feat: add `tab_spanner_delim()` method

[jrycw]

Fix issue: #582.

Hello team:

This PR sets up a scaffold for tab_spanner_delim().

Basic Logic:

1. Use str.split or str.rsplit to split column names, corresponding to delimi= and limit= parameters.
2. Iterate through selected columns, splitting each name and recording its level and root_col (original column).
3. Process the recorded info:
   • If level != 0, call tab_spanner().
   • If level == 0, call cols_label().

Example:

import polars as pl
from great_tables import GT

data = {
    "province.NL_ZH.pop": [1, 2, 3],
    "province.NL_ZH.gdp": [4, 5, 6],
    "province.NL_NH.pop": [7, 8, 9],
    "province.NL_NH.gdp": [10, 11, 12],
}

gt = GT(pl.DataFrame(data))
gt.tab_spanner_delim()

gt.tab_spanner_delim(limit=1)

gt.tab_spanner_delim(split="first", limit=1)

Remaining Tasks:

• [ ] The reverse= logic is not implemented—need the team's help to clarify the expected behavior.
• [ ] Docs are ported from {gt} but need refinement.
• [ ] Tests are missing since the logic isn't fully settled yet.