Ilia Denisov
e17e945b41
build / dawg (pull_request) Successful in 1m33s
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.
444 lines
19 KiB
Python
444 lines
19 KiB
Python
#!/usr/bin/env python3
|
||
"""Stage 2 — the "brain" of the Russian Scrabble word-list pipeline.
|
||
|
||
It reads the Stage-1 base word list (built once by ruwords so the heavy PDF is not
|
||
re-parsed) together with the grammatical notes and the singular/variant structure, runs
|
||
the whole noun-selection logic in memory, and writes a minimal result:
|
||
|
||
sources/scrabble_ru/scrabble.txt — the working dictionary (common nouns, nom. sing.)
|
||
sources/scrabble_ru/undefined.txt — the ambiguous tail, left for manual review
|
||
|
||
(sources/scrabble_ru/all.txt is the Stage-1 base.) Every other bucket — adjectives, verbs,
|
||
the merged note-nouns, singulars, variants — stays in memory. Pass --dump to also write
|
||
them; pass --trace WORD to ask how a single word did or did not reach the dictionary.
|
||
|
||
Note: all.txt is a plain word list, so the grammatical notes, "ед." singulars and "и"
|
||
variants are read from the pdftotext output (slov.txt) and the Stage-1 side files; the
|
||
expensive PDF parse itself runs only once.
|
||
|
||
Sources, most authoritative first: OpenCorpora (mawo-pymorphy3), libmorph (libmorph_check),
|
||
and the orthographic dictionary's own notes. See tools/README.md.
|
||
|
||
Run: ru-venv/bin/python tools/ru_stage2.py [--dump] [--trace WORD]
|
||
"""
|
||
import argparse
|
||
import os
|
||
import re
|
||
import subprocess
|
||
|
||
HERE = os.path.dirname(os.path.abspath(__file__))
|
||
# The curated Russian word lists live in sources/scrabble_ru/ (this tool sits in tools/);
|
||
# the uncommitted pipeline intermediates (orfo/all/debug) are regenerated alongside them.
|
||
OUT_DIR = os.path.join(HERE, "..", "sources", "scrabble_ru")
|
||
SLOV = os.path.join(OUT_DIR, "orfo_dict_2025.txt") # committed pdftotext output (source of truth)
|
||
WL_FROM, WL_TO = 452, 168808 # 1-based inclusive bounds of the column word-list section
|
||
OC_CACHE = "/tmp/oc_nouns.txt"
|
||
LIBMORPH_BIN = os.path.join(HERE, "libmorph_check")
|
||
|
||
ALPHABET = "абвгдеёжзийклмнопрстуфхцчшщъыьэюя"
|
||
ORDER = {c: i for i, c in enumerate(ALPHABET)}
|
||
PROPER = {"Name", "Surn", "Patr", "Geox", "Orgn", "Trad"}
|
||
ABBR_NOUN = {"Abbr", "Fixd"} # indeclinable abbreviation marker (НДС, КПД, сельпо, …)
|
||
FUNCTION_POS = {"CONJ", "PRCL"} # conjunction / particle: a function word, never a Scrabble noun
|
||
LIBMORPH_NOUN_CODES = set(range(7, 22)) | {24} # 7..21 plus 24 (pluralia tantum)
|
||
ADJ_END = {"ая", "яя", "ое", "ее", "ье", "ья", "ьи"}
|
||
VERB3 = ("ет", "ёт", "ит", "ют", "ут", "ает", "яет", "ует", "уют", "нет", "жет", "чет")
|
||
GENPL = ("ов", "ёв", "ев", "ей")
|
||
|
||
|
||
def key(w):
|
||
return [ORDER.get(c, 99) for c in w]
|
||
|
||
|
||
def destress(s):
|
||
return "".join(c for c in s if ord(c) not in (0x0300, 0x0301)).lower()
|
||
|
||
|
||
def cyr_ok(w):
|
||
return 2 <= len(w) <= 15 and all(("а" <= c <= "я") or c == "ё" for c in w)
|
||
|
||
|
||
def load(p):
|
||
return [l.strip() for l in open(p, encoding="utf-8") if l.strip()] if os.path.exists(p) else []
|
||
|
||
|
||
def write(path, words):
|
||
os.makedirs(os.path.dirname(path), exist_ok=True)
|
||
open(path, "w", encoding="utf-8").write("\n".join(sorted(set(words), key=key)) + "\n")
|
||
|
||
|
||
import mawo_pymorphy3 # noqa: E402
|
||
|
||
M = mawo_pymorphy3.MorphAnalyzer()
|
||
D = M._dawg_dict
|
||
|
||
|
||
def oc_noun_lemmas():
|
||
"""Every common-noun lemma (nom. sing. / pluralia tantum) in OpenCorpora's words.dawg,
|
||
excluding indeclinable abbreviations (Abbr+Fixd, e.g. НДС, КПД). Such words are admitted
|
||
only when the orthographic dictionary attests them as nouns (see oc_abbr_fixd_only and the
|
||
Stage-2 loop); abbreviations carried by OpenCorpora alone (ст, ср, кпд, …) are dropped."""
