I want to use a Rust async method in Python. I'm trying to use PyO3 or rust-cpython.
For example, for sync Rust functions, I can use,
#[pyfunction]
fn myfunc(a: String) -> PyResult<String> {
let mut contents = String::new();
contents = a.to_string() + " appended";
Ok((contents))
}
#[pymodule]
fn MyModule(py: Python, m: &PyModule) -> PyResult<()> {
m.add_wrapped(wrap_pyfunction!(urlshot))?;
Ok(())
}
For async methods, how can I do it? For example, I want to call the following method in Python,
async fn hello_world() {
println!("hello, world!");
}
It makes use of Python async features using asyncio/await provided in Python 3. The time and queue modules have been replaced with the asyncio package. This gives your program access to asynchronous friendly (non-blocking) sleep and queue functionality.
An async function uses the await keyword to denote a coroutine. When using the await keyword, coroutines release the flow of control back to the event loop. To run a coroutine, we need to schedule it on the event loop. After scheduling, coroutines are wrapped in Tasks as a Future object.
Python's asyncio package (introduced in Python 3.4) and its two keywords, async and await , serve different purposes but come together to help you declare, build, execute, and manage asynchronous code.
Since there was no easy way of solving this issue (at least, I hadn't found), I converted my async method to sync one. And called it on Python side as,
async fn my_method(s: &str) -> Result<String, Error> {
// do something
}
#[pyfunction]
fn my_sync_method(s: String) -> PyResult<String> {
let mut rt = tokio::runtime::Runtime::new().unwrap();
let mut contents = String::new();
rt.block_on(async {
result = format!("{}", my_sync_method(&s).await.unwrap()).to_string();
});
Ok((result))
}
#[pymodule]
fn MyModule(py: Python, m: &PyModule) -> PyResult<()> {
m.add_wrapped(wrap_pyfunction!(my_sync_method))?;
Ok(())
}
In the Cargo.toml
file, I added the following dependencies,
[dependencies.pyo3]
git = "https://github.com/PyO3/pyo3"
features = ["extension-module"]
After running cargo build --release
, target/release/libMyModule.so
binary file is generated. Rename it as MyModule.so
and it now can be imported from Python.
import MyModule
result = MyModule.my_sync_method("hello")
Using setuptools-rust, I could bundle it as an ordinary Python package.
All of the above code and commands are tested on newly-released Linux Mint 20. On MacOS, the binary file will be libMyModule.dylib
.
If you want to use Python to control Rust's async function, I don't think it will work (Or at least it is very complicated, as you need to connect two different future
mechanism). For async functions, Rust compiler will maintain a state machine to manage the coroutines run correctly under await's control. This is an internal state of Rust applications and Python cannot touch it. Similarly Python interpreter also has a state machine that cannot be touched by Rust.
I do found this topic about how to export an async function using FFI. The main idea is to wrap the async in a BoxFuture
and let C control the timing of returning it to Rust. However, you cannot use BoxFuture
in PyO3 since its pyfunction
macro cannot convert a function returns BoxFuture
to a Python callback. You may try to create a library using FFI and use python's cffi module to load it.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With