col: simulator auto-adjusts spending to local prices via Numbeo+Expatistan

The Monte Carlo used to compare jurisdictions at a flat London-equivalent
spend, which silently overstated the cost-of-living for any move to a
cheaper region. Now every cross-jurisdiction simulation auto-scales
spending_gbp by the real Numbeo/Expatistan ratio between the user's
baseline city and the target city.

Architecture:
- fire_planner/col/baseline.py — 22 cities with headline Numbeo data
  (source URLs + snapshot dates embedded) — fallback when scraper fails
- col/numbeo.py + col/expatistan.py — httpx async scrapers, regex-parsed,
  polite 1.1s rate-limit, EUR/USD anchored
- col/cache.py — PG-backed cache (col_snapshot table, 1-year TTL)
- col/service.py — sync compute_col_ratio() for the simulator; async
  lookup_city_cached() with source reconciliation for the refresh CronJob
- alembic 0005 — col_snapshot table, UNIQUE(city_slug, source_name)

Simulator wiring:
- SimulateRequest gains col_auto_adjust=True (default), col_baseline_city,
  col_target_city. Defaults pick the jurisdiction's representative city.
- _resolve_col_adjustment scales spending_gbp before path-building.
- SimulateResult surfaces col_multiplier_applied + col_adjusted_spending_gbp.

CLIs:
- python -m fire_planner col-seed — loads BASELINES into col_snapshot
  (post-migration seed step)
- python -m fire_planner col-refresh-stale --within-days 7 — used by the
  weekly fire-planner-col-refresh CronJob

268 tests pass. Mypy strict + ruff clean.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

fire_planner/api/schemas.py

@@ -504,6 +504,22 @@ class SimulateRequest(BaseModel):
     annual_real_adjust_pct: Decimal = Decimal("0")
     guardrail_threshold_pct: Decimal | None = None
     guardrail_cut_pct: Decimal = Decimal("0.10")
+    # Cost-of-living auto-adjust: when `col_auto_adjust=True`, the
+    # simulator looks up COL ratio (target/baseline) from `fire_planner.col`
+    # and scales `spending_gbp` BEFORE running paths. Defaults to True so
+    # cross-jurisdiction comparisons are honest by default — earlier
+    # comparisons used hand-wave 0.5x/0.75x multipliers, which were
+    # consistently optimistic vs. actual Numbeo data (Bulgaria is 0.41x,
+    # not 0.50x; Cyprus 0.67x, not 0.75x).
+    #
+    # `col_target_city` defaults to the jurisdiction's representative
+    # city (uk→london, cyprus→limassol, etc.). Set explicitly to anchor
+    # on a different city (e.g. `cyprus`+`paphos` if Limassol is too
+    # expensive a proxy). For `jurisdiction='nomad'` there is no
+    # representative city and auto-adjust is skipped silently.
+    col_auto_adjust: bool = True
+    col_baseline_city: str = "london"
+    col_target_city: str | None = None
 
 
 class SimulateResult(BaseModel):
@@ -516,6 +532,13 @@ class SimulateResult(BaseModel):
     elapsed_seconds: Decimal
     yearly: list[ProjectionPoint]
     goals_probability: list[GoalProbability] = Field(default_factory=list)
+    # When `col_auto_adjust=True`, surface the applied multiplier + the
+    # COL-adjusted spending so the user can see what was used. Null when
+    # auto-adjust was off, jurisdiction had no representative city
+    # (nomad), or baseline==target (London-to-London).
+    col_multiplier_applied: Decimal | None = None
+    col_adjusted_spending_gbp: Decimal | None = None
+    col_target_city: str | None = None
 
 
 class CompareRequest(BaseModel):