built_at stringclasses 1
value | license stringclasses 1
value | license_holder stringclasses 1
value | counts dict | sqlite_path stringclasses 1
value |
|---|---|---|---|---|
2026-05-03T02:39:57.894634+00:00 | CC-BY-SA-2.0-KR | 국립국어원 (National Institute of Korean Language) | {
"stdict": 436574,
"urimalsam": 1204559,
"krdict": 53672,
"norm_articles": 490,
"examples": 80115,
"total_lex": 1694805
} | src/data/lexicon/lexicon.sqlite |
Korean Lexicon Ontology — Tier B (1.7M Headwords SQLite)
한국어 4대 언어자원(표준국어대사전·우리말샘·한국어기초사전·어문 규범)을 단일 SQLite로 통합한 170만 표제어 데이터셋입니다. Companion data for the Korean Lexicon Ontology Explorer.
개요
본 데이터셋은 국립국어원이 공개한 4개 한국어 언어자원을 통합·정규화하여 검색·조회에 최적화된 단일 SQLite 파일로 제공합니다.
| 자원 | 표제어/조항 수 | 비고 |
|---|---|---|
| 표준국어대사전 (stdict) | 436,574 | 표준어 |
| 우리말샘 (urimalsam) | 1,204,559 | 신어·방언·전문어 포함 |
| 한국어기초사전 (krdict) | 53,672 | 외국인 학습자용 |
| 한국어 어문 규범 (kornorms) | 110조항 + 380부록 + 80,115용례 | 한글 맞춤법·외래어·로마자 등 |
| 합계 | 약 1,694,805 표제어 + 80,115 용례 |
파일 구성
| 파일 | 크기 | 설명 |
|---|---|---|
lexicon.sqlite |
810 MB | 통합 표제어 + 용례 SQLite (WAL 모드) |
lexicon.meta.json |
< 1 KB | 버전·생성일·각 자원별 카운트 메타 |
스키마
-- 통합 표제어 테이블
CREATE TABLE lexicon_entries (
id INTEGER PRIMARY KEY,
source TEXT NOT NULL, -- 'stdict' | 'urimalsam' | 'krdict'
headword TEXT NOT NULL, -- 표제어
pos TEXT, -- 품사
definition TEXT, -- 뜻풀이
examples TEXT, -- JSON array (있을 시)
meta TEXT -- JSON object (자원별 확장 메타)
);
CREATE INDEX idx_headword ON lexicon_entries(headword);
CREATE INDEX idx_source ON lexicon_entries(source);
-- 외래어·로마자 용례
CREATE TABLE norm_examples (
id INTEGER PRIMARY KEY,
category TEXT NOT NULL, -- 'foreign' | 'romanization' | ...
original TEXT, -- 원어 또는 한국어
normalized TEXT, -- 정규화/표준 표기
source_ref TEXT -- 출처 조항 ID
);
상세 스키마는 lexicon.meta.json 참고.
사용법
직접 다운로드 (huggingface-cli)
hf download J-Seo/kocca-ontology-tierB \
--repo-type dataset \
--local-dir ./lexicon
Python (datasets 라이브러리)
from huggingface_hub import hf_hub_download
import sqlite3
path = hf_hub_download(
repo_id="J-Seo/kocca-ontology-tierB",
filename="lexicon.sqlite",
repo_type="dataset",
)
con = sqlite3.connect(path)
rows = con.execute(
"SELECT headword, definition FROM lexicon_entries WHERE headword LIKE ? LIMIT 5",
("사랑%",),
).fetchall()
웹앱 (Korean Lexicon Ontology Explorer)
본 데이터셋은 Next.js 기반 탐색 웹앱과 연동되어 있습니다. 전체 셋업 절차는 다음 저장소를 참고하세요.
→ https://github.com/J-Seo/kocca-ontology-explorer-public
라이선스
CC BY-SA 2.0 KR (https://creativecommons.org/licenses/by-sa/2.0/kr/)
원자료 라이선스를 그대로 승계합니다. 재이용 시 다음을 준수해야 합니다.
- 저작자 표시: 국립국어원 및 각 자원명·URL 명시
- 동일조건 변경허락: 재가공물도 동일한 CC BY-SA 라이선스로 공개
- 상업적 이용 가능 (위 두 조건 충족 시)
원자료 출처
- 표준국어대사전 — https://stdict.korean.go.kr
- 우리말샘 — https://opendict.korean.go.kr
- 한국어기초사전 — https://krdict.korean.go.kr
- 한국어 어문 규범 — https://kornorms.korean.go.kr
가공 내역
- 4개 자원의 이질적 스키마를 단일
lexicon_entries테이블로 정규화 headword,source인덱스 추가로 검색 성능 향상- 외래어/로마자 용례 80,115건을 별도 테이블로 추출
- 어문 규범 110조항 + 부록 변경 380건을 구조화 (별도 JSON, 본 SQLite에는 미포함 — repo 참고)
인용
@dataset{korean_lexicon_ontology_tierB_2026,
title = {Korean Lexicon Ontology — Tier B (1.7M Headwords SQLite)},
author = {Seo, Jaehyung},
year = {2026},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/J-Seo/kocca-ontology-tierB},
note = {Derived from open Korean lexical resources by NIKL.
Licensed under CC BY-SA 2.0 KR.}
}
면책
비공식 가공물입니다. 국립국어원의 공식 입장이나 데이터 정합성을 보증하지 않습니다. 정확한 표준 데이터는 위 출처의 원자료를 참고하십시오.
- Downloads last month
- 19