Getting started

Installation

Install the package#

npm install @cyfora/numera

pnpm add @cyfora/numera

yarn add @cyfora/numera

The package ships a precompiled C++ addon for each supported platform. When you import "@cyfora/numera", it loads the matching binary. You do not need a compiler, CMake or Python.

Supported platforms#

Windows and Alpine/musl Linux have no prebuilt binaries yet. The prebuilds use the stable Node-API, so one binary works on every supported Node.js version.

Importing#

numera is an ES module. The default export is the np namespace, the same object you would get from import numpy as np:

import np from "@cyfora/numera";

np.arange(4);   // => [0, 1, 2, 3]

Named imports work too:

import { array, zeros, linalg } from "@cyfora/numera";

zeros([2]);                        // => [0, 0]
linalg.det(array([[1, 2], [3, 4]])); // => -2.0000000000000004

In a CommonJS project, use a dynamic import: const { default: np } = await import("@cyfora/numera").

Check the installation#

import np from "@cyfora/numera";

np.linalg.solve([[3, 1], [1, 2]], [9, 8]); // => [2, 3]
typeof np.linalg.backend();                // => "string"

np.linalg.backend() returns "accelerate" on macOS, where linear algebra runs on Apple Accelerate, and "fallback" elsewhere, where it uses numera's portable C++ implementation.

TypeScript#

Type declarations are included, so you do not need an @types package. See TypeScript for the exported types.