[Python] Simulate data for binary classification in Rust, as Polars DataFrame/Series...

I am trying to translate the following Python code into Rust/Polars. The code creates a dataset for a binary classification: a response variable with values 0/1 and a set of features for every sample with values in the range [0.,1.).

import numpy as np
import pandas as pd

nsmpls = 50
nfeats = 10
smpls_names = [f'smpl{n}' for n in range(nsmpls)]
feats_names = [f'feat{n}' for n in range(nfeats)]

#response variable
y = np.random.randint(0, 2, size=nsmpls)
y = pd.Series(y, index=smpls_names)

#data
X = np.random.random((nfeats, nsmpls))
X = pd.DataFrame(X, columns=smpls_names, index=feats_names)


So far I have managed to create a column of integers for the response variable and a single column of features:

use rand::Rng;
use polars:relude::*;

fn main() {
let nsmpls = 50;
let nfeats=10;

let mut rng = rand::rng();
//unsuccessful attempt at generating sample names
// let smpls_names: Vec<String> = (0..nsmpls).map(|v| "{v}").collect();

//generating response variables
let y: Series = (0..nsmpls).map(|_| (rng.random_bool(0.5) as u32)).collect();

//generating a column of features
let feats_col: Vec<f64> = (0..nsmpls).map(|_| rng.random()).collect();
}


However, I am less successful in

• creating generic sample and feature names (commented line)
• generating required number of generic columns into a DataFrame (which is partially related to the previous issue)

Advice or code suggestions will be greatly appreciated.

Related:
I took some inspiration from this thread: How do I generate a vector of random numbers in a range?

Update
The suggested duplicate only partially answers the question. Indeed, by now I have already overcome this issue, but it still remains to generate a DataFrame with an arbitrary number of columns. Here is my current code:

use rand::Rng;
use polars:relude::*;

fn main() {
let nsmpls = 50;
let nfeats=10;

let mut rng = rand::rng();
//unsuccessful attempt at generating sample names
let smpls_names: Vec<String> = (0..nsmpls).map(|v| format!("smpl{v}")).collect();
let feats_names: Vec<String> = (0..nfeats).map(|v| format!("feat{v}")).collect();

//generating response variables
let status: Vec<u32> = (0..nsmpls).map(|_| (rng.random_bool(0.5) as u32)).collect();
let c1 = Column::new("Sample".into(), &smpls_names);
let c2 = Column::new("status".into(), &status);
let y = DataFrame::new(vec![c1, c2]);

//generating a column of features
let feats_col: Vec<f64> = (0..nsmpls).map(|_| rng.random()).collect();
}

Continue reading...