sci-math-wasm 🧪

High-performance scientific mathematical functions for the web, powered by Rust and WebAssembly.

Overview

sci-math-wasm provides a robust, typed, and highly optimized set of mathematical tools designed specifically for scientific applications, data analysis, and signal processing in browser environments.

Features

• 🚀 Performance: Hand-optimized Rust algorithms compiled to WASM.
• 📈 Analysis: Regression, Calculus, and Signal Processing (FFT).
• 📊 Statistics: Mean, median, variance, standard deviation.
• 🔢 Linear Algebra: Vector and Matrix operations.
• 📁 File I/O: Scientific data parsing (CSV, Excel, DAT, MPT) with auto-detection.
• 📐 Math Tools: Complex numbers, Polynomials, and Trigonometry.
• 📚 Documentation: Full VitePress site with LaTeX support.

Documentation

Full documentation is available at https://velosci.github.io/sci-math-wasm/ (locally in /docs).

Interactive benchmarks comparing WASM vs JavaScript performance for both mathematical computations and file processing are available at the /bench route.

Installation

npm install sci-math-wasm
# or
pnpm add sci-math-wasm

Usage

import init, { mean, fft, magnitude, TextStreamer, sniffFormat } from 'sci-math-wasm';

async function run() {
    await init();
    
    // === Mathematical Operations ===
    const data = new Float64Array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0]);
    
    // Statistics
    const avg = mean(data);
    console.log(`Mean: ${avg}`);
    
    // Signal Processing
    const spectrum = fft(data);
    const mags = magnitude(spectrum);
    console.log('Frequencies:', mags);
    
    // === File Processing ===
    // Auto-detect file format
    const fileHeader = new Uint8Array(await file.slice(0, 2048).arrayBuffer());
    const hint = sniffFormat(fileHeader);
    
    if (!hint.isBinary) {
        // Process text-based scientific files
        const streamer = new TextStreamer()
            .setDelimiter(hint.delimiter)
            .setSkipLines(hint.skipLines)
            .setCommentChar(hint.commentChar);
        
        // Stream processing for large files
        const reader = file.stream().getReader();
        while (true) {
            const { done, value } = await reader.read();
            if (done) break;
            const rows = streamer.processChunk(value);
            // Process rows incrementally
        }
    }
}

run();

Modules

• basic: Fundamental arithmetic and utilities.
• stats: Statistical analysis tools.
• linalg: Linear algebra (Vectors & Matrices).
• signal: Advanced signal processing (FFT).
• io: Scientific file parsing (CSV, Excel, DAT, MPT) with auto-detection.

Development

Build

Requires rust and wasm-pack.

wasm-pack build --target web

Documentation

cargo doc --open

License

MIT © VeloSci