niho

niho is a command-line tool for converting romanized Japanese text to Japanese characters.

$ echo _niho ha, Ro-ma ji_ wo nihongo_ni henkan_surutameno Tu-ru desu. | niho -d dics/sile.jsonl
nihoは、ローマ字を日本語に変換するためのツールです。

Installation

$ cargo install niho
A command-line tool for converting romanized Japanese text to Japanese characters

Usage: niho [OPTIONS]

Options:
      --version                Print version
  -h, --help                   Print help ('--help' for full help, '-h' for summary)
  -t, --tokenize               Output tokenized input as JSON instead of converting to Japanese
  -d, --dictionary-file <PATH> Path to dictionary file [env: NIHO_DICTIONARY_FILE]

Basic Syntax

niho converts romanized Japanese text to Japanese characters using the following syntax:

Input Text Types

• Regular text: Converted to hiragana (e.g., konnnichiwa → こんにちは)
• Capitalized text: Converted to katakana (e.g., Ko-hi- → コーヒー)
• Text ending with _: Converted to kanji using dictionary lookup (e.g., nihongo_ → 日本語)
• Text ending with multiple _: Select specific kanji from multiple candidates (e.g., ka__ → second kanji option for "ka")
• Text prefixed with _: Kept as raw text until whitespace (e.g., _Hello desu → Helloです)
• Whitespace: A whitespace immediately following a non-whitespace character is removed; all other whitespace is preserved as-is in output

Examples

# Convert hiragana
$ echo konnnichiwa | niho
こんにちは

# Convert katakana (use uppercase)
$ echo Ko-hi- | niho
コーヒー

# Convert kanji (use underscore suffix)
$ echo nihongo_ | niho
日本語

# Select from multiple kanji options
$ echo ka__ | niho
掛

# Mix different types
$ echo watashi ha Ko-hi- wo nomimasu | niho
わたしはコーヒーをのみます

# Keep raw text
$ echo '_English  _to  nihongo_' | niho
English to 日本語

Dictionary Format

The dictionary is stored in a JSONL (JSON Lines) format, where each line contains a JSON object representing a character or word mapping. The dictionary contains four types of entries:

• hiragana: Maps romanized text to hiragana characters
• katakana: Maps romanized text to katakana characters
• kanji: Maps hiragana text to kanji characters (with multiple options support)

Example entries:

{"type": "hiragana", "from": "ka", "to": "か"}
{"type": "katakana", "from": "ka", "to": "カ"}
{"type": "kanji", "from": "にほんご", "to": ["日本語"]}

The default dictionary can be found at dictionaries/default.jsonl.

Kanji and Unknown Word Handling

For kanji conversion (text ending with _), the tool performs the following process:

1. Romanized → Hiragana: First converts the romanized text to hiragana
2. Hiragana → Kanji: Then looks up the hiragana in the kanji dictionary
3. Multiple Options: If multiple kanji options exist, use additional _ to select (e.g., ka_ for first option, ka__ for second)
4. Unknown words: If no mapping is found in the kanji dictionary, the hiragana-converted text is output with trailing underscores matching the selection index (e.g., humei_ → ふめい_)