Stop generating synthetic weather data

docs/UPDATE_ROADMAP.md

@@ -173,10 +173,21 @@ running in parallel until each real source is proven, not cut over in one step.
       falls back to older synthetic data via the existing stale-observation
       UI rather than breaking), and Kartes' aggregate query API returning
       sane blended real+synthetic values.
-- [ ] Let the open-data fetch run for a trial period, then remove the private
-      FTP path (`fetch/lvgmc/FetchService.scala`'s `fetchWeatherStations`,
-      its `LVGMC_*` env vars) and the synthetic seed machinery
-      (`dev/seed_weather.sql`, the `seed` Compose service) once confident.
+- [x] Stop synthetic data generation now that real data is verified and this
+      tool may plausibly be used against live broadcasts, where mixed
+      real/fake data — both shown identically as "older observation" — is a
+      real hazard, not just messiness: removed `dev/seed_weather.sql`, the
+      `seed` Compose service, and the README section describing it.
+- [ ] Wipe the synthetic rows already sitting in the `weather` table (a
+      separate step from stopping generation — the seed already ran once
+      before this decision) so history going forward is real data only, not
+      a mix. Cities with no matching open-data station (Valmiera, Cēsis)
+      should show no data rather than falling back to old synthetic rows —
+      deferred by design, revisit once a real source is found or accepted as
+      permanently unavailable.
+- [ ] Once the open-data fetch has run for a trial period, remove the
+      private FTP path (`fetch/lvgmc/FetchService.scala`'s
+      `fetchWeatherStations`, its `LVGMC_*` env vars) if it remains unused.
 - [ ] Register free DMI Open Data credentials and validate DMI STAC discovery
       and EDR GRIB downloads (code already exists and is already scheduled;
       this is credential registration and verification, not new code —