-
released this
2026-06-30 05:10:45 +00:00 | 0 commits to master since this releaseDrops 11 obscene nouns (mat, on the хуй/еб·ёб/пизд/бляд roots) from the Russian Scrabble and Эрудит dictionaries.
Removed:
ахуй, блядь, выблядок, ебальник, еблан, охуение, пизда, пиздец, пиздобол, хуй, хуйня.ru_scrabble: 83027 → 83016 wordsru_erudit: 82985 → 82974 wordsen_sowpods: unchanged
Coarse-but-not-mat (
мудак,мудила) and legitimate homographs (очко,сука,моча,дрочёна) are kept. Vetoed as policy viamanual_reject.txt. See #7.Downloads
-
released this
2026-06-22 12:37:53 +00:00 | 2 commits to master since this releaseDrop abbreviations from the Russian noun list (108 words: вуз, нэп, гост, жэк, спид, огвз, обком, наркомат, …). Lexicalised words rescued via manual_confirm (колхоз, совхоз, рация, спецназ, …, под). en_sowpods unchanged. See PR #6.
Downloads
-
released this
2026-06-20 16:18:09 +00:00 | 4 commits to master since this releaseConsistency pass on v1.2.0
Reviewing the kept low-confidence noun verdicts surfaced adjective-primary words admitted only by colloquial ellipsis, while their twins were already rejected — notably
большой(was kept) vsмаленький(dropped in v1.2.0). Now symmetric.9 more words rejected via
manual_reject.txt(71 total):большой, старшенький, горбатый, городской, одинокий, вольный, каторжный, припадочный, сподручный. Their lexicalized nouns are different words (горбун,горожанин,одиночка,каторжник) or the substantivized sense is marginal/устар.dawg words ru_scrabble.dawg83135 (−9) ru_erudit.dawg83093 (−9) en_sowpods.dawg267752 (unchanged) Asset
scrabble-dawg-v1.2.1.tar.gz— flat:en_sowpods.dawg,ru_scrabble.dawg,ru_erudit.dawg.Downloads
-
released this
2026-06-20 15:16:01 +00:00 | 6 commits to master since this releaseDrop substantivized-adjective false nouns
Adds the
manual_reject.txtmechanism:ru_stage2.pysubtracts hand-vetoed false nouns — substantivized adjectives a dictionary misreads as nouns (нёбный,акцизный,велярный, …) — from the Russian noun list, applied last, after every admission path (OC seed / libmorph / note /manual_confirm/ variant).62 words removed (43 pure adjectives + 19 marginal slang/archaic/jargon); genuine substantivized nouns (
больной,знакомый,учёный,участковый, …) retained.dawg words ru_scrabble.dawg83144 (−62) ru_erudit.dawg83102 (−62) en_sowpods.dawg267752 (unchanged) Asset
scrabble-dawg-v1.2.0.tar.gz— flat:en_sowpods.dawg,ru_scrabble.dawg,ru_erudit.dawg.Downloads
-
released this
2026-06-13 11:54:57 +00:00 | 8 commits to master since this releaseRussian dictionary fixes.
- Restore the lost primary sources (orfo_dict_2025.pdf, orfo_dict_2025.txt, all.txt), recovered byte-identically from scrabble-solver history.
- Drop OpenCorpora indeclinable abbreviations from the noun list: -179 words (ндс, ст, ср, кпд, чп, гибдд, …); lexicalised abbreviations (сельпо, под, ска, роно, врио, тв, фио, суперэвм) are kept via the orthographic note. See tools/README.md.
- Add
ru_stage2.py --trace WORD— a detailed per-signal audit of how a word does or does not reach the dictionary.
Artifact
scrabble-dawg-v1.1.0.tar.gz: en_sowpods (267752), ru_scrabble (83206), ru_erudit (83164).Downloads
-
released this
2026-06-09 13:59:51 +00:00 | 11 commits to master since this releaseDictionary sources tidied into
sources/<variant>/+tools/(PR #2). The DAWG content is unchanged from v1.0.0 — byte-identical en_sowpods/ru_scrabble/ru_erudit. This release just re-cuts the artifact from the tidied master for clean provenance; backends on v1.0.0 need no change. Asset: scrabble-dawg-v1.0.1.tar.gz (unpack into BACKEND_DICT_DIR).Downloads
-
released this
2026-06-05 18:42:21 +00:00 | 13 commits to master since this releaseDictionary DAWG set v1.0.0 (en_sowpods, ru_scrabble, ru_erudit). Built against scrabble-solver v1.0.0 + dafsa/alphabet v1.1.0.
Downloads