Try to_pandas rather than erroring if interchanging to pandas doesn't work?

[MarcoGorelli]

Here's an example of some code which currently raises:

import seaborn as sns
import polars as pl
df = pl.DataFrame({'a': [1,2,3], 'b': [4,5,6], 'c': [[1,2], [4,5], [6,7]]})
sns.catplot(
    data=df,
    x="a",
    y="b",
)

There's a really long error message, but the gist of it is

ValueError: data type List(Int64) not supported by the interchange protocol

Indeed, just doing

pd.api.interchange.from_pandas(df)

would raise the same error

I'd like to suggest that, when converting to pandas, if the interchange protocol fails, there be a to_pandas fallback.

Because that at least works here (for reference, plotly do the same - try interchanging first, and if that fails, check if there's to_pandas, and if so, use that)