使用Py03从Rust调用Python函数时的最佳返回类型

我假设您已经将Rust环境设置为使用Py03和NumPy crate .或者更确切地说，在你的Cargo 中.在下面，设置如下.

[dependencies]
numpy = "0.15.0"  # Specify the version you need
pyo3 = "0.15.0"   # Specify the version you need
ndarray = "0.15.0" # Specify the version you need

use numpy::{PyArray2, IntoPyArray};
use pyo3::{Python, PyResult, PyModule, PyTuple};

fn call_python_function(py: Python) -> PyResult<ndarray::Array2<usize>> {
    let my_py_func_rs = PyModule::import(py, "__main__")?.getattr("my_py_func")?.into();

    // Assume my_input_array is a ndarray::Array1<u8>
    let args = PyTuple::new(py, &[my_input_array.into_pyarray(py)]);

    let output_pyarray = my_py_func_rs.call1(py, args)?.extract::<&PyArray2<i64>>(py)?;
    
    // Convert to ndarray::Array2<usize>
    let output_array = output_pyarray.as_array().mapv(|x| x as usize);

    Ok(output_array)
}

look at pyo3 documentation for datastructure for output array here: https://docs.rs/numpy/latest/numpy/array/struct.PyArray.html and
https://docs.rs/numpy/latest/numpy/array/type.PyArray2.html .

当前版本为NumPy="0.20.0"

此外，您还可以

use ndarray::Array2;

// ...

let output_array: Vec<Vec<usize>> = my_py_func_rs.call1(py, args).unwrap().extract(py).unwrap();
let output_array_nd = Array2::from_pyarray(py, output_array);