Crate featrs 

featrs — feature engineering for Rust, inspired by scikit-learn.

Built on Polars, all transformations operate on DataFrame and preserve column names throughout.

Quick start

use featrs::prelude::*;
use polars::prelude::{Column, DataFrame, NamedFrom, Series};

let col = Column::from(Series::new("x".into(), &[1.0_f64, 2.0, 3.0]));
let df = DataFrame::new(3, vec![col])?;

let mut scaler = StandardScaler::new();
scaler.fit(df.clone())?;
let scaled = scaler.transform(df)?;
assert_eq!(scaled.height(), 3);

Modules

Module	Description
prelude	Convenient glob-import of the most common types
preprocessing	Scaling, encoding, normalization, imputation, binarization, polynomial features, feature hashing, auto-type detection
pipeline	Pipeline (sequential) and ColumnTransformer (per-column transforms)
feature_selection	VarianceThreshold, SelectKBest with ANOVA F-value scoring
traits	Core Fit, Transform, FitTransform traits and error types
time_series	Lag features, rolling windows, difference, cyclical encoding

Re-exports

pub use crate::prelude::*;

Modules

feature_selection

Feature selection transformers.

pipeline

Pipeline composition utilities.

prelude

Convenient glob import of the most common types.

preprocessing

Data preprocessing transformations.

time_series

Time-series feature engineering.

traits

Core traits and error types for the featrs library.

util

Shared helpers for transformers operating on Float64 columns.