|
||
gp, pt = D.get_paradigm, D.parse_tag_string
|
||
para0, tagc = {}, {}
|
||
|
||
def g0(pid):
|
||
r = para0.get(pid)
|
||
if r is None:
|
||
suf0, tag0, pre0 = gp(pid, 0)
|
||
_, gr = pt(tag0)
|
||
r = (pre0, suf0, gr)
|
||
para0[pid] = r
|
||
return r
|
||
|
||
def gt(pid, idx):
|
||
k = (pid, idx)
|
||
r = tagc.get(k)
|
||
if r is None:
|
||
suf, tag, pre = gp(pid, idx)
|
||
pos, gr = pt(tag)
|
||
r = (suf, pre, pos, gr)
|
||
tagc[k] = r
|
||
return r
|
||
|
||
out = set()
|
||
for word, rec in D.words_dawg.iteritems():
|
||
pid, idx = rec
|
||
suf, pre, pos, gr = gt(pid, idx)
|
||
if pos != "NOUN":
|
||
continue
|
||
pre0, suf0, gr0 = g0(pid)
|
||
if (PROPER & gr) or (PROPER & gr0):
|
||
continue
|
||
if ABBR_NOUN <= gr0: # indeclinable abbreviation: not seeded; the loop decides via the note
|
||
continue
|
||
stem = word[len(pre):len(word) - len(suf)] if suf else word[len(pre):]
|
||
out.add(pre0 + stem + suf0)
|
||
return {w for w in out if cyr_ok(w)}
|
||
|
||
|
||
def oc_status(word):
|
||
"""(is_common_noun, in_dictionary) for word, from OpenCorpora only."""
|
||
parses = D.get_word_parses(word)
|
||
if not parses:
|
||
return False, False
|
||
gp, pt = D.get_paradigm, D.parse_tag_string
|
||
for pid, idx in parses:
|
||
suf, tag, pre = gp(pid, idx)
|
||
pos, gr = pt(tag)
|
||
if pos == "NOUN":
|
||
_, tag0, _ = gp(pid, 0)
|
||
_, gr0 = pt(tag0)
|
||
if not (PROPER & gr or PROPER & gr0):
|
||
return True, True
|
||
return False, True
|
||
|
||
|
||
def oc_abbr_fixd_only(word):
|
||
"""True when OpenCorpora's only common-noun reading of word is an indeclinable
|
||
abbreviation (Abbr+Fixd) — e.g. сельпо, ндс, ан. Such a word is not admitted as a noun on
|
||
OpenCorpora's say-so; the Stage-2 loop hands it to the orthographic note (and the function-
|
||
word check) instead, so lexicalised nouns (сельпо, под, ска) survive while bare letter-
|
||
abbreviations (ндс, кпд) and proper/служебные homographs (ан, зато) do not."""
