This is new functionality and should be preferred over 2. as this will circumvent the GIL and will be the way we want to support extending polars.

Parallelism and optimizations are managed by the default polars runtime. That runtime will call into the plugin function. The plugin functions are compiled separately.

We can therefore keep polars more lean and maybe add support for a polars-distance, polars-geo, polars-ml, etc. Those can then have specialized expressions and don't have to worry as much for code bloat as they can be optionally installed.

The idea is that you define an expression in another Rust crate with a proc_macro polars_expr.

That macro can have the following attributes:

• output_type -> to define the output type of that expression
• output_type_func -> to define a function that computes the output type based on input types.

Here is an example of a String conversion expression that converts any string to pig latin:

fn pig_latin_str(value: &str, capitalize: bool, output: &mut String) {
    if let Some(first_char) = value.chars().next() {
        if capitalize {
            for c in value.chars().skip(1).map(|char| char.to_uppercase()) {
                write!(output, "{c}").unwrap()
            }
            write!(output, "AY").unwrap()
        } else {
            let offset = first_char.len_utf8();
            write!(output, "{}{}ay", &value[offset..], first_char).unwrap()
        }
    }
}

#[derive(Deserialize)]
struct PigLatinKwargs {
    capitalize: bool,
}

#[polars_expr(output_type=Utf8)]
fn pig_latinnify(inputs: &[Series], kwargs: PigLatinKwargs) -> PolarsResult<Series> {
    let ca = inputs[0].utf8()?;
    let out: Utf8Chunked =
        ca.apply_to_buffer(|value, output| pig_latin_str(value, kwargs.capitalize, output));
    Ok(out.into_series())
}

On the python side this expression can then be registered under a namespace:

import polars as pl
from polars.utils.udfs import _get_shared_lib_location

lib = _get_shared_lib_location(__file__)


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

    def pig_latinnify(self, capatilize: bool = False) -> pl.Expr:
        return self._expr._register_plugin(
            lib=lib,
            symbol="pig_latinnify",
            is_elementwise=True,
            kwargs={"capitalize": capatilize}
        )

Compile/ship and then it is ready to use:

import polars as pl
import expression_lib

df = pl.DataFrame({
    "names": ["Richard", "Alice", "Bob"],
})


out = df.with_columns(
   pig_latin = pl.col("names").language.pig_latinnify()
)

See the full example in [example/derive_expression]: https://github.com/pola-rs/pyo3-polars/tree/main/example/derive_expression

See the example directory for a concrete example. Here we send a polars DataFrame to rust and then compute a jaccard similarity in parallel using rayon and rust hash sets.

$ cd example && make install $ venv/bin/python run.py

This will output:

shape: (2, 2)
┌───────────┬───────────────┐
│ list_a    ┆ list_b        │
│ ---       ┆ ---           │
│ list[i64] ┆ list[i64]     │
╞═══════════╪═══════════════╡
│ [1, 2, 3] ┆ [1, 2, ... 8] │
│ [5, 5]    ┆ [5, 1, 1]     │
└───────────┴───────────────┘
shape: (2, 1)
┌─────────┐
│ jaccard │
│ ---     │
│ f64     │
╞═════════╡
│ 0.75    │
│ 0.5     │
└─────────┘

$ make install-release

This crate offers a PySeries and a PyDataFrame which are simple wrapper around Series and DataFrame. The advantage of these wrappers is that they can be converted to and from python as they implement FromPyObject and IntoPy.