`dy.Schema.cast` shouldn't cast valid dtype.

[gab23r]

In the following snippet, I would expect the cast call to be a no-op as pl.Int32 is already a valid dy.Integer.

This is related to this issue #314

import dataframely as dy
import polars as pl

class S(dy.Schema):
    a = dy.Integer()

df = pl.DataFrame({"a": range(10)}, schema=({"a": pl.Int32}))

# `pl.Int32` is a valid dtypes for `dy.Integer`
S.a.validate_dtype(df.schema["a"])

# but it got casted to `pl.Int64`
S.cast(df).schema
# Schema([('a', Int64)])

[AndreasAlbertQC]

Good point, IMO mostly an issue with dy.Integer, see #315 (comment)