Drop x and y from the UK area lists and from constituency and local-authority results #726

Current behaviour

.py's packaged UK area lists, src/policyengine/data/uk_geography/constituencies_2024.csv and local_authorities_2023.csv, have x and y columns: each area's integer position on a hex map. compute_longwise_uk_geography_impacts copies them into every result row of compute_uk_constituency_impacts, compute_uk_local_authority_impacts, ConstituencyImpact and LocalAuthorityImpact, as "x" and "y" (None for codes the lookup does not list).

Why remove them

The PolicyEngine app (PolicyEngine/policyengine-app-v2) now places constituencies and local authorities on its hex maps from its own layout files, so nothing downstream should depend on .py's positions. A second copy of the layout in .py can drift from the app's.

Intended result

• constituencies_2024.csv has the columns code,name,region,country, and local_authorities_2023.csv has code,name. The rows, codes, names and order are unchanged.
• Constituency and local-authority result rows have no x or y keys. A lookup CSV passed as constituency_csv_path or local_authority_csv_path may still have x and y columns; they are ignored.
• SOURCES.md, the two functions' docstrings and the docs no longer describe hex positions.

Merge order

The simulation API's executor reads x and y from .py's constituency list and refuses result rows without them until PolicyEngine/policyengine-sim-api#919 is deployed, so the change must merge only after that deployment. It also waits until after the UK Budget on 28 October 2026.