translate-typeform

Translates Typeform forms to Facebook Messenger format.

Number Fields with Locale

To support international number formats in number fields, add locale to the field description in YAML:

locale: de-DE

Supported locales:

• en-US (default) - US format: 1,234.56
• de-DE - European format: 1.234,56
• ar-SA - Arabic Saudi: ١٬٢٣٤٫٥٦
• ar-MA - Arabic Morocco: 1 234,56

Example Typeform Field Description

For a number field expecting European format input:

locale: de-DE

The validator will then accept 1.234,56 as valid input and correctly parse it as 1234.56.

Number Parsing API

The parseNumber(str, locale) function parses numeric strings with locale-aware formatting.

Usage

const { parseNumber } = require('@vlab-research/translate-typeform')

parseNumber('1,234.56')           // 1234.56 (US format, default)
parseNumber('1.234,56', 'de-DE')  // 1234.56 (European format)
parseNumber('١٢٣', 'ar-SA')       // 123 (Arabic-Indic numerals)

Supported Formats

| Locale | Thousands | Decimal | Example | |--------|-----------|---------|---------| | en-US (default) | , | . | 1,234.56 | | de-DE | . | , | 1.234,56 | | ar-SA | ٬ | ٫ | ١٬٢٣٤٫٥٦ | | ar-MA | (space) | , | 1 234,56 |

Unicode Numeral Support

Automatically normalizes digits from 30+ numeral systems:

• Arabic-Indic: ١٢٣ -> 123
• Devanagari: १२३ -> 123
• Persian: ۱۲۳ -> 123
• Bengali, Tamil, Thai, and many more

Cross-Locale Behavior

Important: The locale determines how separators are interpreted. Mismatched locale and input format will produce incorrect results.

Works: ar-SA locale with US-style ASCII input

parseNumber('100.50', 'ar-SA')    // 100.5 (correct)
parseNumber('1,234.56', 'ar-SA')  // 1234.56 (correct)

This works because ar-SA separators (٬ thousands, ٫ decimal) normalize to the same ASCII characters as US format (, and .).

Works: Default locale with Arabic script input

parseNumber('١٠٠.٥٠')        // 100.5 (correct)
parseNumber('١٬٢٣٤٫٥٦')      // 1234.56 (correct)

This works because Arabic digits and separators normalize to US-style ASCII.

Broken: ar-MA locale with US-style input

parseNumber('100.50', 'ar-MA')    // 10050 (wrong!)
parseNumber('1,234.56', 'ar-MA')  // 1.23456 (wrong!)

ar-MA uses European-style formatting (comma=decimal, space=thousands). When given US-style input (dot=decimal, comma=thousands), the parser misinterprets the separators.

Rule: Always ensure the input format matches the specified locale.

Returns

• number - The parsed numeric value
• null - If parsing fails (invalid input, letters, etc.)