Sentence Conversion

Sentence Conversion

num2words_sentence converts numbers embedded in natural-language text while keeping the surrounding sentence intact.

from num2words2 import num2words_sentence

num2words_sentence("I bought 6 apples.")
# 'I bought six apples.'

Aliases:

from num2words2 import convert_sentence, sentence_to_words

Both aliases call the same implementation.

What It Handles

The sentence converter recognizes common number contexts:

• regular numbers
• decimals
• negative values
• ordinals such as 1st, 2nd, 3rd, 4th
• dates such as April 5, 2024
• year-like values in date contexts
• temperatures such as 25C, 25 C, 25°C, and language-specific degree words
• currency symbols such as $, €, £, and ¥

Examples

num2words_sentence("The temperature is -5 degrees.")
# 'The temperature is minus five degrees.'

num2words_sentence("On April 5, 2024, I paid $12.50.")
# 'On April fifth, twenty twenty-four, I paid twelve dollars, fifty cents.'

num2words_sentence("The 1st place winner got $100.")
# 'The first place winner got one hundred dollars.'

Pass lang= when you know the sentence language:

num2words_sentence("J'ai acheté 5 livres pour 20 euros.", lang="fr")

num2words_sentence("Tengo 100 dólares.", lang="es")

Language Detection

If lang is omitted, the converter tries language detection when optional detection libraries are installed. It falls back to simple keyword heuristics and then English.

For deterministic production systems, pass lang= explicitly.

Output Mode

The to= argument is passed through to individual conversions where appropriate.

num2words_sentence("Items 1, 2, and 3", to="ordinal")

Context-specific detections such as dates, temperatures, and currencies may override the generic mode when that gives a more natural result.

Limitations

Sentence conversion is heuristic. It is designed for common text-normalization cases, not full natural-language parsing. For high-stakes or domain-specific text, run representative tests for each language and content type you plan to process.