|
||
parses = D.get_word_parses(word)
|
||
if not parses:
|
||
return False
|
||
gp, pt = D.get_paradigm, D.parse_tag_string
|
||
has_noun = has_plain = False
|
||
for pid, idx in parses:
|
||
suf, tag, pre = gp(pid, idx)
|
||
pos, gr = pt(tag)
|
||
if pos != "NOUN":
|
||
continue
|
||
_, tag0, _ = gp(pid, 0)
|
||
_, gr0 = pt(tag0)
|
||
if PROPER & gr or PROPER & gr0:
|
||
continue
|
||
has_noun = True
|
||
if not (ABBR_NOUN <= gr0):
|
||
has_plain = True
|
||
return has_noun and not has_plain
|
||
|
||
|
||
def oc_function_word(word):
|
||
"""True when OpenCorpora reads word as a conjunction or particle — a function word that
|
||
shares a spelling with an abbreviation/proper noun (ан «ан нет», зато). It is never a
|
||
Scrabble noun even if the orthographic note looks noun-like (ан → «самолёт Антонова»)."""
|
||
gp, pt = D.get_paradigm, D.parse_tag_string
|
||
for pid, idx in (D.get_word_parses(word) or []):
|
||
if pt(gp(pid, idx)[1])[0] in FUNCTION_POS:
|
||
return True
|
||
return False
|
||
|
||
|
||
def libmorph_analyze(words):
|
||
"""Map each word to (known, noun_lemma, codes) per libmorph; noun_lemma is None when it
|
||
is not a common noun there. Empty result if the helper binary is not built."""
|
||
words = list(words)
|
||
if not words or not os.path.exists(LIBMORPH_BIN):
|
||
return {}
|
||
proc = subprocess.run([LIBMORPH_BIN], input="\n".join(words), capture_output=True, text=True)
|
||
out = {}
|
||
for w, line in zip(words, proc.stdout.split("\n")):
|
||
fields = line.split("\t")
|
||
known = fields[:1] == ["1"]
|
||
codes, noun_lemmas = set(), []
|
||
for field in fields[1:]:
|
||
code, _, lex = field.partition(":")
|
||
if code.isdigit():
|
||
codes.add(int(code))
|
||
if int(code) in LIBMORPH_NOUN_CODES:
|
||
noun_lemmas.append(lex)
|
||
lemma = (w if w in noun_lemmas else noun_lemmas[0]) if noun_lemmas else None
|
||
out[w] = (known, lemma, codes)
|
||
return out
|
||
|
||
|
||
def build_notes():
|
||
"""Map each headword (destressed, lowercased) to its grammatical note."""
|
||
def is_hw(ch):
|
||
o = ord(ch)
|
||
return (0x0430 <= o <= 0x044F) or (0x0410 <= o <= 0x042F) or o in (0x0401, 0x0451, 0x0300, 0x0301)
|
||
|
||
hmap = {}
|
||
lines = open(SLOV, encoding="utf-8").read().split("\n")
|
||
for l in lines[WL_FROM - 1:WL_TO]:
|
||
s = l.lstrip()
|
||
e = 0
|
||
for ch in s:
|
||
if is_hw(ch):
|
||
e += 1
|
||
else:
|
||
break
|
||
hw = destress(s[:e])
|
||
if hw and hw not in hmap:
|
||
hmap[hw] = destress(s[e:]).strip()
|
||
return hmap
|
||
|
||
|
||
def classify(w, note):
|
||
"""Coarse part of speech of an out-of-dictionary word from its PDF note."""
