koroman 1.0.16

Korean Romanizer with pronunciation rules based on 국립국어원 표기법

KOROMAN - Korean Romanizer

KOROMAN is a multilingual Romanizer for Korean text, based on the Revised Romanization system (국립국어원 표기법) with additional pronunciation rules. It converts Hangul syllables into Romanized Latin script across multiple languages: JavaScript, Python, and Java.

🌐 Live Demo

• 한국어 버전
• English version

---

📦 Features

• Supports Revised Romanization of Korean
• Applies key Korean phonological rules:
  • Liaison (연음화)
  • Nasal assimilation (비음화)
  • Lateralization (유음화)
  • Aspiration / consonant cluster simplification — refined in 1.0.15
• Casing options (lower, upper, capitalized) — accepts full names, short aliases, or numeric codes (1.0.14+)
• Optional hyphen at ambiguous syllable boundaries (use_hyphen, 1.0.15+) — e.g. 중앙 → jung-ang, 반구대 → ban-gudae, 해운대 → hae-undae
• User dictionary (custom_dictionary, 1.0.15+) — per-call or persistent module-level store; match Korean words and emit user-provided Romanization with the highest priority (casing/phonological rules are not applied to dictionary values)
• Fully tested in each language

---

🆕 1.0.16 changes

• 29항 n-insertion (-잎 trigger), cluster liaison, 밟-, ㄷ+ㅎ+ㅣ→치
• Persistent dictionary starts empty — see CHANGELOG.md

1.0.15 changes (heads-up)

• Aspiration policy changed. Receiver-final ㄱ/ㄷ/ㅂ + initial ㅎ is now romanized as a single fortis (k/t/p/ch) consistently. The previous noun-only exception (e.g. 묵호 → Mukho, 집현전 → Jiphyeonjeon) required morphological analysis we cannot perform reliably, so it has been dropped:
  • 묵호 → muko, 집현전 → jipyeonjeon, 잡혀 → japyeo, 놓다 → nota
• Consonant cluster simplification (자음군 단순화) fixed: 흙→heuk, 닭→dak, 삶→sam, 읊→eup, 값→gap.
• ᆶ + ᄋ liaison preserved: 잃어→ireo (previously ileo).
• Final ㄷ / ㅎ map to t in standard mode (previously d / h). The legacy mapping is kept for use_pronunciation_rules=False to preserve 1.0.14 behaviour.

---

🚀 Getting Started

Python

pip install koroman

from koroman import romanize

# Basic usage
romanize("한글")  # → "hangeul"

# With pronunciation rules disabled
romanize("해돋이", use_pronunciation_rules=False)  # → "haedodi"

# With pronunciation rules enabled (default)
romanize("해돋이")  # → "haedoji"

# Casing options (full names)
romanize("한글", casing_option="uppercase")  # → "HANGEUL"
romanize("안녕 한글", casing_option="capitalize-word")  # → "Annyeong Hangeul"
romanize("안녕\n한글 로마자 변환", casing_option="capitalize-line")  # → "Annyeong\nHangeul romaja byeonhwan"

# 1.0.14+ : short aliases / numeric codes are also accepted
romanize("한글", casing_option="u")    # → "HANGEUL"
romanize("한글", casing_option="uc")   # → "HANGEUL"
romanize("한글", casing_option=1)      # → "HANGEUL"
romanize("안녕 한글", casing_option="cw")   # → "Annyeong Hangeul"
romanize("안녕\n한글 로마자 변환", casing_option=2)  # → "Annyeong\nHangeul romaja byeonhwan"

# Combining options
romanize("해돋이", use_pronunciation_rules=False, casing_option="uc")  # → "HAEDODI"
romanize("해돋이", use_pronunciation_rules=False, casing_option=1)     # → "HAEDODI"

# 1.0.15+ : use_hyphen — insert '-' at ambiguous syllable boundaries
romanize("중앙", use_hyphen=True)    # → "jung-ang"
romanize("반구대", use_hyphen=True)  # → "ban-gudae"
romanize("해운대", use_hyphen=True)  # → "hae-undae"

# 1.0.15+ : custom_dictionary — per-call user dictionary (highest priority, casing-safe)
romanize("나는 김철수입니다",
         custom_dictionary={"김철수": "Kim Chul-soo"})  # → "naneun Kim Chul-sooimnida"

# 1.0.15+ : persistent dictionary (module-level store)
from koroman import (
    set_custom_dictionary, add_custom_dictionary,
    remove_custom_dictionary_entry, clear_custom_dictionary, get_custom_dictionary,
)
set_custom_dictionary({"김철수": "Kim Chul-soo", "서울": "Seoul"})
add_custom_dictionary("부산", "Busan")
romanize("서울에 사는 김철수")                         # → "Seoule saneun Kim Chul-soo"
romanize("부산", use_custom_dictionary=False)         # → "busan"
clear_custom_dictionary()

casing_option aliases (1.0.14+)

Canonical	Aliases	Numeric
lowercase	lower, l, lc	0
uppercase	upper, u, uc	1
capitalize-line	cap-line, cline, cl	2
capitalize-word	cap-word, cword, cw	3

Case-insensitive. Unknown / None falls back to lowercase.

Persistent custom_dictionary API (1.0.15+)

API	Description
set_custom_dictionary(dict)	Replace the entire store. None/empty dict clears it.
add_custom_dictionary(dict) / add_custom_dictionary(key, value)	Merge/insert entries.
remove_custom_dictionary_entry(key)	Remove a single key (returns True on success).
clear_custom_dictionary()	Empty the store.
get_custom_dictionary()	Shallow snapshot of the current store.

Combine with the per-call custom_dictionary option to override the store on a single call (same-key entries in the option win). Set use_custom_dictionary=False on a call to ignore the persistent store entirely.

---

📜 LICENSE

MIT License

2025 ⓒ Donghe Youn (Daissue)