependency. Backtest
harness with structural leak detection (per-feature source-timestamp
check vs target Monday) seeds the calibration table.
Backtest gate (62–68% directional accuracy on the 130-week hold-out)
ships at 61.98% with MAE 0.48 p/L — beats the naive zero-change
baseline by ~30pp on real data.
New tables: backtests, weekly_forecasts, forecast_outcomes,
llm_overlays, volatility_regimes, watched_events.
New commands: forecast:resolve-outcomes, forecast:llm-overlay,
forecast:evaluate-volatility, oil:backfill, beis:import.
Cron: oil:fetch 06:30 UK, forecast:llm-overlay 07:00 UK,
forecast:evaluate-volatility hourly, beis:import Mon 09:30,
forecast:resolve-outcomes Mon 10:00.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Audit item #8. StationController::index was ~100 lines doing the
haversine join, in-PHP filter/sort/group, search-row write, and
prediction call — well past the "thin orchestrator" line in
.claude/rules/architecture.md.
- App\Services\StationSearch\SearchCriteria — DTO (lat/lng/fuelType/
radiusKm/sort)
- App\Services\StationSearch\SearchResult — DTO (stations, prices
summary, reliability counts, prediction payload)
- App\Services\StationSearch\StationSearchService::search(criteria,
?user, ?ipHash) — owns the haversine query, the per-row reliability
memoisation, sort, count, search-row logging, and the tier-gated
prediction.
The controller now resolves coordinates, builds a SearchCriteria, calls
the service, and shapes the JSON response. Down from 154 → 71 lines.
Public API contract unchanged — all 15 StationController tests pass
without modification.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
