text-processing-rs

Inverse Text Normalization (ITN) for Rust — convert spoken-form ASR output to written form.

What it does

Converts spoken-form ASR output to written form:

Usage

Rust

use text_processing_rs::normalize;

let result = normalize("two hundred");
assert_eq!(result, "200");

let result = normalize("five dollars and fifty cents");
assert_eq!(result, "$5.50");

Swift

import NemoTextProcessing

let result = NemoTextProcessing.normalize("two hundred")
// result is "200"

let money = NemoTextProcessing.normalize("five dollars and fifty cents")
// money is "$5.50"

Compatibility

98.6% compatible with the original NeMo test suite (1200/1217 tests passing).

Features

• Cardinal and ordinal number conversion
• Decimal numbers with scale words (million, billion)
• Currency formatting (USD, with scale words)
• Measurements including temperature (°C, °F, K) and data rates (gbps)
• Date parsing (multiple formats)
• Time parsing with AM/PM and timezone preservation
• Email and URL normalization
• Phone numbers, IP addresses, SSN
• Case preservation for proper nouns and abbreviations

Building

Rust

cargo build
cargo test

Swift (XCFramework)

# Install Rust targets
rustup target add aarch64-apple-darwin x86_64-apple-darwin
rustup target add aarch64-apple-ios aarch64-apple-ios-sim

# Build XCFramework
./build-xcframework.sh

Output:

• output/NemoTextProcessing.xcframework - Add to Xcode project
• output/NemoTextProcessing.swift - Swift wrapper

License

Apache 2.0

Acknowledgments

This project is a Rust implementation based on the inverse text normalization grammars from NVIDIA NeMo Text Processing. All credit for the original algorithms and test cases goes to the NVIDIA NeMo team.