|
||
if note is None:
|
||
return "amb"
|
||
n = re.sub(r"\([^)]*\)", "", note).strip() # drop domain/etymology parentheticals
|
||
if "кр. ф" in n or "кр.ф" in n or "прич." in n or "прил." in n:
|
||
return "adj"
|
||
ends = re.findall(r"-([а-яё]+)", n)
|
||
if any(e in ADJ_END for e in ends):
|
||
return "adj"
|
||
if "сов." in n or "несов." in n or "безл." in n:
|
||
return "verb"
|
||
if w.endswith("ся"): # reflexive: no Russian noun ends in -ся
|
||
return "verb"
|
||
if any(e.endswith(VERB3) for e in ends) and not any(m in n for m in ("ед.", "тв.", "род.", "м.", "ж.", "с.")):
|
||
return "verb"
|
||
if n == "" and w.endswith(("ый", "ий", "ой", "ая", "ое", "ые", "ие", "яя", "ее")):
|
||
return "adj"
|
||
if "нескл" in n:
|
||
return "noun" if any(g in n for g in ("м.", "ж.", "с.", "мн.")) else "amb"
|
||
if ends:
|
||
return "noun"
|
||
if n == "" and w.endswith(("ать", "ять", "еть", "ить", "оть", "уть", "ыть", "ти", "чь")):
|
||
return "verb"
|
||
return "amb"
|
||
|
||
|
||
def singular(w, note):
|
||
"""Nominative singular of a noun headword from the PDF note (authoritative) or, for a
|
||
plural headword without an explicit singular, the mawo lemma; pluralia tantum kept."""
|
||
n = note or ""
|
||
full = re.search(r"ед\.\s+([а-яё]+)", n)
|
||
if full:
|
||
return full.group(1)
|
||
suf = re.search(r"ед\.\s+-([а-яё]+)", n)
|
||
if suf:
|
||
s = suf.group(1)
|
||
i = w.rfind(s[0])
|
||
return w[:i] + s if i > 0 else w
|
||
ends = re.findall(r"-([а-яё]+)", re.sub(r"\([^)]*\)", "", n))
|
||
if ends and ends[0].endswith(GENPL):
|
||
for p in M.parse(w):
|
||
if str(p.tag.POS) == "NOUN":
|
||
return p.normal_form
|
||
return w
|
||
return w
|
||
|
||
|
||
def build():
|
||
"""Run the whole pipeline in memory. Returns the result sets plus a `fate` map giving
|
||
every word's outcome, so a word's path can be traced or the buckets dumped."""
|
||
oc = set(load(OC_CACHE)) or oc_noun_lemmas()
|
||
if not os.path.exists(OC_CACHE):
|
||
write(OC_CACHE, oc)
|
||
hmap = build_notes()
|
||
all_words = load(os.path.join(OUT_DIR, "all.txt"))
|
||
ed_nouns = set(load("/tmp/ru_singulars.txt"))
|
||
pairs = [tuple(p) for l in load("/tmp/ru_variants.txt") if len(p := l.split("\t")) == 2]
|
||
pdf = [w for w in all_words if cyr_ok(w)]
|
||
lm = libmorph_analyze(pdf)
|
||
|
||
def to_singular(w):
|
||
s = singular(w, hmap.get(w))
|
||
return s if cyr_ok(s) else w
|
||
|
||
fate = {}
|
||
scrabble = set(oc)
|
||
adj, verb, amb = [], [], []
|
||
for w in pdf:
|
||
if oc_abbr_fixd_only(w):
|
||
# Indeclinable abbreviation in OpenCorpora: do not trust OC's noun verdict. Drop it
|
||
# if OC also reads it as a function word (ан, зато — служебное/пропер); otherwise let
|
||
# the orthographic note decide, keeping lexicalised nouns (сельпо, под, ска) and
|
||
# dropping the rest (кпд, чп, гор, …).
|
||
if oc_function_word(w):
|
||
fate[w] = "отброшено: служебное слово (CONJ/PRCL в OpenCorpora)"
|
||
elif classify(w, hmap.get(w)) == "noun":
|
||
s = to_singular(w)
|
||
scrabble.add(s)
|
||
fate[w] = "scrabble: несклон. аббрев.-сущ. по помете орфословаря" + ("" if s == w else f" → {s}")
|
||
else:
|
||
fate[w] = "отброшено: аббревиатура без подтверждающей пометы орфословаря"
|
||
continue
|
||
oc_noun, oc_known = oc_status(w)
|
||
if oc_noun:
