PE run-job #376: client-review workbook inspection (run 2026-10-20)¶
Provenance (MMDEP-PM check-in, scheduled run, 20 Oct 2026 ~11:05 IST):
- Issue: https://github.com/moneymarche-dev/portfolio-engine/issues/376
- Repository
moneymarche-dev/portfolio-engine, branchfeat/issue-309-client-review-excel, checked-out head700f65a6457f0a436a9a1c6ed57b67172cb162e9(PR #310, still open). - Exact-head CI: workflow run 39095669184 (#559)
completed / successon that head, re-verified via GitHub REST before the run. - Heartbeat:
control/HEARTBEAT.mdlast collector run 2026-10-20T05:16Z, active until 2026-10-20T07:46Z: ACTIVE at execution (~05:35Z). - Input:
docs/relay/data/pe-client-review-workbook-2026-10-17/PHASE1_25FUND_CLIENT_PORTFOLIO_REVIEW_2026-10-17.xlsx, copied unchanged; SHA-256 verifiedebe1caa58791e0677a2ab685b1e1c8bae0377025e7673abe1731691420bba322. Relay original untouched; not regenerated; no live API. - Runner: MMDEP-PM cloud sandbox (Linux, Python,
python -I, stdin from /dev/null), not the laptop. The inspector is read-only and makes no network calls, so the PowerShell form in the issue was run as its Linux equivalent:python tools/jobs/inspect_client_review_workbook.py <workbook> --out <out> - Exit code: 0; status PASS; 8 sheets; missing expected sheets none; formula/error tokens 0.
- Output hashes (SHA-256):
inspection_summary.json69a5fd3cab9c92fe88a31ab4d7491a96b79d63b7772e3c2672b42fc75fb1cf2dinspection_report.md932de177cd07f8d911b9d04fcac4d7c02120da9b7b9d9192c07b32b0217552d7sheets/*.csv: every non-empty cell of each of the 8 worksheets.
Observations for PE (not acceptance; PE reviews the exports):
- Review chart: 1 LineChart, 1 series, title
'Chart Data'!B1, values'Chart Data'!$B$2:$B$915, categories$A$2:$A$915(914 observations, matching "Observations: 914"). - Label search "starting value" / "ending value" found 0 hits because the workbook says "Starting portfolio value" / "Ending portfolio value"; the labels exist (Review, Performance). The search term list repeats "UNAPPROVED", so that block is printed twice.
- Benchmark Evidence rows are marked UNAPPROVED; Methodology says portfolio benchmark UNAVAILABLE. Consistent with frozen methodology.
- No client identifiers in the exports (model portfolio of public schemes).