Ty
49c22c9385
feat(greenfield): rebuild SlabUploader M1-M3 on clean foundation, all gates green
...
Clean re-implementation of the full product (not a port of the old M3 tree):
- P0 foundation: 13-table schema + per-call sqlite factory (thread-safe),
FastAPI app (health, JSON logging + secret redaction, CORS, startup guard).
- P1 M1 Capture: mask/sheetKeep, geometry/rect (min-area rect), measure,
png crop/inventory/overlay, EXIF decodeOrientedBitmap.
- P2 M2 Photos: stateful capture session (batch, pick-cal, apply mask/crop
to batch) + photos-ready lineup.
- P3 M3 Create: real backend (crypto, woo client, taxonomy sync, pricing,
inference Call1/Call2, review, publish, settings, slabs, photos) + real FE
review/publish/settings screens consuming the API (no hardcoded Woo ids).
- P4 gate exit: check_requirements + check_coverage both exit 0
(137/155 IDs, 0 gaps); backend 130/130, frontend 150/150, build green.
Cite: docs/REQUIREMENTS.md (155 rows), docs/MILESTONES.md M1-M3.
2026-09-25 06:57:29 -07:00
Ty
7287a55707
fix(inference): #86 tighten Call 1 prompt and closed lists
...
Ship locked expert Call 1 seed: order cats before edge, species=best
confidence, mill-count edge rules, Wild for cookies/burl, multi figure,
feature_tag_ids only (no free-text bags, no grade). LISTS keys match the
seed; feat- prefix stripped for the model. Gate and review UI apply
feature_tag_ids. Seed auto-activates on deploy when still seed-authored.
2026-09-20 22:54:36 -07:00
Ty
cec03c5bd3
fix(m3): AI-assist scope + confidence gate + grouped wood categories
...
- inference: _taxonomy_lists hands the model the synced species leaves + wood
leaves + case-insensitive attribute terms (was furniture roots / empty)
- inference: CALL1_SEED pins exact flat schema + *_confidence keys, drops grade
- inference: _gate maps field->confidence key correctly (was species_id_confidence),
reads nested shape as safety net, omits grade (METADATA-40)
- review.svelte: figure gets AI assist, grade dropped from AI (owner judgment)
- MultiSelect: grouped mode (parents non-selectable headers, selectable leaves)
- METADATA-40 updated: grade not suggested by Call 1
2026-09-20 21:47:36 -07:00
Ty
8a8a89ed3e
feat(inference): S7 Call 1/Call 2 + prompt revision REST + test-inference + Generate text UI
...
- backend/inference.py: prompt seed (rev 0) + append-only revision REST
(POST /prompts/{call}/revisions, /activate, GET; METADATA-105/106). Call 1 vision
classification (1024px processed photos + synced taxonomy, 0.7 confidence gate,
METADATA-35/40/45), Call 2 long description (user-triggered, METADATA-55/60/65),
inference_log per call (METADATA-66), inference-off server_description (METADATA-70),
test-inference probe (SETTINGS-45). Stdlib urllib, no new deps.
- frontend: create/inference/GenerateText.svelte (control visible only when content
enabled + endpoint reachable, METADATA-110; Call 2 + Call 1 buttons) + dev toggle.
- tests: 6 inference tests (seed+revision, call1 gate, call2 stores desc, off->409,
server desc, test-inference unconfigured).
Verified: pytest 54 passed; gates exit 0; npm build ok.
2026-09-19 17:04:01 -07:00
Ty