str.slice() does not respect codepoints for negative offsets

[mcrumiller]

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

s = pl.Series("你好世界")
s.str.n_chars()    # 4
s.str.len_bytes()  # 12 (3 bytes each)

# positive n
s.str.slice(1)   # "好世界"  - ok
s.str.slice(2)   #   "世界"  - ok
s.str.slice(3)   #     "界"  - ok
s.str.slice(4)   #       ""  - ok

# negative n
s.str.slice(-1)  #       ""  - incorrect
s.str.slice(-2)  #       ""  - incorrect
s.str.slice(-3)  #  "好世界"  - ok
s.str.slice(-4)  # "你好世界" - ok

Log output

No response

Issue description

When str.slice is provided negative indices, it does not properly extract the last n code points.

Expected behavior

s.str.slice(-1)  # "界"
s.str.slice(-2)  # "世界"

Installed versions

main