SPB Git

spb/wp29_uqo

Public
4.5 KB · 64 lines json
1{2 "P0": {3  "status": "passed",4  "checks": [5   "TASK.md read fully",6   "contract.md written with estimands, outputs, info rules, invalidating conditions"7  ],8  "notes": ""9 },10 "P1": {11  "status": "passed",12  "checks": [13   "SHA-256 of all 23 data files recomputed and match README",14   "row/col counts, date ranges, missing values, duplicates, gaps recorded in data_audit.md",15   "first_release.csv levels and yoy recomputed and matched shipped vintages",16   "last(V) per origin tabulated; all 20 targets present",17   "exclusions.csv written (0 removed, 80->80)"18  ],19  "notes": "RSAFS_current and first_release flagged as post-origin; used only for (d) / as actuals."20 },21 "P2": {22  "status": "passed",23  "checks": [24   "naive, snaive, AR(1) baselines computed from vintage V only; RMSE/MAE written to results.json (rmse_naive_h1=0.997, rmse_naive_h3=1.420)"25  ],26  "notes": "Baselines computed by out/analysis.py with the same evaluation rule as the model."27 },28 "P3": {29  "status": "passed",30  "checks": [31   "out/analysis.py reads only data/, writes results.json, forecasts.csv, table1.csv, figure1.png",32   "seed 20261030 fixed for bootstrap",33   "printed output consistent with results.json"34  ],35  "notes": "Model: base-effect drift, no estimated parameters, specification fixed before evaluation."36 },37 "P4": {38  "status": "blocked",39  "checks": [40   "clean re-run: results.json, forecasts.csv, table1.csv byte-identical: PASS",41   "independent recomputation of AR(1) (statsmodels), model, RMSE/MAE, DM, revision: PASS (see checks.md)",42   "first_release_vintage > origin for all 80 rows: PASS",43   "origin_date < target_date: FAIL for 40 h=1 rows by construction of TASK.md target definition (not leakage)",44   "gate run 1 receipt out/state/gate_receipts/20261030T060311.428944Z.json: failed (report/claims/phases missing at that time + forecasts_chronology)",45   "gate run 2 (end of P4/P5) receipt out/state/gate_receipts/20261030T060441.223227Z.json: 64/65 pass; only forecasts_chronology fails (task-definition conflict)",46   "gate run 3 (controller resume) receipt out/state/gate_receipts/20261030T060507.995429Z.json: only forecasts_chronology fails. Not fixable honestly: for every h=1 row the target month (last(V)+1, per TASK.md) lies entirely on or before the month-end origin (6 origins: whole target month before origin; 4 origins: target month = origin month). No date inside the target month satisfies origin_date < target_date, so passing would require writing a date outside the target month, which would falsify forecasts.csv.",47   "gate run 4 (second controller resume) receipt out/state/gate_receipts/20261030T060535.228254Z.json: only forecasts_chronology fails; unchanged conflict with TASK.md target definition; not relabelled (would falsify target_date).",48   "gate run 5 (third controller resume) receipt out/state/gate_receipts/20261030T060548.765750Z.json: only forecasts_chronology fails; same TASK.md/gate conflict; target_date not falsified."49  ],50  "notes": "Blocked solely by the gate's forecasts_chronology rule, which contradicts TASK.md's definition target=last(V)+h with target_date=YYYY-MM-01 for h=1. Not relabelled to avoid misrepresenting targets."51 },52 "P5": {53  "status": "blocked",54  "checks": [55   "report.md written with all six sections and results-key citations",56   "claims.json written (6 claims, all referencing existing keys)",57   "gate run 2 receipt out/state/gate_receipts/20261030T060441.223227Z.json: only forecasts_chronology fails",58   "gate run 3 (controller resume) receipt out/state/gate_receipts/20261030T060507.995429Z.json: only forecasts_chronology fails. Not fixable honestly: for every h=1 row the target month (last(V)+1, per TASK.md) lies entirely on or before the month-end origin (6 origins: whole target month before origin; 4 origins: target month = origin month). No date inside the target month satisfies origin_date < target_date, so passing would require writing a date outside the target month, which would falsify forecasts.csv.",59   "gate run 4 (second controller resume) receipt out/state/gate_receipts/20261030T060535.228254Z.json: only forecasts_chronology fails; unchanged conflict with TASK.md target definition; not relabelled (would falsify target_date).",60   "gate run 5 (third controller resume) receipt out/state/gate_receipts/20261030T060548.765750Z.json: only forecasts_chronology fails; same TASK.md/gate conflict; target_date not falsified."61  ],62  "notes": "All deliverables complete; blocked only because the latest gate receipt fails forecasts_chronology (see P4)."63 }64}
Files · main
Menu

Theme