[Python] BUG: `pyarrow.compute.utf8_slice_codeunits` allows for float start, stop arguments

### Describe the bug, including details regarding any error messages, version, and platform.

```
In [3]: pa.__version__
Out[3]: '11.0.0'

In [4]: arr = pa.array(["abc"])

In [5]: pa.compute.utf8_slice_codeunits(arr, start=1.2, stop=2.2)
Out[5]:

[
"b"
]
```

I would expect this to raise a `ValueError` as the documentation says these arguments should be ints

Related:

* https://github.com/apache/arrow/issues/34929

### Component(s)

Python

Start by reproducing the reported call to pyarrow.compute.utf8_slice_codeunits with float start and stop arguments, then trace the Python entry point and its argument validation. Check the existing documentation and tests for these parameters; done means non-integer arguments raise ValueError while integer slicing remains unchanged.

Written by the indexing model from the issue text.