How to do proper conversion from Polars Dataframe to Smartcore Densematrix in rust

Question

So i have been playing around with polars and smartcore but i have some trouble with converting from a Dataframe into a DenseMatrix.

In my code I'm reading in some simple csv data consisting of several columns containing numeric data.

This gets read into a Dataframe (polars) but a DenseMatrix (smartcore) wants as input (for instance) a Vec<Vec<f64>>.

So i do the conversion but it seems very troublesome, and non ideomatic.

// select some columns and concat the data into a single column containg lists of the column data.

// |    Data   |
// | [1, 2, 3] |
// | [1, 2, 3] |
// | [1, 2, 3] |
// | [1, 2, 3] |

let n = [concat_list([col("Pclass"), col("SibSp"), col("Parch")])?.alias("data")];
let df = df.clone().lazy().select(n).collect()?;

// iterate through the column and convert each number in the nested list into an f64
let data = df
    .column("data")?
    .list()?
    .into_no_null_iter()
    .map(|n| {
        n.i64()?
        .into_no_null_iter()
        .map(|n| n as f64)
        .collect::<Vec<f64>>()
    })
    .collect::<Vec<Vec<f64>>>();

let matrix = DenseMatrix::from_2d_vec(&data)

This solution above works, but it doesn't feel ideomatic, and the documentation for polars is outdated and lacking when it comes to rust. Does anyone have a better solution to this type of problem?