every_holiday/VRCT
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[Update] Enhance kanji and okurigana processing in Transliterator class

Browse Source
This commit is contained in:
misyaguziya
2025-09-15 05:22:57 +09:00
parent 2efedfb572
commit 0d121ecaad
1 changed files with 91 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

105
src-python/models/transliterate/transliterate_japanese.py
View File

@@ -46,29 +46,51 @@ class Transliterator:
# 読みを分配
kana_left = reading_kana
for is_kan, part in blocks:
for i, (is_kan, part) in enumerate(blocks):
if is_kan:
# 仮ルール：残りの読みのうち、送り仮名分を除いた前半を充てる
# ex. "美しい"(うつくしい): 漢字=美, 残り送り仮名=しい
okuri_len = len(blocks[-1][1]) if not blocks[-1][0] else 0
kana_for_kan = kana_left[:-okuri_len] if okuri_len else kana_left
# 漢字ブロックの処理
if len(blocks) == 1:
# 単一ブロック（全て漢字）の場合
kana_for_kan = kana_left
elif i == len(blocks) - 1:
# 最後のブロック（漢字）の場合
kana_for_kan = kana_left
else:
# 中間の漢字ブロックの場合
# 後続の非漢字ブロックの文字数を計算
remaining_non_kanji = sum(len(p) for is_k, p in blocks[i+1:] if not is_k)
if remaining_non_kanji > 0 and len(kana_left) > remaining_non_kanji:
kana_for_kan = kana_left[:-remaining_non_kanji]
else:
# 漢字1文字あたり最低1文字の読みを割り当て
min_kana = len(part)
kana_for_kan = kana_left[:max(min_kana, len(kana_left) - remaining_non_kanji)]
# 空の読みを避ける
if not kana_for_kan and kana_left:
kana_for_kan = kana_left[:1]
result.append(
{
"orig": part,
"kana": kana_for_kan,
"hira": Transliterator.kata_to_hira(kana_for_kan),
"hepburn": katakana_to_hepburn(kana_for_kan, use_macron=True)
}
)
kana_left = kana_left[len(kana_for_kan):]
else:
# 送り仮名部分 → そのまま残りを割り当てる
kana_for_okuri = kana_left
# 非漢字部分（送り仮名など）
kana_for_okuri = kana_left[:len(part)]
result.append(
{
"orig": part,
"kana": kana_for_okuri,
"hira": Transliterator.kata_to_hira(kana_for_okuri),
"hepburn": katakana_to_hepburn(kana_for_okuri, use_macron=True)
}
)
kana_left = ""
kana_left = kana_left[len(kana_for_okuri):]
return result
@@ -78,14 +100,66 @@ class Transliterator:
results = []
for t in tokens:
surface = t.surface()
parts = self.split_kanji_okurigana(surface, t.reading_form())
for p in parts:
reading = t.reading_form()
# 単純に1文字ずつ処理
if len(surface) == 1:
# 1文字の場合はそのまま
results.append({
"orig": p["orig"],
"kana": p["kana"],
"hira": self.kata_to_hira(p["kana"]),
"hepburn": katakana_to_hepburn(p["kana"], use_macron=use_macron)
"orig": surface,
"kana": reading,
"hira": self.kata_to_hira(reading),
"hepburn": katakana_to_hepburn(reading, use_macron=use_macron)
})
else:
# 複数文字の場合は文字種別で分割
i = 0
reading_pos = 0
while i < len(surface):
char = surface[i]
if self.is_kanji(char):
# 漢字の場合、連続する漢字をまとめて処理
kanji_block = ""
while i < len(surface) and self.is_kanji(surface[i]):
kanji_block += surface[i]
i += 1
# 漢字ブロックの読みを推定
if i < len(surface):
# 後に文字がある場合、送り仮名を考慮
remaining_chars = len(surface) - i
kanji_reading = reading[reading_pos:-remaining_chars] if remaining_chars > 0 else reading[reading_pos:]
else:
# 最後の漢字ブロックの場合
kanji_reading = reading[reading_pos:]
results.append({
"orig": kanji_block,
"kana": kanji_reading,
"hira": self.kata_to_hira(kanji_reading),
"hepburn": katakana_to_hepburn(kanji_reading, use_macron=use_macron)
})
reading_pos += len(kanji_reading)
else:
# 非漢字の場合
non_kanji_block = ""
while i < len(surface) and not self.is_kanji(surface[i]):
non_kanji_block += surface[i]
i += 1
# 非漢字部分の読み（通常は文字数分）
non_kanji_reading = reading[reading_pos:reading_pos + len(non_kanji_block)]
results.append({
"orig": non_kanji_block,
"kana": non_kanji_reading,
"hira": self.kata_to_hira(non_kanji_reading),
"hepburn": katakana_to_hepburn(non_kanji_reading, use_macron=use_macron)
})
reading_pos += len(non_kanji_reading)
return results
# --- テスト ---
@@ -103,6 +177,9 @@ if __name__ == "__main__":
"ラーメンを食べる",
"チョコレートが好き",
"SessionIDを取得する",
"取り敢えず検索してみる",
"見知らぬ土地で冒険する",
"彼は優れたエンジニアです",
]
transliterator = Transliterator()