|
||
# A common-noun reading already put w in the seed (scrabble = set(oc)) — unless its
|
||
# lemma is a different word, i.e. w is an inflected form (ан = род. мн. от «ана»);
|
||
# such a form is not a headword and stays out.
|
||
fate[w] = ("scrabble: сущ. по OpenCorpora" if w in scrabble
|
||
else "отброшено: словоформа OpenCorpora (лемма — другое слово)")
|
||
continue
|
||
lm_known, lm_lemma, _ = lm.get(w, (False, None, frozenset()))
|
||
if lm_lemma is not None:
|
||
s = lm_lemma if cyr_ok(lm_lemma) else to_singular(w)
|
||
scrabble.add(s)
|
||
fate[w] = "scrabble: сущ. по libmorph" + ("" if s == w else f" → {s}")
|
||
continue
|
||
if oc_known or lm_known:
|
||
fate[w] = "отброшено: словарь знает как не-существительное"
|
||
continue
|
||
if w in ed_nouns:
|
||
scrabble.add(w)
|
||
fate[w] = "scrabble: ед.ч. по помете «ед.»"
|
||
continue
|
||
c = classify(w, hmap.get(w))
|
||
if c == "noun":
|
||
s = to_singular(w)
|
||
scrabble.add(s)
|
||
fate[w] = "scrabble: сущ. по помете орфословаря" + ("" if s == w else f" → {s}")
|
||
elif c == "adj":
|
||
adj.append(w)
|
||
fate[w] = "отброшено: прилагательное (помета орфословаря)"
|
||
elif c == "verb":
|
||
verb.append(w)
|
||
fate[w] = "отброшено: глагол (помета орфословаря)"
|
||
else:
|
||
amb.append(w)
|
||
fate[w] = "undefined: неоднозначное (нет в словарях, помета не определяет)"
|
||
|
||
# Manual confirmations: nouns the maintainer approved from the undefined tail.
|
||
for w in load(os.path.join(OUT_DIR, "manual_confirm.txt")):
|
||
if cyr_ok(w):
|
||
scrabble.add(w)
|
||
fate[w] = "scrabble: подтверждено вручную (manual_confirm.txt)"
|
||
|
||
# Variant rescue: a word joined by "и" to a confirmed noun is itself a noun.
|
||
pending = set(amb) - scrabble
|
||
changed = True
|
||
while changed:
|
||
changed = False
|
||
for a, b in pairs:
|
||
for x, y in ((a, b), (b, a)):
|
||
if x in scrabble and y in pending:
|
||
scrabble.add(y)
|
||
pending.discard(y)
|
||
fate[y] = f"scrabble: вариант от «{x}» (через «и»)"
|
||
changed = True
|
||
|
||
# Manual rejections: words the maintainer vetoed — substantivized-adjective false nouns
|
||
# a dictionary misreads as nouns (нёбный, акцизный, …). Subtracted last, so it overrides
|
||
# every admission path above (OC seed / libmorph / note / manual_confirm / variant).
|
||
reject = {w for w in load(os.path.join(OUT_DIR, "manual_reject.txt")) if cyr_ok(w)}
|
||
scrabble -= reject
|
||
for w in reject:
|
||
fate[w] = "отброшено: ручной запрет (manual_reject.txt)"
|
||
|
||
undefined = [w for w in amb if w not in scrabble and w not in reject]
|
||
return {
|
||
"oc": oc, "scrabble": scrabble, "undefined": undefined,
|
||
"adjectives": adj, "verbs": verb, "singulars": ed_nouns,
|
||
"fate": fate, "all": set(all_words), "hmap": hmap,
|
||
}
|
||
|
||
|
||
def trace(word, r):
|
||
"""A detailed, per-signal audit of how WORD did or did not reach the dictionary: its
|
||
Stage-1 (all.txt) membership, every OpenCorpora reading, libmorph, the orthographic note
|
||
with its classify verdict, and the final outcome. Used by `--trace WORD`."""
