- End-of-session page for /boucle: the session's curves, the analyst's
labels against the detectors' events per category, a plain score table
and the benchmark (comparaison.py, D-066)
- Detectors moved to vues-deplacements/detectors/ (D-067)
- Detectors must detect everything; what the client sees is a separate
layer (D-068)
- Benchmark: each case keeps its labels from creation (etiquettes.json),
benchmark.py scores the detectors on every case, history in
vues-deplacements/benchmark/historique.csv; run by /boucle at each lot,
after each change and at the end, read by the upgrader and validator;
labels for C-001..C-005 from one analyst pass corrected by their files,
C-006..C-008 from lot 01 (D-069)
- Fixes from the releveur's notes: /boucle arguments and other sessions,
branch name in the spec, detector name in categories.py output, learned
case header, anonymiser.py (camera/, no empty CSVs, --etiquettes),
ensemble.py (clock-aligned steps, title), plages.py (same-sign merge),
labeling request written into /nouveau-cas
- Also carries the other session's uncommitted D-065 edits in the same
files; data/, .gitignore, pomper.py and stations.py are left out
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Each batch now leaves etiquettes.json next to the critique: the analyst's
own reading, written before the detectors' output, one entry per moment,
checked by critiquer-lot/etiquettes.py against the event map and the
station's sensors. Groundwork for a future learned model.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The agent now serves one precise case, reads the project's displacement
knowledge and the whole case base, starts from the detectors' events
without judging them, and no longer has conclusions imposed on it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Copy the station's displacement export and images into the case folder,
with coded sensor names; the station takes the case number instead of
S01, and graphique.py / vignettes.py drop --alias (D-050).
Rewrite C-001 as the expert's diagnosis of this event only: summary,
symptom, verdict, signature with commands, diagnosis, evidence. Drop
exclusions, expected output, unknowns and test history; matching no
longer uses exclusions (D-051). The corr_x jumps at the reference
switches are judged negligible.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add the anonymiser-cas skill: a growing registry gives stable codes
(S01, R1, C1), a stdlib script rewrites a case and renames its files,
and graphique.py / vignettes.py take --alias to plot with codes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Reports become a folder with rapport.md and its figures: a verdict, one
short section per anomaly (series plot with shaded window, full-resolution
image crops around the sensors), constant behaviour in one line each.
graphique.py gains --zone and --sortie; new vignettes.py crops images at
the sensors' raw x/y. Old reports removed; the blind rescan of PS des
Coudres finds the single anomaly flagged by the specialist (REF_G/REF_D).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Inclusive --fin at the displayed minute, UTC for naive time_utc, neutral examples,
5-significant-digit values, 'vide' vs '-' cells, no empty or gap-bridging plots,
timestamped report names, and an up-to-date CLAUDE.md next step.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New data will be added per station without mixing with the working set.
All scripts now default to data/VERMONT/; docs and D-029 follow.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Record decision D-027: a server-side LLM agent layer next to the
unchanged deterministic rules, aimed at unknown anomalies and rule
edge cases. The agent annotates and never rewrites the day color.
Home-made harness (Messages API + Tool Runner), staged start with a
manual Energy pilot on VERMONT history. D-003 is partly replaced.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>