Convert two bare db.rollback() calls on the shared per-request session
(best_layouts _fetch_supply_only_layouts + Step 2b core-attribution) to
db.begin_nested() — a bare rollback there discards ALL uncommitted work of
the request and orphans the outer SessionTransaction. Also wrap the swallow
sites in market_metrics (_query_stock, _query_sales_window, _price_sensitivity,
compute_offer_price_trend), best_layouts step-5 supply, future_supply
_query_supply_rows, own_portfolio _query_current/_query_future, and
custom_pois get_overlaps_for_scoring so a failed statement no longer poisons
the shared tx for later steps. Per-site regression tests assert fallback
preserved, begin_nested used, bare rollback NOT called, and session usable
after failure.
Refs #2464 (cluster A session-poisoning, Wave 2).
EOF2
echo "HANDLED"
Cold §22 forecast measured ~215-233s on prod: §9.x layers re-execute the same
horizon/segment-invariant DB loads with identical args hundreds of times per
report (profiled: get_competitors x69, market_metrics x124, get_monthly_macro
x290). Add a per-report ContextVar cache (forecast_cache(), opened once in the
orchestrator) + @cached(key_builder) on the expensive §9.x loaders so each
unique load runs ONCE and reuses the same frozen, read-only instance.
Output is byte-identical (memoized producers are frozen dataclasses / read-only
Pydantic, callers never mutate; cache is per-report, discarded on exit; no-op
outside the report build). No concurrency, no signature changes.
- forecast_request_cache.py: ContextVar cache + cached() decorator (no-op
outside context, reentrant, _MISS sentinel for cached None)
- @cached on competitors/future_supply/market_metrics/macro_series/
sales_series/macro_coefficient/demand_normalization/regression loaders
- orchestrator: wrap build_site_finder_report in forecast_cache()
- 58 tests: key discrimination (call-counting regression guard), no-op-outside,
per-report isolation, reentrancy, frozen-producer canary, amplification proof
(real get_monthly_macro xN->1)
code-reviewer APPROVE (keys correct, mutation-safe, output identical). 1265
forecast/cache tests green. No new deps. Refs #1129.
