@xsynaptic/word-count

A fast, dependency-free word counter with multilingual support. It handles whitespace-delimited scripts (Latin, Cyrillic, and so on) and scriptio-continua scripts where characters or syllables are themselves words: Chinese, Japanese, Korean, Thai, Lao, Burmese, Khmer, Javanese, and Vai.

Counting uses a Uint8Array bitmap that maps Unicode codepoints to word boundaries. Storing one bit per codepoint instead of one byte keeps the lookup table around 25.7KB.

Note: this package is ESM-only.

Install

npm install @xsynaptic/word-count

Use

import { countWords } from '@xsynaptic/word-count';

countWords('The quick brown fox'); // 4
countWords('中文 can appear anywhere'); // 5 (中, 文, can, appear, anywhere)
countWords('テスト'); // 3 (Japanese kana count per character)
countWords('สวัสดีครับ'); // counts Thai syllables, which have no spaces

API

countWords(input: string): number

Returns the number of words in input. For whitespace-delimited scripts a word is a run of non-boundary characters. For scriptio-continua scripts each in-range character counts as a word, since those scripts do not separate words with spaces.

countWordsBreakdown(input: string): { words, scriptChars, total }

Same count as countWords, split by how each unit was counted: words are whitespace-delimited (Latin, Korean, Cyrillic, …), scriptChars are per-character-script units (CJK, Japanese kana, Thai, Lao, Burmese, Khmer, Javanese, Vai), and total equals words + scriptChars. Use this when the two classes need different treatment, such as separate reading-speed rates.

Acknowledgements

The bitmap approach and the Unicode range tables are derived from alfaaz by Abdullah Atta, used under the MIT License. See LICENSE for the full notice.

License

MIT