NumRS2 0.1.1 Released — The Installable First-Stable Baseline

First-stable, now genuinely installable.

Today we released NumRS2 0.1.1 — a same-day point release that locks in our first stable as the installable baseline.

Same pure-Rust promise as 0.1.0: no C, no Fortran, no system BLAS/LAPACK — just N-dimensional arrays that compile to a single static binary or WASM module.

What changed

This is an honest, unglamorous one: a same-day quality and packaging stabilization on top of this morning’s 0.1.0 first-stable. As the release went out, a few first-day rough edges surfaced around publishing and packaging, and the cleanly resolvable version on crates.io landed at 0.1.1.

What that means for you:

• cargo add numrs2 now resolves to 0.1.1 — the genuinely installable baseline.
• The public API is unchanged. This is a drop-in over 0.1.0; nothing in your code needs to move.
• The bar holds. Zero compilation warnings, zero clippy errors, and 1,111+ unit tests still passing across the same ~155,000 lines.

There is no separate feature changelog to dress up. The CHANGELOG groups 0.1.0 and 0.1.1 together as the initial release for exactly this reason — 0.1.1 is the version that installs cleanly.

Getting Started

cargo add numrs2

use numrs2::prelude::*;

let a = Array::from_vec(vec![1.0, 2.0, 3.0, 4.0]).reshape(&[2, 2]);
let b = a.add(&a);          // element-wise, NumPy-style
println!("{}", b);

Tips

• Pin the installable baseline. Use numrs2 = "0.1.1" so cargo add and lockfiles land on the version that resolves cleanly.
• The defaults are already useful. matrix_decomp and scirs are on out of the box — decompositions and the SciRS2 SIMD/distribution integration come for free.
• Add the rest only when you need it. Reach for features = ["arrow"] for Apache Arrow zero-copy, ["python"] for PyO3 bindings, or ["gpu"] for wgpu acceleration — no need to carry them otherwise.

NumRS2 remains the NumPy-class N-dimensional array core of the COOLJAPAN stack, launched alongside same-day siblings OptiRS (optimization) and PandRS (dataframes), on top of OxiBLAS and OxiCode.

Repository: https://github.com/cool-japan/numrs

Star the repo if a clean, installable pure-Rust array core is what your stack has been missing.

— KitaSan at COOLJAPAN OÜ December 30, 2025