Korean Grapheme-to-Phoneme (g2pkk)

Common errors

• LookupError: Resource 'punkt' not found. Please use the NLTK Downloader to obtain the resource:

  cause The NLTK 'punkt' tokenizer data, required by g2pkk, has not been downloaded.
  fix Execute `import nltk; nltk.download('punkt')` in your Python environment. This typically only needs to be done once.

• ModuleNotFoundError: No module named 'g2pkk'

  cause The `g2pkk` package is not installed in your current Python environment.
  fix Install the package using pip: `pip install g2pkk`.

• AttributeError: 'G2pkk' object has no attribute 'some_method'

  cause Attempting to call a method that does not exist or has been removed/renamed in the `G2pkk` class.
  fix Refer to the official g2pkk documentation or GitHub README for the correct API usage. If you've updated the library, check for breaking changes in the release notes.

Warnings

• gotcha The `nltk` library, a dependency of g2pkk, requires the 'punkt' tokenizer resource to be downloaded separately. Failure to do so will result in a `LookupError`.
  Fix: Run `import nltk; nltk.download('punkt')` once before using g2pkk. It's good practice to include this check at the start of your application.
• gotcha g2pkk is in early development (version 0.1.x), which means its API might undergo non-backward-compatible changes in future minor or patch releases as it approaches a stable 1.0 version.
  Fix: Pin your `g2pkk` dependency to exact versions (e.g., `g2pkk==0.1.2`) in your `requirements.txt` to prevent unexpected breaking changes with new releases. Review release notes for any new versions carefully.
• gotcha Input text must be primarily Korean. While it handles numbers and some English, its core functionality and accuracy are designed for Korean grapheme-to-phoneme conversion.
  Fix: Ensure that the majority of the input text is Korean for optimal results. Mixed language input might produce less accurate or less natural-sounding romanizations for non-Korean parts.

Install

• pip install g2pkk Install g2pkk

Imports

• G2pkk

  from g2pkk import G2pkk

Quickstart

This quickstart demonstrates how to initialize the G2pkk converter and use it to transform Korean text into its romanized phonetic representation. It includes a crucial step to ensure the NLTK 'punkt' resource is downloaded, which is a common prerequisite for many NLP libraries.

import nltk
from g2pkk import G2pkk

# Important: Download NLTK 'punkt' resource if not already done
try:
    nltk.data.find('tokenizers/punkt')
except nltk.downloader.DownloadError:
    print("Downloading NLTK 'punkt' resource...")
    nltk.download('punkt')

# Initialize the G2P converter
g2p = G2pkk()

# Convert Korean text to its phoneme representation
text = "안녕하세요 g2p 입니다. 반갑습니다. 123."
result = g2p(text)
print(f"Original: {text}")
print(f"Phonetic: {result}")

# Example with specific romanization
text_roman = g2p("한국어")
print(f"'한국어' phonetic: {text_roman}")