Finish Scottish Census 2022 tenure inputs for BRMA rent targets #326

The Scottish Census 2022 tenure input from #313 section 5 remains outside PR #314. This follow-up tracks source acquisition, registration and emitted facts for Scottish private-renter household counts used by PolicyEngine/microcosm#1090, PolicyEngine/microcosm#1092 and PolicyEngine/microcosm#1123.

Ward tenure by bedrooms

The original request is the publisher table-builder export from household2022ftb: all 355 electoral wards on 2022 boundaries, households by detailed tenure and bedrooms.

• Publisher: Scotland’s Census table builder.
• Microcosm #1092 pins census_2022_ward2022_tenure_bedrooms.csv, SHA-256 03d6157d34bd3df44e79ded1c03510954dee55e9e05d3c276b26ad463a565677.
• The original CSV was not found in the inspected repositories, releases or caches. The publisher export requires guest-terms acceptance. Obtain the cached original or an authorized new export; a new export must retain its actual acquisition date and checksum.
• Register the immutable artifact and emit direct publisher cells with tenure, bedroom category, ward code, reference period, universe and source-cell guards. Preserve the census date of 20 March 2022 separately from publication and acquisition dates.

Public output-area count alternative

Issue #313 explicitly offers output-area tenure alone as an alternative for the renter counts. National Records of Scotland publishes UV404 for 2022 output areas under OGL v3:

• Publisher landing page.
• Public UV404 CSV.
• CSV SHA-256: d6abb0ee1a8f21e17160a517088b1ca62ddcb728a35bf5ecbf82c7f400b6378f. The UK Data Service copy matches the publisher ZIP member byte for byte.
• The table contains 46,363 output areas and counts all occupied households. Physical column I is the directly published Private rented: Total; use that cell rather than summing its detail categories.
• The private-rented total has 41,453 numeric cells and 4,910 - markers. Confirm the marker meaning against primary publisher documentation and preserve the original symbols; no primary zero convention has been established.

PR #314 links this download but does not register it or emit its output-area facts. UV404 can support bedroom-independent renter counts and avoid proportional ward splitting. It does not provide a tenure-by-bedroom joint distribution.

Completion criteria

• Resolve the joint-table request by acquiring and ingesting its direct cells, or record an explicit decision to use UV404 for the count requirement while retaining the joint-bedroom limitation.
• Register the chosen original artifact and emit guarded facts with the publisher’s geography vintage, census date, household universe, units, labels and lineage.
• If using UV404, resolve marker handling from primary evidence and preserve direct published totals and detail cuts without reconciliation.
• Add source regressions for coverage, physical-cell provenance, missing markers and dimensional validity; document coverage and remaining limitations in the source catalogue.

BRMA assignment, aggregation, uprating and calibration target selection remain in Microcosm. Chronicle must not synthesize joint counts from independent tenure and bedroom margins or reconstruct publisher originals from derived BRMA outputs.

Remaining source work split from #313; implementation of the other requested inputs is in #314.