numerals-to-text-ua
v1.0.3
Published
A lightweight Ukrainian numeral parser that reliably extracts and converts number words from free-form text
Readme
numerals-to-text-ua
A zero-dependency TypeScript library for extracting and parsing Ukrainian number words from natural language text.
It understands:
- grammatical cases
- ordinal and cardinal forms
- big numbers above billions
- apostrophe variants
Designed for NLP, document parsing, legal texts, financial reports, and OCR pipelines.
Features
- Extracts number expressions directly from text
- Converts Ukrainian number words to integers
- Supports:
- declensions (чотирьохсот, тисячами)
- composite numbers (сто двадцять три)
- large numbers (мільйон, мільярд)
- apostrophe variants (п’ять, п'ять, п`ять)
- No runtime dependencies
- Deterministic and fast
API
extractNumbers(text: string): Match[]
Finds all number expressions inside a text.
Returns:
- start — start index
- end — end index
- word — matched numeral
Example:
import { extractNumbers, parseNumber } from 'numerals-to-text-ua'
const text = 'У мене є двадцять три яблука і чотириста груш.'
extractNumbers(text)
// [{ start: 9, end: 21, word: 'двадцять три' }, { start: 34, end: 43, word: 'чотириста' }]parseNumber(input: string): number
Converts numbers to integers
parseNumber('двадцять три') // 23
parseNumber('сто двадцять пʼять') // 125
parseNumber('чотирьмастами') // 400
parseNumber('два мільйони') // 2000000MIT License