Crate sprs

source ·
Expand description

sprs is a sparse linear algebra library for Rust.

It features a sparse matrix type, CsMat, and a sparse vector type, CsVec, both based on the compressed storage scheme.

§Features

  • sparse matrix/sparse matrix addition, multiplication.
  • sparse vector/sparse vector addition, dot product.
  • sparse matrix/dense matrix addition, multiplication.
  • sparse triangular solves.
  • powerful iteration over the sparse structure, enabling easy extension of the library.
  • matrix construction using the triplet format, vertical and horizontal stacking, block construction.
  • sparse cholesky solver in the separate crate sprs-ldl.
  • fully generic integer type for the storage of indices, enabling compact representations.
  • planned interoperability with existing sparse solvers such as SuiteSparse.

§Quick Examples

Matrix construction:

use sprs::{CsMat, TriMat};

let mut a = TriMat::new((4, 4));
a.add_triplet(0, 0, 3.0_f64);
a.add_triplet(1, 2, 2.0);
a.add_triplet(3, 0, -2.0);

// This matrix type does not allow computations, and must to
// converted to a compatible sparse type, using for example
let b: CsMat<_> = a.to_csr();

Constructing matrix using the more efficient direct sparse constructor

use sprs::{CsMat, CsVec};
let eye : CsMat<f64> = CsMat::eye(3);
let a = CsMat::new_csc((3, 3),
                       vec![0, 2, 4, 5],
                       vec![0, 1, 0, 2, 2],
                       vec![1., 2., 3., 4., 5.]);

Matrix vector multiplication:

use sprs::{CsMat, CsVec};
let eye = CsMat::eye(5);
let x = CsVec::new(5, vec![0, 2, 4], vec![1., 2., 3.]);
let y = &eye * &x;
assert_eq!(x, y);

Matrix matrix multiplication, addition:

use sprs::{CsMat, CsVec};
let eye = CsMat::eye(3);
let a = CsMat::new_csc((3, 3),
                       vec![0, 2, 4, 5],
                       vec![0, 1, 0, 2, 2],
                       vec![1., 2., 3., 4., 5.]);
let b = &eye * &a;
assert_eq!(a, b.to_csc());

Re-exports§

Modules§

  • approx
    Traits for comparing vectors and matrices using the approx traits
  • array_backend
    Fixed size arrays usable for sparse matrices.
  • binop
    Sparse matrix addition, subtraction
  • errors
    Error type for sprs
  • indexing
    Abstraction over types of indices
  • io
    Serialization and deserialization of sparse matrices
  • linalg
    Sparse linear algebra
  • num_kinds
    Trait to be able to know at runtime if a generic scalar is an integer, a float or a complex.
  • num_matrixmarket
  • prod
    Sparse matrix product
  • smmp
    Implementation of the paper Bank and Douglas, 2001, Sparse Matrix Multiplication Package (SMPP)
  • special_mats
    Common sparse matrices
  • stack
  • vec
  • visu

Structs§

Enums§

Traits§

  • DenseVector
    A trait for types representing dense vectors, useful for expressing algorithms such as sparse-dense dot product, or linear solves.
  • DenseVectorMut
    Trait for dense vectors that can be modified, useful for expressing algorithms which compute a resulting dense vector, such as solvers.
  • MulAcc
    Trait for types that have a multiply-accumulate operation, as required in dot products and matrix products.
  • SparseMat
    A trait for common members of sparse matrices

Functions§

  • assign_to_dense
    Assign a sparse matrix into a dense matrix
  • bmat
    Specify a sparse matrix by constructing it from blocks of other matrices
  • hstack
    Construct a sparse matrix by horizontally stacking other matrices
  • is_symmetric
  • kronecker_product
    Compute the Kronecker product between two matrices
  • perm_is_valid
  • permute_cols
    Compute the matrix resulting from the product A * P
  • permute_rows
    Compute the matrix resulting from the product P * A
  • transform_mat_papt
    Compute the square matrix resulting from the product P * A * P^T
  • transform_mat_paq
    Compute the matrix resulting from the product P * A * Q, using the same number of operations and allocation as for computing either P * A or A * Q by itself.
  • vstack
    Construct a sparse matrix by vertically stacking other matrices

Type Aliases§