larzmatrix

Small-matrix linear algebra in pure Python. No numpy, zero dependencies.

A Matrix type with the operations you need for small systems: arithmetic, transpose, determinant, inverse, and solving Ax = b by Gaussian elimination with partial pivoting. For a handful of dimensions - transforms, geometry, least squares, teaching - this is all you need, and it installs anywhere.

from larzmatrix import Matrix

A = Matrix([[2, 1], [1, 3]])
A.determinant()                 # 5.0
A.inverse()
A.solve([3, 5])                 # x such that A x = [3, 5]
A * Matrix([[1, 0], [0, 1]])    # matrix product
A.T                             # transpose

Why

• No heavyweight dependency. numpy is wonderful and enormous; when you just need to invert a 3x3 or solve a small system, larzmatrix is a few hundred readable lines and zero install cost.
• The core operations, correct. Determinant, inverse (Gauss-Jordan), and solve all use partial pivoting for numerical stability, and singular matrices raise instead of returning nonsense.
• Clean API. + - * (scalar and matrix), .T, .trace(), .identity(n), tolerant ==.

Install

pip install larzmatrix

Usage

from larzmatrix import Matrix

Matrix.identity(3)
Matrix([[1, 2], [3, 4]]) * Matrix([[5, 6], [7, 8]])   # [[19, 22], [43, 50]]
Matrix([[1, 2], [3, 4]]).determinant()                # -2.0
Matrix([[4, 7], [2, 6]]).inverse()
Matrix([[2, 1], [1, 3]]).solve([5, 10])               # [1.0, 3.0]

Tests

python -m unittest discover -s tests -v   # 19 tests incl. det/inverse/solve

The Larz stack

One of 30+ pure-Python, zero-dependency libraries at github.com/larz-scripter - pairs with larzstats.

License

MIT (c) larz-scripter