`binary_replace_slice` should not work with `string` types

`binary_replace_slice` can give in invalid output when used with string types. Given that there is `utf8_replace_slice`, I think `binary_replace_slice` should not support string types.

If a user actually wants to play with bytes for string type, they should explicitly cast to binary type and use `binary_replace_slice`.
```java

>>> pc.binary_replace_slice(["hé"], 1, 2, "x")

[
"hx�"
]
>>> pc.binary_replace_slice(["hé"], 1, 2, "x").validate(full=True)
Traceback (most recent call last):
...
ArrowInvalid: Invalid UTF8 sequence at string index 0
```

Ref:

cc: @pitrou  

**Reporter**: [Kshiteej K](https://issues.apache.org/jira/browse/ARROW-18319) / @kshitij12345

_{**Note**: *This issue was originally created as [ARROW-18319](https://issues.apache.org/jira/browse/ARROW-18319). Please see the [migration documentation](https://github.com/apache/arrow/issues/14542) for further details.*}

Start by reproducing the Python examples for binary_replace_slice and utf8_replace_slice, including validate(full=True), and trace how string inputs are handled. Done means binary_replace_slice no longer accepts string types that can produce invalid UTF-8, while callers can explicitly cast to binary and use the operation successfully.

Written by the indexing model from the issue text.