Segmentation fault

### Issue
Like a few others before me, I was getting a segmentation fault when trying to embed python within rust.
But the Rust compiler was not catching the error.

### Discussion
After a lot of debugging, I believe the issue was the lifetime for the `Result`.

The easiest solution I came to was to keep the config, the interpreter, and the consumer all within the same function.

I had tried defining the lifetimes of variables, but that didn't work.

```rs
async fn runner(input: &str) -> String {
let mut config: OxidizedPythonInterpreterConfig<'_> = OxidizedPythonInterpreterConfig::default();
config.interpreter_config.isolated = Some(true);
config.interpreter_config.filesystem_encoding = Some("utf-8".to_string());
config.set_missing_path_configuration = false;
config.interpreter_config.parse_argv = Some(false);
config.argv = Some(vec!["python".into()]);
config.interpreter_config.executable = Some("python".into());

let embeded_python = MainPythonInterpreter::new(config).unwrap();
let py_ai: &str = include_str!(concat!(env!("CARGO_MANIFEST_DIR"), "/python/ai.py"));

let from_python: Result, PyErr> = embeded_python.with_gil(|py: Python<'_>| -> PyResult> {
let app: Py = PyModule::from_code(py, py_ai, "", "")?
.getattr("run")?.into();

app.call1(py, (input,))
});

match from_python {
Ok(result) => {
return result.to_string();
},
Err(err) => {
eprintln!("Error: {:?}", err);
return String::new();
}
}
}
```

**NOTE** The above configs worked for me. The official docs do not provide any examples of working configs.

### Recommendation
Learn from my folly. Hopefully this helps others who are trying to embed python but do not have example configurations.

Start with the embedded Rust example in the issue, including the configuration fields and python/ai.py, then inspect the official embedding documentation for where examples belong. Done means the documented configuration and embedding example explain a working setup and help prevent the reported segmentation fault; no specific documentation file or test is named.

Written by the indexing model from the issue text.