|
||
w = destress(word)
|
||
note = r["hmap"].get(w)
|
||
parses = D.get_word_parses(w) or []
|
||
gp, pt = D.get_paradigm, D.parse_tag_string
|
||
pos = sorted({pt(gp(pid, idx)[1])[0] for pid, idx in parses})
|
||
oc_noun, _ = oc_status(w)
|
||
|
||
out = [f"{word} → {'В СЛОВАРЕ' if w in r['scrabble'] else 'НЕ в словаре'}"]
|
||
out.append(f" Stage-1 / all.txt (строчный заголовок РАН): {'да' if w in r['all'] else 'нет'}")
|
||
if parses:
|
||
kind = ("только несклон. аббрев. (Abbr+Fixd)" if oc_abbr_fixd_only(w) else
|
||
"обычное сущ." if oc_noun else
|
||
"не существительное")
|
||
out.append(f" OpenCorpora: части речи {pos}; как сущ. — {kind}; "
|
||
f"служебное (CONJ/PRCL): {'да' if oc_function_word(w) else 'нет'}")
|
||
else:
|
||
out.append(" OpenCorpora: нет в словаре")
|
||
lm = libmorph_analyze([w]).get(w)
|
||
if lm:
|
||
known, lemma, _ = lm
|
||
out.append(" libmorph: " + (f"сущ. → {lemma}" if lemma else ("известно, не сущ." if known else "нет")))
|
||
else:
|
||
out.append(" libmorph: helper отсутствует — в анализе не участвует")
|
||
out.append(" Орфословарь РАН: " + (f"помета «{note}» → classify={classify(w, note)}"
|
||
if note is not None else "нет заголовка"))
|
||
reason = r["fate"].get(w)
|
||
if reason is None:
|
||
if w in r["scrabble"]:
|
||
reason = "лексикон OpenCorpora (обычное сущ.)"
|
||
elif oc_abbr_fixd_only(w) and w not in r["all"]:
|
||
reason = "исключено из посева: несклон. аббрев. (Abbr+Fixd), нет в all.txt"
|
||
elif not parses and w not in r["all"]:
|
||
reason = "нет ни в OpenCorpora, ни в all.txt"
|
||
else:
|
||
reason = "не прошло ни один путь отбора"
|
||
out.append(f" ИТОГ: {reason}")
|
||
return "\n".join(out)
|
||
|
||
|
||
def main():
|
||
ap = argparse.ArgumentParser(description="Stage 2 brain: build the noun dictionary, trace a word, or dump buckets.")
|
||
ap.add_argument("--dump", action="store_true", help="also write the in-memory buckets (adjectives, verbs, singulars, variants, fate)")
|
||
ap.add_argument("--trace", metavar="WORD", help="report how WORD did or did not reach the dictionary, then exit")
|
||
args = ap.parse_args()
|
||
|
||
r = build()
|
||
if args.trace:
|
||
print(trace(args.trace, r))
|
||
return
|
||
|
||
write(os.path.join(OUT_DIR, "scrabble.txt"), r["scrabble"])
|
||
print(f"=> sources/scrabble_ru/scrabble.txt {len(r['scrabble'])}")
|
||
print(f" undefined kept in memory: {len(set(r['undefined']))} (use --dump to write it)")
|
||
if args.dump:
|
||
write(os.path.join(OUT_DIR, "undefined.txt"), r["undefined"])
|
||
write(os.path.join(OUT_DIR, "adjectives.txt"), r["adjectives"])
|
||
write(os.path.join(OUT_DIR, "verbs.txt"), r["verbs"])
|
||
write(os.path.join(OUT_DIR, "singulars.txt"), r["singulars"])
|
||
fate_path = os.path.join(OUT_DIR, "fate.tsv")
|
||
os.makedirs(OUT_DIR, exist_ok=True)
|
||
with open(fate_path, "w", encoding="utf-8") as f:
|
||
for w in sorted(r["fate"], key=key):
|
||
f.write(f"{w}\t{r['fate'][w]}\n")
|
||
print(f" dumped: undefined.txt ({len(set(r['undefined']))}), adjectives.txt, verbs.txt, singulars.txt, fate.tsv")
|
||
|
||
|
||
if __name__ == "__main__":
|
||
main()
|