List(Categorical/Enum).contains(Categorical/Enum) is broken since 0.20.2

[Object905]

Checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of Polars.

Reproducible example

import polars as pl

# pl.enable_string_cache()
# exception with both - global string cache and without

df = pl.DataFrame([(["a", "b"], "c"), (["a", "b"], "a")], schema={"li": pl.List(pl.Categorical), "x": pl.Categorical})

print(df.select(pl.col("li").list.contains(pl.col("x").cast(pl.Utf8))))
# ok

print(df.select(pl.col("li").list.contains(pl.col("x"))))
# polars.exceptions.InvalidOperationError: is_in operation not supported for dtypes `cat` and `list[cat]`


enum = pl.Enum(["a", "b", "c"])

df = pl.DataFrame([(["a", "b"], "c"), (["a", "b"], "a")], schema={"li": pl.List(enum), "x": pl.Categorical})


print(df.select(pl.col("li").list.contains(pl.col("x").cast(pl.Utf8))))
# ok

print(df.select(pl.col("li").list.contains(pl.col("x"))))
# polars.exceptions.InvalidOperationError: is_in operation not supported for dtypes `cat` and `list[enum]`

Log output

No response

Issue description

Broken for Categorical-Categorical, Enum-Enum and Categorical-Enum.
Both with and without pl.enable_string_cache.

Expected behavior

To work,, like in 0.20.2

Installed versions

--------Version info---------
Polars:               0.20.9
Index type:           UInt32
Platform:             Linux-6.7.4-arch1-1-x86_64-with-glibc2.39
Python:               3.12.1 (main, Feb  5 2024, 17:00:58) [GCC 13.2.1 20230801]

----Optional dependencies----
adbc_driver_manager:  <not installed>
cloudpickle:          3.0.0
connectorx:           <not installed>
deltalake:            <not installed>
fsspec:               <not installed>
gevent:               24.2.1
hvplot:               <not installed>
matplotlib:           <not installed>
numpy:                1.26.4
openpyxl:             <not installed>
pandas:               2.2.0
pyarrow:              15.0.0
pydantic:             2.6.1
pyiceberg:            <not installed>
pyxlsb:               <not installed>
sqlalchemy:           2.0.27
xlsx2csv:             0.8.2
xlsxwriter:           3.1.9

[mcrumiller]

There are a few missing is_in implementations:

• String is in Categorical
• String is in Enum

also fails, but perhaps should fail:

• Categorical is in Enum
• Enum is Categorical

I don't know if those last two should be implemented. Both Categorical and Enums should coerce strings into their revmap representation, but if your data is already a Categorical and/or Enum, then it means that you have explicitly placed them into particular sets, and those sets are not always comparable.

I do think we could benefit from a slightly more informative error message than, for example:

polars.exceptions.InvalidOperationError: `is_in` cannot check for Enum(Some(local), Physical) values in String data

e.g. explain why it cannot check. Also the enum innards are probably not very useful in the error message.