MT cycle-1 review remediation: dwelling references, household age semantics, provenance metadata

Task metadata

Latest summary

Built MT cycle-1 remediation bundle /home/synthestat/output/runs/MT/mt_population_review_cycle1_8f28c2dd_seed420987 and addressed the reviewer blockers: dwelling household refs, HH_SINGLE age semantics, enriched provenance statuses, and seeded degeneracy diagnostics. Verified contract/readability and HARD pass_exact status, ran tests/test_release_review.py (9 passed), and created synth-reviewer follow-up t_f47640f3.

Body

Reviewer verdict for MT population synthesis QA cycle 1 is NEEDS_MODEL_FIX.

Review report: /home/synthestat/workspace/manager_handoffs/reviewer/2026-05-19_164843Z_MT_population_review_cycle1.md
Bundle reviewed: /home/synthestat/output/runs/MT/mt_population_review_cycle1_0a09edfb_seed420987
Run ID: mt_population_review_cycle1_0a09edfb_seed420987

Blocking fixes requested:
1. Fix dwelling-household reference emission and/or validation semantics: synthetic_dwellings.parquet has household_id null for all 8 occupied/observed dwellings, while households and synthetic_building_assignments.parquet link those dwellings to households and diagnostics imply occupied-dwelling household reference checks pass.
2. Fix household type/member-age semantics: HH_SINGLE_E/M/Y are emitted but all persons are born 1991-06-15. Either generate ages coherent with the labels, or provide/encode an explicit codebook showing these labels are not age-coded and adjust validation accordingly.
3. Strengthen source_provenance.json to include per-source URL or explicit manual/local data_uri status, retrieval timestamp or explicit seeded/manual timestamp status, licence/license, reference period, and source universe for registry/source entries.
4. Add/strengthen seeded-slice degeneracy diagnostics: repeated birth_date and identical socioeconomic person attributes should be explicitly marked as toy seeded limitations, not distribution-realism evidence.

After remediation, produce a new contract-complete review bundle and hand off to synth-reviewer.

Runs

Events

Comments

modeler remediation handoff:
{
  "country": "MT",
  "new_run_id": "mt_population_review_cycle1_8f28c2dd_seed420987",
  "bundle_path": "/home/synthestat/output/runs/MT/mt_population_review_cycle1_8f28c2dd_seed420987",
  "builder_path": "/home/synthestat/workspace/manager_handoffs/modeller/build_mt_review_bundle.py",
  "handoff_files": [
    "/home/synthestat/workspace/manager_handoffs/modeller/2026-05-19_1655_build_status.md",
    "/home/synthestat/workspace/manager_handoffs/modeller/2026-05-19_1655_missing_requirements.md",
    "/home/synthestat/workspace/manager_handoffs/modeller/latest.md"
  ],
  "remediations": [
    "Backfilled synthetic_dwellings.household_id for all occupied/observed dwellings from synthetic_households dwelling links.",
    "Added HH_SINGLE_Y/M/E codebook and remediated seeded birth dates to representative young/middle/elderly ages.",
    "Enriched source_provenance.json with source_records and per-registry explicit source_url/data_uri/retrieval/licence/reference_period/source_universe statuses.",
    "Added seeded_degeneracy_diagnostics and explicit guardrails that repeated/generated attributes are toy seeded limitations, not distribution-realism evidence."
  ],
  "verification": {
    "contract_files_checked": 16,
    "hard_constraint_status": "pass_exact",
    "null_dwelling_household_ids": "0/8",
    "bad_dwelling_household_refs": 0,
    "person_birth_dates": ["1951-06-15", "1981-06-15", "2001-06-15"],
    "source_records": 8,
    "registry_entries_with_required_status_fields": 38,
    "pytest": "python -m pytest tests/test_release_review.py -q => 9 passed"
  },
  "reviewer_child_task": "t_f47640f3"
}