Files

T

Ty d6aa79b593 docs(m3): regroup METADATA and remove the false fallback framing

METADATA was ordered as it was written, not by topic, and used Call 1 /
Call 2 without defining them.

- Add a short METADATA preamble stating the inference flow: Call 1 vision
  proposes taxonomy, the name and short description are deterministic, Call 2
  writes the long description.
- Regroup the rows: taxonomy, then Call 1, then deterministic copy, then
  Call 2, then cross-cutting.
- METADATA-70 said the deterministic name and short description were an
  inference-off fallback; they are unconditional. It now covers only the
  inference-off description.
- METADATA-81 folded into METADATA-80 (same rule, two strings).
- METADATA-45 now names the below-threshold outcome.
- METADATA-90 now states what the fact block drives.

153 requirements; check_requirements and check_coverage both pass.

2026-09-19 12:10:11 -07:00

4.5 KiB

Raw Blame History

Prompts contract

Status: SPEC · 2026-09-18. Owns prompt storage, prompt inputs, and the two inference calls.

Requirements impact

Requirement	Impact on this design
METADATA-5, METADATA-10, METADATA-15, METADATA-30	taxonomy choices and cardinality
METADATA-35, METADATA-40, METADATA-45, METADATA-50	Call 1 inputs, constraints, pre-fill, precedence
METADATA-55, METADATA-60, METADATA-65, METADATA-70	Call 2 output, inputs, dimension safety, inference-off path
METADATA-75, METADATA-80, METADATA-85, METADATA-90, METADATA-95	non-blocking, deterministic copy, pattern selection, fact block, overlay exclusion
METADATA-100, METADATA-105, METADATA-110, METADATA-115	feat mapping, prompt storage, button, failure
METADATA-66	generation logging
SETTINGS-30	brand voice and geo context
REVIEW-40, REVIEW-50	editability and field limits

1. Storage and revisions

Prompts live in the database, not in text files, with all revisions retained. Tables are in docs/contracts/DATA-MODEL.md §7. No prompt files ship with the application; the database is the only prompt source.
Each call reads the active revision fresh.
Editing a prompt creates a new revision; history is never overwritten.
Every generation is logged with its inputs, the request/response when a model was used, and the final stored content.

2. Call 1 — vision classification

When. Runs automatically after a calibrated draft uploads only when inference_enabled is true; also on demand or retry. With inference off it is skipped and the fields are filled by hand.

Inputs.

The clean processed photos at 1024 px long edge or less.
The synced taxonomy: species leaves, Edge Type, Figure, Grade, Thickness, Moisture terms, each with its Woo id. These are the only legal choices.
User metadata: SKU, length, thickness.

Constraints.

The model may only choose values from the supplied synced lists; it never invents categories or terms.
Free-text observations (character, inclusions, voids, checks) are mapped to existing synced feat-* tags only; Woo is authoritative for those tags.

Result fields. species_id + species_confidence; wood_category_ids; edge_type_term_id; figure_term_ids + figure_confidence; grade_term_id + grade_confidence; character, inclusions, voids, checks. Grade, edge, thickness, and moisture are singular.

Pre-fill. At or above the threshold (default 0.7) the field is pre-filled and remains editable; below it the field stays empty and is not shown as a low-confidence guess. Typed values always win. Call 1 is assist-only and never gates publish. A runtime failure is classified recoverable or unrecoverable for the owner.

3. Call 2 — content prose

When. User-triggered only, via the Generate text control; never automatic. The control appears only when content generation is enabled and the endpoint is reachable.

Inputs.

The clean processed photos.
The metadata confirmed or changed in review, including the primary and secondary wood categories and the parent-category path.
The literal dimension fact block, exact strings for species, figure term, length, thickness, average width, width min–max, and board feet; no length-by-width-by-thickness rectangle assumption.
Brand voice and geo context.

No provider thread id is used, and Call 1's raw prompt and result are not sent (METADATA-60).

Constraints.

The model writes prose only and never invents or alters a dimension; every number comes from the fact block (METADATA-65).
The dimension-overlay image is never sent (METADATA-95).

Output. The long description prose. The server assembles name and short_description (METADATA-55, METADATA-80). The primary wood category selects the specialized prompt (METADATA-85). The owner can edit the result (REVIEW-40).

4. Fallback and templates

With inference off, the server produces the long description so create never lacks prose (METADATA-70).
The deterministic title and short-description patterns live in docs/contracts/CONTENT-WOO.md (METADATA-80).
The primary wood category selects the specialized Call 2 prompt (METADATA-85).
All generated text stays within the field limits in docs/contracts/WOO-API.md §9 (REVIEW-50).

5. Prompt management

The active Call 1 and Call 2 revisions are editable; saving creates a new revision.
A revision identifies its content, author, and timestamp; inference_log records the revision used for each call.

4.5 KiB Raw Blame History Unescape Escape