SciRS2: Scientific Computing in Rust

SciRS2 is a comprehensive scientific computing library for Rust, inspired by SciPy and designed to provide a complete ecosystem for numerical computation, statistical analysis, and scientific algorithms.

Important: From v0.1.0, SciRS2 follows the SciRS2 POLICY for consistent ecosystem architecture with layered abstractions.

Overview

This is the main SciRS2 crate, which provides a convenient facade over the ecosystem of specialized sub-crates. Each sub-crate focuses on a specific domain of scientific computing, while this crate re-exports their functionality in a unified interface.

Features

SciRS2 brings together a large collection of scientific computing tools:

• Core Utilities: Common functionality through scirs2-core
• Linear Algebra: Matrix operations, decompositions, and solvers via scirs2-linalg
• Statistics: Distributions, hypothesis testing, and statistical functions in scirs2-stats
• Optimization: Minimization, root finding, and curve fitting with scirs2-optimize
• Integration: Numerical integration and ODE solvers through scirs2-integrate
• Interpolation: Various interpolation methods in scirs2-interpolate
• Special Functions: Mathematical special functions via scirs2-special
• Fourier Analysis: FFT and related transforms in scirs2-fft
• Signal Processing: Filtering, convolution, and spectral analysis in scirs2-signal
• Sparse Matrices: Efficient sparse matrix formats and operations via scirs2-sparse
• Spatial Algorithms: Spatial data structures and algorithms in scirs2-spatial
• Image Processing: Multidimensional image processing in scirs2-ndimage
• Machine Learning: Clustering, metrics, and neural networks via multiple sub-crates
• Data I/O: Reading and writing various scientific data formats via scirs2-io
• And more: Various additional modules for specific applications

Installation

Add the following to your Cargo.toml:

[dependencies]
scirs2 = "0.1.0"  # Main package with default features

You can enable only the features you need:

[dependencies]
scirs2 = { version = "0.1.0", features = ["linalg", "stats", "optimize"] }

Or use specific modules directly:

[dependencies]
scirs2-core = "0.1.0"
scirs2-linalg = "0.1.0"
scirs2-stats = "0.1.0"

Basic usage examples:

use scirs2::prelude::*;
use ndarray::array;

fn main() -> CoreResult<()> {
    // Linear algebra operations
    let a = array![[1., 2.], [3., 4.]];
    let eig = linalg::eigen::eig(&a)?;
    println!("Eigenvalues: {:?}", eig.eigenvalues);
    println!("Eigenvectors: {:?}", eig.eigenvectors);
    
    // Statistical distributions
    let normal = stats::distributions::normal::Normal::new(0.0, 1.0)?;
    let samples = normal.random_sample(1000, None)?;
    let mean = stats::descriptive::mean(&samples)?;
    let std_dev = stats::descriptive::std_dev(&samples, None)?;
    println!("Sample mean: {}, std dev: {}", mean, std_dev);
    
    // Optimization
    let f = |x: &[f64]| x[0].powi(2) + x[1].powi(2);
    let df = |x: &[f64], grad: &mut [f64]| {
        grad[0] = 2.0 * x[0];
        grad[1] = 2.0 * x[1];
    };
    let result = optimize::unconstrained::minimize(
        f, df, &[1.0, 1.0], "L-BFGS-B", None, None)?;
    println!("Optimization result: {:?}", result);
    
    // Special functions
    let gamma = special::gamma::gamma(5.0)?;
    println!("Gamma(5) = {}", gamma);
    
    // FFT
    let signal = array![1.0, 2.0, 3.0, 4.0];
    let fft_result = fft::fft(&signal)?;
    println!("FFT result: {:?}", fft_result);
    
    Ok(())
}

Feature Flags

This crate uses feature flags to control which sub-crates are included:

• core: Core utilities (always enabled)
• linalg: Linear algebra operations
• stats: Statistical functions and distributions
• optimize: Optimization algorithms
• integrate: Numerical integration and ODEs
• interpolate: Interpolation methods
• fft: Fast Fourier Transform
• special: Special functions
• signal: Signal processing
• sparse: Sparse matrices
• spatial: Spatial algorithms
• ndimage: N-dimensional image processing
• cluster: Clustering algorithms
• datasets: Dataset utilities
• io: I/O utilities
• neural: Neural networks
• graph: Graph algorithms
• transform: Data transformation
• metrics: Evaluation metrics
• text: Text processing
• vision: Computer vision
• series: Time series analysis
• autograd: Automatic differentiation

Architecture

SciRS2 follows a modular architecture where each domain of scientific computing is implemented in a separate crate. This main crate provides a unified interface by re-exporting their functionality.

SciRS2 POLICY

Starting with v0.1.0, SciRS2 follows a strict layered architecture:

• Core-Only Dependencies: Only scirs2-core uses external dependencies directly
• Unified Abstractions: All other crates use scirs2-core abstractions for rand, ndarray, etc.
• Ecosystem Consistency: Ensures consistent APIs, centralized version control, and type safety
• Performance Benefits: Enables better optimization through centralized abstractions

See the SciRS2 Ecosystem Policy for complete details.

Architecture Benefits

• Modular Development: Each domain can be developed and tested independently
• Reduced Compilation Time: Users can include only the features they need
• Flexible Dependencies: Sub-crates can have different dependency requirements
• Focused Documentation: Each domain has its own focused documentation
• Consistent API: Unified interfaces through scirs2-core abstractions

Module Structure

scirs2
├── core       // Core utilities
├── linalg     // Linear algebra
├── stats      // Statistics
├── optimize   // Optimization
├── integrate  // Integration
├── interpolate // Interpolation
├── fft        // Fourier transforms
├── special    // Special functions
├── signal     // Signal processing
├── sparse     // Sparse matrices
├── spatial    // Spatial algorithms
├── ndimage    // Image processing
└── ...        // Other modules

Performance

SciRS2 is designed with performance in mind:

• Uses optimized BLAS and LAPACK implementations where appropriate
• Leverages SIMD operations when available
• Provides parallel processing capabilities
• Uses memory-efficient algorithms
• Employs caching for expensive computations

Contributing

See the CONTRIBUTING.md file for contribution guidelines.

License

This project is dual-licensed under:

• MIT License
• Apache License Version 2.0

You can choose to use either license. See the LICENSE file for details.