orthos 0.1.3

High-performance linear algebra library written in Rust

orthos

High-performance linear algebra library written in Rust with Python bindings.

Installation

pip install orthos

Quick Start

from orthos import Matrix, Vector

# Create matrices
a = Matrix([[1.0, 2.0], [3.0, 4.0]])
b = Matrix.identity(2)

# Matrix operations
c = a @ b           # Matrix multiplication
d = a + b           # Addition
e = a.transpose()   # Transpose
f = a.inverse()     # Inverse

# Properties
print(a.determinant())  # -2.0
print(a.trace())        # 5.0
print(a.shape)          # (2, 2)

# Element access
print(a[0, 0])      # 1.0
a[0, 0] = 10.0      # Set element

# Vectors
v = Vector([1.0, 2.0, 3.0])
w = Vector([4.0, 5.0, 6.0])
print(v.dot(w))         # 32.0
print(v.norm())         # 3.7416...
print(v.normalize())    # Unit vector

# NumPy interoperability
import numpy as np
np_array = np.array([[1.0, 2.0], [3.0, 4.0]])
m = Matrix.from_numpy(np_array)
back_to_numpy = m.to_numpy()

Features

Matrix Operations

• Creation: zeros, ones, identity, from lists, from NumPy
• Arithmetic: +, -, * (element-wise), @ (matrix multiply)
• Methods: transpose(), inverse(), determinant(), trace(), scale()
• Indexing: matrix[row, col]

Vector Operations

• Creation: zeros, ones, from lists, from NumPy
• Arithmetic: +, -, scalar multiply
• Methods: dot(), norm(), norm_squared(), normalize(), scale()
• Indexing: vector[index]

Matrix-Vector Operations

• matvec(matrix, vector) - Matrix-vector multiplication
• matrix.matvec(vector) - Method form

NumPy Interoperability

• Matrix.from_numpy(array) - Create from NumPy array
• matrix.to_numpy() - Convert to NumPy array
• Vector.from_numpy(array) - Create from NumPy array
• vector.to_numpy() - Convert to NumPy array

Supported Platforms

Platform	Architectures
Linux	x86_64, aarch64
macOS	x86_64, Apple Silicon
Windows	x86_64

Python Version Support

• Python 3.9
• Python 3.10
• Python 3.11
• Python 3.12
• Python 3.13

Performance

orthos is built on faer, a high-performance linear algebra library in Rust. It provides competitive performance especially for small to medium matrices.

Links

• Source Code
• Issue Tracker
• Rust crate

License

Apache-2.0