Crate rulinalg

Source
Expand description

§The rulinalg crate.

A crate that provides high-dimensional linear algebra implemented entirely in Rust.

This crate provides two core data structures: Matrix and Vector. These structs are designed to behave as you would expect with relevant operator overloading.

The library currently contains (at least) the following linear algebra methods:

  • Matrix inversion
  • LUP decomposition
  • QR decomposition
  • SVD decomposition
  • Cholesky decomposition
  • Eigenvalue decomposition
  • Upper Hessenberg decomposition
  • Linear system solver
  • Other standard transformations, e.g. Transposing, concatenation, etc.

§Usage

Specific usage of modules is described within the modules themselves. This section will highlight the basic usage.

We can create new matrices.

use rulinalg::matrix::Matrix;

// A new matrix with 3 rows and 2 columns.
let a = Matrix::new(3, 2, vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0]);

The matrices are stored in row-major order. This means in the example above the top row will be [1,2,3].

We can perform operations on matrices.

use rulinalg::matrix::Matrix;

// A new matrix with 3 rows and 2 columns.
let a = Matrix::new(3, 2, vec![1.0, 2.0, 3.0, 4.0, 5.0, 6.0]);
let b = Matrix::new(3, 2, vec![6.0, 5.0, 4.0, 3.0, 2.0, 1.0]);

// Produces a 3x2 matrix filled with sevens.
let c = a + b;

Sometimes we want to construct small matrices by hand, usually for writing unit tests or examples. For this purpose, rulinalg provides the matrix! macro:

// Remember to enable macro usage in rulinalg!
#[macro_use]
extern crate rulinalg;

// Construct a 3x3 matrix of f64
// Commas separate columns and semi-colons separate rows
let mat = matrix![1.0, 2.0, 3.0;
                  4.0, 5.0, 6.0;
                  7.0, 8.0, 9.0];

Of course the library can support more complex operations but you should check the individual modules for more information.

§Matrix Slices

Often times it is desirable to operate on only a sub-section of a Matrix without copying this block. Rulinalg allows this via the MatrixSlice and MatrixSliceMut structs. These structs can be created from Matrix structs and follow all of the borrowing rules of Rust.

Note finally that much of the Matrix/MatrixSlice/MatrixSliceMut functionality is contained behind the BaseMatrix/BaseMatrixMut traits. This allows us to be generic over matrices or slices.

Re-exports§

pub use norm::VectorNorm;
pub use norm::MatrixNorm;
pub use norm::VectorMetric;
pub use norm::MatrixMetric;

Modules§

convert
The convert module.
error
Error handling for the linalg module.
macros
Macros for the linear algebra modules.
matrix
The matrix module.
norm
The norm module
ulp
Tools for ULP-based comparison of floating point numbers.
utils
Linear algebra utils module.
vector
The vector module.

Macros§

assert_matrix_eq
Compare matrices for exact or approximate equality.
assert_vector_eq
Compare vectors for exact or approximate equality.
matrix
The matrix! macro enables easy construction of small matrices.
vector
The vector! macro enables easy construction of small vectors.