Example: matrix-math

This is an example of how to use componentize-py to build a CLI app that does matrix multiplication using NumPy inside a sandboxed environment. This demonstrates using a non-trivial Python package containing native extensions within a guest component.

Prerequisites

• wasmtime 38.0.0 or later
• componentize-py 0.25.0
• NumPy, built for WASI

Note that we use an unofficial build of NumPy since the upstream project does not yet publish WASI builds.

Below, we use Rust's cargo to install Wasmtime. If you don't have cargo, you can download and install from https://github.com/bytecodealliance/wasmtime/releases/tag/v38.0.0.

cargo install --version 38.0.0 wasmtime-cli
pip install componentize-py==0.25.0
curl -OL https://github.com/dicej/wasi-wheels/releases/download/v0.0.2/numpy-wasi.tar.gz
tar xf numpy-wasi.tar.gz

Running the demo

componentize-py -d ../../wit -w matrix-math componentize app -o matrix-math.wasm
wasmtime run matrix-math.wasm '[[1, 2], [4, 5], [6, 7]]' '[[1, 2, 3], [4, 5, 6]]'

The second command above should print the following:

matrix_multiply received arguments [[1, 2], [4, 5], [6, 7]] and [[1, 2, 3], [4, 5, 6]]
[[9, 12, 15], [24, 33, 42], [34, 47, 60]]

If you run into any problems, please file an issue!