Are there any scientific packages for Lua comparable to Scipy?
There is the basis for one in Numeric Lua.
You should try Torch7 (github).
Torch7 has a very nice and efficient vector/matrix/tensor numerical library with a Lua front-end. It also has a bunch of functions for computer vision and machine learning.
It's pretty recent but getting better quickly.
One can always use Lunatic Python and access scipy inside lua.
> require("python")
> numpy = python.import("numpy")
> numpy.array ... etc ..
You have some options:
- Numeric Lua - C module for Lua 5.1/5.2, provides matrices, FFT, complex numbers and others
- GSL Shell - Modification of Lua (supports Lua libraries) with a nice syntax. Provides almost everything that Numeric Lua does, plus ODE solvers, plotting capabilities, and other nice things. Has a great documentation.
- SciLua - Pure LuaJIT module. Aims to be a complete framework for scientific computing in Lua. Provides vectors and matrices, random numbers / distributions, optimization, others. Still in early development.
- Lua Numerical Algorithms - Pure LuaJIT module (uses blas/lapack via LuaJIT FFI). Provides matrices / linear algebra, FFT, complex numbers, optimization algorithms, ODE solver, basic statistics (+ PCA, LDA), and others. Still in early development, but has a somewhat complete documentation and test suits.
There is the basis for one in Numeric Lua.
I'm not sure if it is comparable to Scipy, but there is GSL Shell which is based on LuaJIT and GNU Scientific Library, which offers many numerical algorithms and vector/matrix linear algebra operations.
There's a Numpy-like extension for Lua which runs without dependencies at
https://github.com/jzrake/lunum
In the future it will provide FFT's and linear algebra like Numpy+Scipy. Presently it supports numeric array manipulation like in Numpy.
© 2022 - 2024 — McMap. All rights reserved.