Skip to content

F3b-complete — bundle the real NKJV verse-count table - #11

Merged
kbennett2000 merged 1 commit into
mainfrom
chore/f3b-complete-verse-table
Jun 10, 2026
Merged

F3b-complete — bundle the real NKJV verse-count table#11
kbennett2000 merged 1 commit into
mainfrom
chore/f3b-complete-verse-table

Conversation

@kbennett2000

Copy link
Copy Markdown
Owner

F3b-complete — generate & commit the real verse-count table

Completes the human-in-the-loop checklist from PR #10 (merged at its stop point before the table landed). The bundled asset app/src/main/assets/anchor/verse_counts.json was confirmed absent on main; this PR adds it so SpanResolver has real data and VerseCountsSchemaTest runs.

This PR opens READY (not draft) — it is complete by construction (no further human step). It also adds the new methodology rule that human-step PRs open as drafts going forward.

Generated

Run on Kris's LAN box against Concord (a dev-time dependency only — the app never calls Concord at runtime):

python3 tools/generate_verse_counts/generate_verse_counts.py \
  --base-url http://192.168.1.62:8000 \
  --translation NKJV \
  --concord-version v1.2.0 \
  --out app/src/main/assets/anchor/verse_counts.json

Connectivity smoke-test first: GET /v1/translations → HTTP 200, NKJV present.

Emitted meta:

{
  "book_count": 66,
  "concord_version": "v1.2.0",
  "generated_at": "2026-06-10T00:55:07Z",
  "source_translation": "NKJV"
}

Contiguity warnings: none — NKJV's versification is contiguous (it retains the traditional verses), so len(verses) == max(verse) for every chapter.

Sanity (reported)

  • meta.book_count == 66; 66 book keys, all matching BookMap.
  • books.GEN = [31, 25, 24, 26, 32, 22, ...] (50 chapters).
  • books.PSA = 150 entries (Ps 23 = 6).
  • books.REV = 22 entries.
  • No non-positive counts.

Provenance caveat

The table is generated from NKJV, a licensed translation living in Concord's data/private/. Bare verse counts are canonical structure (not text), so NKJV is a fine source — but regeneration requires Kris's private Concord deployment. Recorded in CLAUDE.md.

Tests — skipped → passed

  • VerseCountsSchemaTest: previously Assume-skipped; now passes against the real asset (asserts 66 keys == BookMap, every count a positive int, four meta fields, book_count==66, production VerseTable parses it).
  • Clean :app:assemblePaddleDebug + :app:testPaddleDebugUnitTestgreen. Full suite 1128 tests, 0 failures, 0 errors, 0 skipped (was 1 skipped).

Scope

2 files changed: app/src/main/assets/anchor/verse_counts.json (+1331) and CLAUDE.md (+4/−1). No frozen-core, res/, or UI changes.

🤖 Generated with Claude Code

Completes PR #10's human-in-the-loop checklist: generate and commit the real
canon-structural verse-count table the SpanResolver consumes.

Generated with tools/generate_verse_counts against Concord on the LAN
(--translation NKJV --concord-version v1.2.0). meta.book_count=66; 66 book keys
matching BookMap; GEN=[31,25,24,...]; PSA 150 chapters; REV 22; no contiguity
warnings (NKJV versification is contiguous); all counts positive.

VerseCountsSchemaTest flips skipped -> passed against the real asset; full unit
suite green (1128 tests, 0 skipped).

CLAUDE.md: add the NKJV/private-deployment provenance caveat to the F3b regen
note, and a methodology line — PRs at a human-step stop point open as GitHub
drafts (mechanically unmergeable until Kris marks ready).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@kbennett2000
kbennett2000 merged commit 51d7e09 into main Jun 10, 2026
2 of 3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant