I have a simple pl.DataFrame with a number of columns that only contain boolean values.

import polars as pl

df = pl.DataFrame(
    {"s1": [True, True, False], "s2": [False, True, True], "s3": [False, False, False]}
)

shape: (3, 3)
┌───────┬───────┬───────┐
│ s1    ┆ s2    ┆ s3    │
│ ---   ┆ ---   ┆ ---   │
│ bool  ┆ bool  ┆ bool  │
╞═══════╪═══════╪═══════╡
│ true  ┆ false ┆ false │
│ true  ┆ true  ┆ false │
│ false ┆ true  ┆ false │
└───────┴───────┴───────┘

I need to add another column that contains lists of varying length. A list in any individual row should contain the column name where the values of the columns S1, s2, and s3 have a True value.

Here's what I am actually looking for:

shape: (3, 4)
┌───────┬───────┬───────┬──────────────┐
│ s1    ┆ s2    ┆ s3    │ list         │
│ ---   ┆ ---   ┆ ---   │ ---          │
│ bool  ┆ bool  ┆ bool  │ list[str]    │
╞═══════╪═══════╪═══════╡══════════════╡
│ true  ┆ false ┆ false │ ["s1"]       │
│ true  ┆ true  ┆ false │ ["s1", "s2"] │
│ false ┆ true  ┆ false │ ["s2"]       │
└───────┴───────┴───────┴──────────────┘