The results folder and the report¶
<dir>/campaign.json the spec, with lattice_sha256 (deck; for a .lgproj the deck it points to) and spec_sha256
circuits.json the circuit map the campaign used
pins.json the frozen-phase pins per engine (envelope, mp)
lattice/<deck copy> provenance only — runs use the pinned original
plan/manifest.json the static plan; manifest_comp / manifest_mp / manifest_fos the waves planned at run time
legs/faults/runs/<id>/{results.h5,status.json}
legs/faults/compensation/<case>_<strategy>.json
legs/faults/{faults.csv, compensation.csv, criticality_map.csv, unrecoverable.json}
legs/imperfections/{draws/seed_NNNN.json, runs/…, seeds.csv, faults_on_seeds.csv, robustness.csv}
legs/foil/{runs/…, foil.csv}
legs/availability/{blocks.csv, fault_classes.json, availability.json, availability.csv, trip_histogram.csv, sensitivity.csv}
figures/*.png
summary.json
report.html
import.json written by a job import
Item ids read A_<case>_<bracket>_<env|mp> (fault leg), A_comp_<case>_<strategy>,
C_draw_NNNN, C_seed_NNNN_<model>, C_fos_<case>_seed_NNNN_<before|after>_<model>,
D_<scenario>_<model>, B_availability.
status.json of a scan-pool item: status (ok / failed), finished,
elapsed, error, the item (id, leg, case, class, bracket, label,
elements, model, seed), the overrides applied, metrics (the worker row:
sizes, emittances, transmission, exit energy, loss power) and extras (the
exit RF clock, the treaty-point energy, phase, sizes and Twiss, loss power
per section, the foil spot, power density and stripping populations). A
.part results file is an orphan of an interrupted run and is swept on
resume.
summary.json carries the provenance (deck path and digest, spec digest,
preset, when), the item counts per leg, the rule, the fault leg's verdicts
(cases, critical, by class, recovered-by counts, the unrecoverable list, the
top ten), the compensation tally, the imperfection statistics and robustness
rows, the foil rows, and the availability per variant.
report.html is one file with the figures embedded as data URIs and only the
HTML subset a Qt text view renders (no scripts, no CSS grid): provenance, the
rule, every leg's verdicts and first rows with the CSV named for the rest.
summarize rebuilds all of it from the per-item facts at any time.