A WebAssembly interpreter written in pure Python, no third-party libraries are used.
The wasm
version currently in use is: WebAssembly Specification, Release 2.0 (Draft 2025-04-25).
The wasi
version currently in use is: WASI Preview 1.
Also requires Python version >= 3.12.
$ pip install pywasm
-
First we need a wasm module! Grab our
example/fibonacci/bin/fibonacci.wasm
file and save a copy in a new directory on your local machine. -
Now, instantiate WebAssembly modules directly from underlying sources. This is achieved using the
Runtime.instance_from_file
method.
import pywasm
pywasm.log.lvl = 1
runtime = pywasm.core.Runtime()
m = runtime.instance_from_file('example/fibonacci/bin/fibonacci.wasm')
r = runtime.invocate(m, 'fibonacci', [10])
print(f'fibonacci(10) = {r[0]}')
A brief description for example
File | Description |
---|---|
example/blake2b.py | Blake2b hashing algorithm |
example/blake2b_direct.py | Make the hash result returned as a value, not as an output parameter |
example/fibonacci.py | Fibonacci, which contains loop and recursion |
example/fibonacci_env.py | Call python/native function in wasm |
example/pi.py | Calculate π using the agm algorithm |
example/wasi_httpbin.py | Execute a wasi program and http get from http://httpbin.org |
example/wasi_ll.py | Execute a wasi program and list information about a directory |
example/wasi_stdout.py | Execute a wasi program and capture stdout |
example/wasi_zen.py | Execute a wasi program and print poem "The Zen of Python" |
$ python script/build_wabt.py # Download wabt tools at res/wabt.
$ python script/build_spec.py # Download spec tests at res/spec.
$ python script/build_wasi.py # Download wasi tests at res/wasi-testsuite.
$ python test/example.py # Test example.
$ python test/spec.py # Test spec.
$ python test/wasi.py # Test wasi.
Pywasm is an interpreter implemented using pure python and its standard library, capable of running on any python platform. Due to the inherent characteristics of python, you should not have overly high expectations for pywasm's performance.
However, pywasm's design makes it jit-friendly. Compared to cpython, pywasm's performance on pypy can improve by approximately 10 times, significantly surpassing the average 3-fold improvement.