Model/harness
Khalim Conn-Kowlessar 5e4906dd70 feat(modelling): cohort builder + cohort-level ParityReport
build_property_reports models a dump in order (errors captured per-cert);
parity_report_for aggregates the lodged-vs-calculated SAP across the cohort
into the existing ParityReport (MAE/RMSE/bias/worst-N), excluding certs that
couldn't be mapped or scored. Residual convention is the calculator's own
(predicted - actual), the negative of PropertyReport.sap_error.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-04 11:13:04 +00:00
..
__init__.py feat(modelling): sense-check table for a Plan in the DB-less harness 2026-06-04 08:06:53 +00:00
cohort.py feat(modelling): turnkey offline cohort script (tables + CSV) 2026-06-04 09:30:53 +00:00
console.py feat(modelling): robust offline modelling inspection (run_modelling) 2026-06-04 09:19:18 +00:00
plan_table.py feat(modelling): wire Valuation Uplift onto the Plan 2026-06-04 08:59:04 +00:00
report.py feat(modelling): cohort builder + cohort-level ParityReport 2026-06-04 11:13:04 +00:00
sample_catalogue.json feat(modelling): robust offline modelling inspection (run_modelling) 2026-06-04 09:19:18 +00:00