Drop substantivized-adjective false nouns via manual_reject.txt #4

2026-06-20T15:11:03Z

developer commented

2026-06-20 15:11:03 +00:00

Problem

Stage 2 trusts a noun reading from OpenCorpora (and libmorph) above the РАН orthographic note. The abbreviation fix only carved out Abbr+Fixd, so substantivized-adjective false nouns a dictionary misreads still reached the dictionary — e.g. нёбный, акцизный, велярный. The РАН note classifies them as adjectives, but that check runs last and never fires for a word already seeded as a noun.

Fix

tools/ru_stage2.py: read sources/scrabble_ru/manual_reject.txt and subtract it last — after every admission path (OC seed / libmorph / note / manual_confirm / variant). Mirror of manual_confirm.txt; no-op when the file is absent (backward compatible).
sources/scrabble_ru/manual_reject.txt: 62 hand-reviewed words.
tools/README.md: document manual_reject.txt (Outputs table + a Stage-2 Manual overrides step).

The 62

43 pure adjectives (no nominal use): акцизный, нёбный, велярный, батальонный, брюшнотифозный, …
19 marginal (slang/archaic/jargon, vetoed): чумной, уголовный, фланговый, …
Genuine substantivized nouns stay (217 reviewed kept): больной, знакомый, учёный, участковый, рулевой, гласный, …

Method: 25 confirmed straight from the РАН note (-ого / в знач. сущ.); the rest by lexical review.

Effect

scrabble.txt: 83206 → 83144 (−62)
erudit.txt: re-folded, −62 (incl. небный)
DAWGs are gitignored; CI rebuilds them as the validation gate. Local make dawg-ru dawg-erudit succeeds (83144 / 83102 words).

Verified: нёбный and the reject sample are gone; увеличение, управление, знакомый, учёный, суперэвм retained.

## Problem Stage 2 trusts a noun reading from OpenCorpora (and libmorph) above the РАН orthographic note. The abbreviation fix only carved out `Abbr+Fixd`, so **substantivized-adjective false nouns** a dictionary misreads still reached the dictionary — e.g. `нёбный`, `акцизный`, `велярный`. The РАН note classifies them as adjectives, but that check runs last and never fires for a word already seeded as a noun. ## Fix - `tools/ru_stage2.py`: read `sources/scrabble_ru/manual_reject.txt` and subtract it **last** — after every admission path (OC seed / libmorph / note / `manual_confirm` / variant). Mirror of `manual_confirm.txt`; no-op when the file is absent (backward compatible). - `sources/scrabble_ru/manual_reject.txt`: 62 hand-reviewed words. - `tools/README.md`: document `manual_reject.txt` (Outputs table + a Stage-2 *Manual overrides* step). ## The 62 - **43 pure adjectives** (no nominal use): `акцизный, нёбный, велярный, батальонный, брюшнотифозный, …` - **19 marginal** (slang/archaic/jargon, vetoed): `чумной, уголовный, фланговый, …` - Genuine substantivized nouns stay (217 reviewed kept): `больной, знакомый, учёный, участковый, рулевой, гласный, …` Method: 25 confirmed straight from the РАН note (`-ого` / `в знач. сущ.`); the rest by lexical review. ## Effect - `scrabble.txt`: 83206 → 83144 (−62) - `erudit.txt`: re-folded, −62 (incl. `небный`) - DAWGs are gitignored; CI rebuilds them as the validation gate. Local `make dawg-ru dawg-erudit` succeeds (83144 / 83102 words). Verified: `нёбный` and the reject sample are gone; `увеличение`, `управление`, `знакомый`, `учёный`, `суперэвм` retained.

developer added 1 commit 2026-06-20 15:11:04 +00:00

Drop substantivized-adjective false nouns via manual_reject.txt

build / dawg (pull_request) Successful in 1m33s

Details

e17e945b41

OpenCorpora/libmorph hand a noun reading to Stage 2 above the РАН note, so
substantivized-adjective false nouns a dictionary misreads (нёбный, акцизный,
велярный, …) reached scrabble.txt with no veto path — the earlier abbreviation
filter only carved out Abbr+Fixd.

ru_stage2.py now subtracts sources/scrabble_ru/manual_reject.txt last, after
every admission path (OC seed / libmorph / note / manual_confirm / variant),
mirroring manual_confirm.txt. The list seeds 62 hand-reviewed words — 43 pure
adjectives plus 19 marginal (slang/archaic/jargon) substantivizations; genuine
substantivized nouns (больной, знакомый, учёный, участковый, …) stay.

scrabble.txt -62 (83206->83144); erudit.txt re-folded -62. The DAWGs are
gitignored and rebuild from these lists in CI.

owner approved these changes 2026-06-20 15:12:14 +00:00

developer merged commit c901fee994 into master

2026-06-20 15:14:02 +00:00

developer deleted branch feature/manual-reject-list

2026-06-20 15:14:02 +00:00

Sign in to join this conversation.

2 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: developer/scrabble-dictionary#4