feat(python): add parametric testing support for the Array dtype

[alexander-beedie]

In addition to List (already supported) can now declare/use the Array dtype in parametric tests.

Example

import polars as pl
from hypothesis import given
from polars.testing.parametric import column, create_array_strategy

@given(
    df=dataframes(
        cols=[
            # basic use; declare inner dtype and width
            column("colx", dtype=pl.Array(pl.UInt8, width=3)),
            # use a more sophisticated strategy
            column(
                name="coly", 
                strategy=create_array_strategy(
                    inner_dtype=pl.Array(pl.String, width=2),
                    select_from=["a", "b", "c"],
                    width=2,
            ))]
))
def test_array_stuff(df: pl.DataFrame) -> None:
    print(df)
    
# shape: (6, 2)
# ┌────────────────┬──────────────────────────┐
# │ colx           ┆ coly                     │
# │ ---            ┆ ---                      │
# │ array[u8, 3]   ┆ array[array[str, 2], 2]  │
# ╞════════════════╪══════════════════════════╡
# │ [242, 131, 57] ┆ [["c", "c"], ["c", "a"]] │
# │ [220, 78, 82]  ┆ [["a", "a"], ["b", "a"]] │
# │ [153, 19, 100] ┆ [["a", "c"], ["b", "a"]] │
# │ [154, 195, 82] ┆ [["a", "a"], ["a", "a"]] │
# └────────────────┴──────────────────────────┘