kiwipiepy-model

What it is and what it does

kiwipiepy-model is a model data package that supplies pre-trained linguistic models required by the morphological analyzer. It is automatically installed as a dependency and contains the dictionary, part-of-speech tag definitions, and learned models needed to tokenize Korean text and assign grammatical tags to each morpheme.

The package itself is not directly imported or called by user code—it works behind the scenes to provide the data layer. When a morphological analyzer instance calls tokenize() on a Korean sentence, these models are loaded to break the text into morphemes and label them with Sejong Corpus-based part-of-speech tags. The models support normalization of irregular Korean consonants and can be extended with user-defined words and their inflected forms.

Use it for:

• Tokenizing Korean text into morphemes for downstream NLP tasks like named entity recognition or sentiment analysis.
• Extracting nouns, verbs, and other parts of speech from Korean documents for text mining or search indexing.
• Preprocessing Korean social media or user-generated text with normalization options to handle irregular consonants.
• Building Korean language search or filtering systems that need to recognize word boundaries and grammatical roles.
• Analyzing Korean sentences with sentence splitting and per-token metadata like position and part-of-speech tag.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides pre-trained morphological analysis models for Korean text processing, enabling tokenization and part-of-speech tagging when installed as a dependency.

Yes, with conditions. Install if you need Korean morphological analysis and are willing to handle compilation requirements on platforms without pre-built wheels. The package is actively maintained, has no known vulnerabilities, and is in the top 15000 on PyPI by downloads. The LGPL v3 copyleft license requires disclosure if you modify or redistribute the package itself, but is permissive for typical library use. High install friction is the main trade-off.

Install

kiwipiepy-model on PyPI

pip install kiwipiepy-model

uv add kiwipiepy-model

poetry add kiwipiepy-model

Installing kiwipiepy-model

Before you install

High install friction: requires compilation from source in environments without pre-built binaries, necessitating CMake 3.12 or later and a C++17-compatible compiler. Maintenance is active with recent commits, but binary distribution availability varies by platform.

License in practice

Licensed under LGPL v3 (copyleft), which requires that any modifications to this package be shared under the same license and that derivative works disclose their use of this library.

Quickstart

pip install kiwipiepy-model
from kiwipiepy import Kiwi
kiwi = Kiwi()
tokens = kiwi.tokenize("안녕하세요")

CMake 3.12 or later required for source compilation if pre-built wheels unavailable; Windows Vista or later, Linux, or macOS 10.12 or later.

Verify before relying

• Whether pre-built wheels are available for all major platforms and Python versions.
• Performance characteristics and memory footprint on large texts.
• Exact Python version compatibility beyond what source build documentation specifies.

Package facts

Evidence: kiwipiepy_model-0.23.0.tar.gz

Keywords: Korean, morphological, analysis

Tags

More Libraries packages