polars_expr_preprocessing.py

import polars as pl

@pl.api.register_expr_namespace("preprocessing")
class Preprocessing:
    def __init__(self, expr: pl.Expr):
        self._expr = expr

    def scale(self) -> pl.Expr:
        return (self._expr - self._expr.min()) / (self._expr.max() - self._expr.min())
    
    def normalize(self) -> pl.Expr:
        return (self._expr - self._expr.mean()) / self._expr.std()

pl.DataFrame(data=[1, 2, 3]).select(
    [
        pl.all().preprocessing.scale().alias("scaled"),
        pl.all().preprocessing.normalize().alias("normalized"),
    ]
)

# shape: (3, 1)   shape: (3, 2)
# ┌──────────┐    ┌────────┬────────────┐
# │ column_0 │    │ scaled ┆ normalized │
# │ ---      │    │ ---    ┆ ---        │
# │ i64      │    │ f64    ┆ f64        │
# ╞══════════╡ >> ╞════════╪════════════╡
# │ 1        │    │ 0.0    ┆ -1.0       │
# │ 2        │    │ 0.5    ┆ 0.0        │
# │ 3        │    │ 1.0    ┆ 1.0        │
# └──────────┘    └────────┴────────────┘