VERIFY PASS pe corpus k-NN exemple etichetate (seed real 17181 Haiku, comis
in 756f777): suita 1392 passed, 1 deselected (live); smoke init_db seeder
(17181/NUL=2200/idempotent); toate codurile in nomenclator.
US-007 (cerere user la CLOSE) — badge sursa pe sugestia fuzzy din editor:
- _mapari.html: chip confirmat (GOLD) / similar (SILVER+k-NN) / non-operatie (NUL)
- base.html: .sugg-sursa--{confirmat,similar,nul} pe tokeni de tema (color-mix)
- routes.py: cheia `nul` adaugata in surse_sugestie default (finding cross-file)
- tests/test_web_badge_sursa.py: gold/silver/nul/fara-sursa (4 teste)
- E2E render live verificat in serverul real (/_fragments/mapari)
CLOSE /code-review high (main..HEAD, 3 finder x 8 unghiuri) — runtime curat,
invariant #13 intact; 3 findings low/cosmetic REPARATE + lock-uite:
- shared_store.seed_suggestions: cod whitespace -> NULL (era ''), + test lock
- genereaza_seed.py: with open(...) in loc de open().read() (FD leak tool offline)
- embeddings.py: docstring-uri aliniate la [{cod, is_nul, similaritate}]
ROADMAP: 5.18 LIVRAT. PRD: raport VERIFY/CLOSE scris.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
5.15 (propagare design + dashboard editare) si 5.14 (mapare LLM distilata)
inchise dupa /code-review high. 8 buguri reparate TDD:
- HIGH modal nu se deschidea pe randul slim (base.html: trimitere-slim)
- HIGH /repune trunchia prestatii (declaratie incompleta la RAR) -> iterare
peste existing, codes pozitional
- HIGH embeddings incarca model ~230MB degeaba pe corpus gol -> poarta has_corpus()
- HIGH picker chips gol pe re-render eroare -> conn/account_id pe toate ramurile
- MED obs re-derivat dupa stergere explicita -> _merge_override pastreaza obs=''
- MED mapare salvata fara denumire poluă GOLD -> _record_gold_validation guard
- MED typo nome_prestatie -> nume_prestatie in select /repune
- MED bucketare timp +3h gresita iarna -> SQLite localtime + TZ=Europe/Bucharest
Embeddings WIRE-uit functional (PRD #15, decizie user): ensure_embeddings_corpus
construieste corpus din nomenclator, gated pe AUTOPASS_EMBEDDINGS_ENABLED (default
off). Marime model corectata ~50MB->~230MB (estimare PRD gresita).
Cleanup: hoist load_* din bucla bulk-fix; import re la top.
Regresie: 1256 passed, 1 deselected (live), 0 failed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PRD 5.14 rescris cu pivotul arhitectural: LLM doar etichetator OFFLINE,
runtime = clasificator local fara API (fuzzy + embeddings), baza de
cunostinte GOLD partajata cross-account (validarea unui service ajuta
toate). Decizia 8 (corpus per-cont) SUPERSEDED.
Tooling nou OpenRouter (free, familia NVIDIA Nemotron): or_common.py
(client + corpus pe frecventa, cheie din .env) + or_modeltest.py
(comparatie modele, acord ensemble vs Groq). Masurat: super-120b +
nano-9b fiabile, 3/3 unanim pe 87% volum; ultra-550b aruncat.
Corpus real (4 CSV service, coloana NR=frecventa) + etichete Groq
bootstrap incluse ca date de masurare.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
