nix-conf/user/bin/furigana

49 lines
1.7 KiB
Plaintext
Raw Normal View History

2025-07-13 15:42:38 -04:00
"""
ぁあぃいぅうぇえぉおかがきぎくぐけげこごさざしじすずせぜそぞただちぢっつづてでとどなにぬねのはばぱひびぴふぶぷへべぺほぼぽまみむめもゃやゅゆょよらりるれろわをんーゎゐゑゕゖゔゝゞ・「」。、
ァアィイゥウェエォオカガキギクグケゲコゴサザシジスズセゼソゾタダチヂッツヅテデトドナニヌネノハバパヒビピフブプヘベペホボポマミムメモャヤュユョヨラリルレロワヲンーヮヰヱヵヶヴヽヾ・「」。、
"""
2025-07-26 17:07:57 -04:00
# vi: filetype=python
2025-07-13 15:42:38 -04:00
import fugashi
import sys
2025-07-26 17:07:57 -04:00
import re
2025-07-13 15:42:38 -04:00
if len(sys.argv) != 2:
2025-07-26 17:07:57 -04:00
print("usage: furigana <word>")
2025-07-13 15:42:38 -04:00
exit(1)
2025-07-26 17:07:57 -04:00
H, K = __doc__.strip().split("\n")
trans = {ord(a): ord(b) for a, b in zip(K, H)}
2025-07-13 15:42:38 -04:00
2025-07-26 17:07:57 -04:00
def align_kana(s: list[str], k: str):
if not s:
return ""
if not k:
return "".join(s)
if s[-1][0] in __doc__:
return align_kana(s[:-1], k[:-len(s[-1])]) + s[-1]
if s[0][0] in __doc__:
return s[0] + align_kana(s[1:], k[len(s[0]):])
if len(s) == 1:
return f" {s[0]}[{k}]"
if any(x[0] in __doc__ for x in s):
i = next(i for i, x in enumerate(s) if x[0] in __doc__)
j = k.find(s[i].translate(trans))
if j != -1:
x = align_kana(s[:i], k[:j])
y = align_kana(s[i+1:], k[j+len(s[i]):])
return x + s[i] + y
return f" {''.join(s)}[{k}]"
output = []
2025-07-13 15:42:38 -04:00
for x in fugashi.Tagger().parseToNodeList(sys.argv[1]):
2025-07-26 17:07:57 -04:00
if x.feature.kana:
s = re.findall("[ぁ-ゔァ-ヴー]+|[^ぁ-ゔァ-ヴー]+", x.surface)
k = x.feature.kana.translate(trans)
output.append(align_kana(s, k))
2025-07-13 15:42:38 -04:00
else:
2025-07-26 17:07:57 -04:00
output.append(x.surface)
print("".join(output).strip())