kana2p 0.1.2

Japanese Katakana-Phonemes Conversion

Japanese Katakana-Phonemes Conversion

Installation

pip install kana2p

Usage

from kana2p import katakana2phonemes

# Convert Katakana to Phonemes
katakana = "ハロー、ワールデョ！"
phonemes = katakana2phonemes(katakana)
print(phonemes)
# Output: ["h", "a", "r", "o", ":", "、", "w", "a", ":", "r", "u", "dy", "o", "！"]
print(katakana2phonemes(katakana, normalize=True))
# Output: ["h", "a", "r", "o", ":", "w", "a", ":", "r", "u", "dy", "o"]

from kana2p import phonemes2katakana

# Convert Phonemes to Katakana
phonemes = ["s", "i", ":", "kw", "a", ":", "s", "a", ":"]
katakana = phonemes2katakana(phonemes)
print(katakana)
# Output: "スィークァーサー"

Note

General Notes

• Only full-width Katakana are processed.
• If a character is not found in the mapping and normalize is False, the character is passed through unchanged.
• If normalize is True, characters not present in the mapping (e.g., non-Katakana characters) are excluded.

Special Characters

• The sokuon ッ is converted to the phoneme q.
• The prolonged sound mark ー is converted to the phoneme :.
• The syllabic nasal ン is converted to the phoneme N.
• ヲ is converted to ["o"] rather than ["w", "o"].
• ヶ is converted to ["k", "a"] rather than ["k", "e"].

Rare Consonants

• Some rare consonants like shw, dy, tsy, bw, ... are included (e.g. シュェ -> ["shw", "e"], デョ -> ["dy", "o"]).
• See src/kana2p/const.py for the full list of phoneme mappings.
• See also: https://ja.wikipedia.org/wiki/%E6%8D%A8%E3%81%A6%E4%BB%AE%E5%90%8D