How to pass rust polars DataFrame to python object?

I know there's this question. However the code I want is a little different.

[dependencies]

ordered-float: "4.4"

polars: {version="0.43", features = ["lazy", "parquet", "dtype-struct", "dtype-u8", "round_series"]}

itertools = "0.13"

chrono: "0.4"

pyo3: { version = "0.21", features = ["extension-module"] }

pyo3-polars: "0.18"

I want this to work:

use pyo3::prelude::*;
use pyo3::{Python, IntoPy};
use polars::prelude::*;
use pyo3_polars::PyDataFrame;

fn get_df() -> DataFrame{
    let mut symbols: Vec<i32> = Vec::new();
    let mut prices: Vec<i32> = Vec::new();
    df!(
        "Symbol" => &symbols,
        "Price" => &prices
    ).unwrap()
}
fn to_python() -> PyResult<PyDataFrame> {
    let df:DataFrame = get_df(); 
    let pydf = PyDataFrame(df);
    Python::with_gil(|py| {
        Ok(pydf)
    })
}

However I got:

error[E0308]: mismatched types
   --> src/lib.rs:62:32
    |
62  |         let pydf = PyDataFrame(df);
    |                    ----------- ^^ expected `polars_core::frame::DataFrame`, found `polars::prelude::DataFrame`
    |                    |
    |                    arguments to this struct are incorrect
    |
    = note: `polars::prelude::DataFrame` and `polars_core::frame::DataFrame` have similar names, but are actually distinct types

How to solve this? Thank you!