Ko Speech Tools

0.1.0 verified Fri May 01 auth: no python

Korean speech and NLP tools for tokenization, romanization, and pronunciation analysis. Version 0.1.0, early development with monthly release cadence.

pip install ko-speech-tools

Common errors

error ModuleNotFoundError: No module named 'ko_speech_tools' ↓

cause Python environment does not have the package installed or the package name misspelled as 'ko-speech-tools' (with hyphen).

fix

Run 'pip install ko-speech-tools' and import with underscores: 'import ko_speech_tools'.

error AttributeError: module 'ko_speech_tools' has no attribute 'KoSpeechTokenizer' ↓

cause Older version of the package (pre-0.1.0) did not export KoSpeechTokenizer at top level, or the import path was wrong.

fix

Upgrade to 0.1.0: 'pip install --upgrade ko-speech-tools' and use 'from ko_speech_tools import KoSpeechTokenizer'.

error TypeError: romanize() missing 1 required positional argument: 'text' ↓

cause Called romanize without arguments or passed keyword argument incorrectly.

fix

Call romanize(text) or romanize(text=your_string).

Warnings

gotcha The library is in early alpha (0.1.0); expect breaking changes in minor versions. Pin to exact version in production. ↓

fix Use ko-speech-tools==0.1.0 in requirements.txt.

gotcha Romanization does not handle all ambiguous Korean syllables (e.g., homographs) correctly. Always verify output for sensitive applications. ↓

fix Use an additional disambiguation step or a more mature romanizer for critical tasks.

deprecated The function 'tokenize_syllables' is renamed to 'syllable_tokenize' in v0.2.0 (upcoming). ↓

fix Use 'KoSpeechTokenizer.syllable_tokenize()' instead of 'tokenize_syllables()'.

Imports

KoSpeechTokenizer

wrong

from ko_speech_tools.tokenizer import KoSpeechTokenizer

correct

from ko_speech_tools import KoSpeechTokenizer

Top-level import is preferred; submodule import may break in future versions.

romanize

wrong

from ko_speech_tools.romanization import romanize

correct

from ko_speech_tools import romanize

Function exported at package level, not from submodule.

pronounce

wrong

from ko_speech_tools.pronunciation import pronounce

correct

from ko_speech_tools import pronounce

Direct top-level import is intended.

Quickstart

Basic usage: tokenize, romanize, and get pronunciation of Korean text.

from ko_speech_tools import KoSpeechTokenizer, romanize, pronounce

text = "안녕하세요"
tokenizer = KoSpeechTokenizer()
tokens = tokenizer.tokenize(text)
print("Tokens:", tokens)

romanized = romanize(text)
print("Romanized:", romanized)

pron = pronounce(text)
print("Pronunciation:", pron)