Kiana Sheibani
983733f241
LArrays were originally devised for efficiency reasons, but now that Delayed array representations exist there isn't any need for it. |
||
---|---|---|
src/Data | ||
.gitignore | ||
LICENSE | ||
numidr.ipkg | ||
README.md | ||
sirdi.json |
NumIdr
NumIdr is a linear algebra and data manipulation library for Idris 2. It features an efficient, type-safe array data structure, as well as utilities for working with vector spaces and matrices.
The name is pronounced like "num-idge".
Features
-
A type-safe and efficient array type, based on NumPy's arrays.
-
Compile-time checked indexing operations to avoid run-time overhead.
-
Utility functions and operations for working with vectors, matrices, homogeneous coordinates, and other linear algebra computations.
-
Transform types for working with rotations, reflections, isometries, and other types of affine maps.
Inspiration
NumIdr is inspired by many different data science and linear algebra libraries, including Python's NumPy, Rust's nalgebra, and Haskell's massiv. It aims to combine the most useful features of each library.
Documentation
Most of the exported utility functions have docstrings. There are also plans for an in-depth tutorial on NumIdr's features, though that is not available yet.
Usage
Basic package install:
idris2 --install numidr.ipkg