Files
fuel-alert/app/Services/BrentPriceFetcher.php
Ovidiu U ddd591ad47 feat(forecasting): build calibrated weekly forecast stack with LLM overlay and volatility detector
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>
2026-05-03 08:40:05 +01:00

65 lines
1.7 KiB
PHP

<?php
namespace App\Services;
use App\Models\BrentPrice;
use App\Services\BrentPriceSources\BrentPriceFetchException;
use App\Services\BrentPriceSources\EiaBrentPriceSource;
use App\Services\BrentPriceSources\FredBrentPriceSource;
final readonly class BrentPriceFetcher
{
public function __construct(
private readonly EiaBrentPriceSource $eia,
private readonly FredBrentPriceSource $fred,
) {}
/**
* Fetch from EIA and persist. Throws on failure.
*/
public function fetchFromEia(): void
{
$rows = $this->eia->fetch();
if ($rows === null) {
throw new BrentPriceFetchException('EIA fetch returned no data');
}
BrentPrice::upsert($rows, ['date'], ['price_usd']);
}
/**
* Fetch from FRED and persist. Throws on failure.
*/
public function fetchFromFred(): void
{
$rows = $this->fred->fetch();
if ($rows === null) {
throw new BrentPriceFetchException('FRED fetch returned no data');
}
BrentPrice::upsert($rows, ['date'], ['price_usd']);
}
/**
* One-shot Brent backfill via FRED's observation_start/end. Used to
* seed `brent_prices` going back to 2018 so Phase 9's volatility
* detector and Phase 8's LLM overlay have proper context.
*
* @return int rows inserted/updated
*/
public function backfillFromFred(string $from, string $to): int
{
$rows = $this->fred->fetchRange($from, $to);
if ($rows === null) {
throw new BrentPriceFetchException("FRED backfill ({$from}{$to}) returned no data");
}
BrentPrice::upsert($rows, ['date'], ['price_usd']);
return count($rows);
}
}