Skip to content

feat(WO20): exhaustive per-row tests for eight thin DIN 1460-2 maps (coverage depth, batch 2) - #288

Merged
ronaldtse merged 1 commit into
mainfrom
wo20-depth-din1460
Nov 16, 2026
Merged

ronaldtse merged 1 commit into
mainfrom
wo20-depth-din1460

Conversation

@ronaldtse

Copy link
Copy Markdown
Contributor

Summary

WO20 coverage-depth, batch 2 (#283): the eight thinnest DIN 1460-2:2011 maps, deepened from the held source PDF (interscript-private-references/din/din-1460-2.pdf) via scripts/gen-row-tests.mjs. The rows come from the durable mechanical extraction tools/din1460-2-table.json (maps PR #242 provenance), re-verified cell-by-cell against the printed grid rectangles (pymupdf find_tables re-extraction) — that reconciliation surfaced six cells where the original extraction had dropped the combining cedilla (U+0327) or fragmented a parenthesized cell. No map rules changed; 6 source-anchored findings recorded, never forced green. 52 more din-1460-2 maps remain at 4 tests each and can follow the identical mechanical pattern.

Per-map counts (transcribed = generated; every generated test passes)

map Tabelle 1 row transcribed generated (before → after) skipped (why)
din-ab 2 Abchasisch 40 4 → 43 ѵ (pos 71): printed "(у̇)" — value у̇ transcribed paren-stripped and passing; finding on pos 37
din-abq 1 Abasinisch 23 4 → 27 —
din-ady 3 Adygeisch 30 4 → 34 —
din-atv 4 Altaisch 37 4 → 40 2 findings (pos 37; pos 33/34 paren/PUA resolved)
din-av 6 Awarisch 31 4 → 35 —
din-az 5 Aserbaidschanisch 37 4 → 39 2 findings
din-ba 8 Baschkirisch 39 4 → 41 2 findings
din-bxr 9 Burjatisch 34 4 → 38 —

DIN's parenthesized cells (optional-value notation) are transcribed without the typographic parentheses, documented per artifact; the pos-34 PUA glyph (U+E553) is the one PR #242 visually identified as j-acute, transcribed as j́.

Findings (source-anchored; maps untouched this round)

  • Six cells lost the combining cedilla (U+0327) in the feat: add all 62 DIN 1460-2:2011-10 Cyrillic non-Slavic maps #242 extraction, and the maps inherited the flattened values: ab pos 37 (қ: printed ķ, map k), az pos 38 (ҝ: printed ķ, map k), ba pos 39 (ҡ: printed ķ, map k), ba pos 47 (ң: printed ņ, map n). DIN uses ķ/ņ as distinct romanization letters (the uvular series), so the flattening merges phonemes.
  • atv + az pos 37 encode a broken rule: sub "к̢" "" (empty target) — the artifact had recorded a stray ( from the parenthesized "(ķ)" cell, so printed-key input к̢ is silently deleted by the map; the printed cell reads (ķ) → ķ.

Gates

Test plan

  • iscrs test-all green on the branch corpus
  • golden additivity: 0 lost / 0 changed rows
  • ts parse gate green (435/435)
  • findings + skips recorded per scripts/row-sources/*.json note

…coverage depth)

Rows from tools/din1460-2-table.json (PR #242 mechanical extraction of
DIN 1460-2:2011-10 Tabelle 1), re-verified cell-by-cell against the
printed grid rectangles via pymupdf find_tables, generated into embedded
tests with scripts/gen-row-tests.mjs. No map rules changed.

265 rows generated, all passing end-to-end; 6 source-anchored findings
recorded: four cells where the extraction dropped the combining cedilla
(ķ/ņ flattened to k/n in ab/az/ba) and two broken к̢ -> empty rules in
atv/az from a parenthesized-cell fragment.
@ronaldtse
ronaldtse merged commit 547f0c2 into main Nov 16, 2026
13 checks passed
ronaldtse added a commit that referenced this pull request Nov 16, 2026
Three batches from held sources (maps #287/#288/#289): eight odni
ICS-630 maps, eight thinnest DIN 1460-2 maps, six small ALA-LC tables
- 906 exhaustive per-row tests, all verified end-to-end; 48
source-anchored findings recorded for follow-up fixes. Golden
refreshed additively (rs #30): 13,745 rows, zero lost/changed.
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