f5bfa3dca5
All checks were successful
CI / changes (pull_request) Successful in 9s
CI Trade-In / changes (pull_request) Successful in 10s
CI Trade-In / frontend-checks (pull_request) Has been skipped
CI / backend-tests (pull_request) Has been skipped
CI / frontend-tests (pull_request) Has been skipped
CI / openapi-codegen-check (pull_request) Has been skipped
CI Trade-In / backend-tests (pull_request) Successful in 1m45s
M2 root-cause fix already landed in PR #2152/#2141 — honest-ratio recompute runs unconditionally after all median mutations. Gap found: segment-multiplier (#2255) mutates both sides after that recompute with independent rounding, untested at the ratio-invariant level. Added regression guard. M3 — median_bargain_pct (house-level exposure discount) and asking_to_sold_ratio (estimate-level asking-vs-sold) are genuinely different metrics with no honest common basis. Chose explicit cross-referencing docstrings over merging. |
||
|---|---|---|
| .. | ||
| app | ||
| data/sql | ||
| scripts | ||
| tests | ||
| .dockerignore | ||
| Dockerfile | ||
| pyproject.toml | ||