Add English text quality and completeness checks #110

Open
opened 2026-07-01 00:54:20 +00:00 by manlycucumber · 1 comment
manlycucumber commented 2026-07-01 00:54:20 +00:00 (Migrated from github.com)

Problem

Ingested English text can have gaps, encoding errors, or mismatched references. Automated checks are needed before text ships.

Acceptance criteria

  • Checks detect missing translations for referenced parts
  • Encoding and stray-markup checks run on every string
  • Reference-key integrity is validated against the resolver
  • Failures block the build

Notes

Keeps the pilot office complete and clean.

## Problem Ingested English text can have gaps, encoding errors, or mismatched references. Automated checks are needed before text ships. ## Acceptance criteria - [ ] Checks detect missing translations for referenced parts - [ ] Encoding and stray-markup checks run on every string - [ ] Reference-key integrity is validated against the resolver - [ ] Failures block the build ## Notes Keeps the pilot office complete and clean.

Status check against the tree (develop @ 8d5408c) — partially done, staying open.

Met:

  • Encoding checks on every string ✓ — NFC normalization at the ingest boundary plus non-empty-after-normalization (tools/generator/src/ingest.mjs:31,69)
  • Failures block the build ✓ — ingest and the provenance gate both throw from build.mjs
  • Missing translations flagged, not dropped ✓ — via #104's withheld-leaf path

Not yet met:

  • Stray-markup checks — no check exists; NFC normalization is not a markup scan
  • Reference-key integrity validated against the resolververify.mjs checks MANIFEST sha integrity and build.mjs:559 guards per-edition referential integrity, but neither validates text reference keys against what the resolver actually addresses

Because this is open, its epic #107 stays open too.

Status check against the tree (`develop` @ `8d5408c`) — **partially done, staying open.** Met: - Encoding checks on every string ✓ — NFC normalization at the ingest boundary plus non-empty-after-normalization (`tools/generator/src/ingest.mjs:31,69`) - Failures block the build ✓ — ingest and the provenance gate both throw from `build.mjs` - Missing translations flagged, not dropped ✓ — via #104's withheld-leaf path Not yet met: - **Stray-markup checks** — no check exists; NFC normalization is not a markup scan - **Reference-key integrity validated against the resolver** — `verify.mjs` checks MANIFEST sha integrity and `build.mjs:559` guards per-edition referential integrity, but neither validates text reference keys against what the resolver actually addresses Because this is open, its epic #107 stays open too.
Sign in to join this conversation.
No milestone
No project
No assignees
2 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Directorium/Core#110
No description provided.