Bug: `PlDataFrame._get_cell()` type inconsistency breaks gt-extras `gt_plt_bar_stack`

[KristijanArmeni]

Prework

• Read and agree to the code of conduct and contributing guidelines.
• If there is already a relevant issue, whether open or closed, comment on the existing thread instead of posting a new issue.

Description

When running the gt_plt_bar_stack example from gt-extras docs with Polars, the gt_plt_bar_stack function throws a TypeError: the truth value of a Series is ambiguous but works correctly with Pandas data frames. While the issue is thrown when using gt-extras, I think it's possibly best addressed as an inconsistency between the _get_cell() implementations for Pandas vs. Polars in great-tables. Full error message and MRE below.

Show error message

File "/Users/kriarm/code/great-tables/great_tables/_export.py", line 452, in save
  html_content = as_raw_html(self)
                 ^^^^^^^^^^^^^^^^^
File "/Users/kriarm/code/great-tables/great_tables/_export.py", line 229, in as_raw_html
  built_table = self._build_data(context="html")
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/kriarm/code/great-tables/great_tables/gt.py", line 319, in _build_data
  built = self._render_formats(context)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/kriarm/code/great-tables/great_tables/gt.py", line 312, in _render_formats
  new_body.render_formats(self._tbl_data, self._formats, context)
File "/Users/kriarm/code/great-tables/great_tables/_gt_data.py", line 189, in render_formats
  result = eval_func(_get_cell(data_tbl, row, col))
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/kriarm/code/gt-extras/gt_extras/plotting.py", line 2030, in <lambda>
  lambda x: _make_bar_stack_svg(
            ^^^^^^^^^^^^^^^^^^^^
File "/Users/kriarm/code/gt-extras/gt_extras/plotting.py", line 1941, in _make_bar_stack_svg
  if not values:
     ^^^^^^
File "/Users/kriarm/project/gtextras_bug/.venv/lib/python3.12/site-packages/polars/series/series.py", line 666, in __bool__
  raise TypeError(msg)
TypeError: the truth value of a Series is ambiguous

Here are some things you might want to try:
- instead of `if s`, use `if not s.is_empty()`
- instead of `s1 and s2`, use `s1 & s2`

Problem: To my checking, when a cell entry is a container data type, PlDataFrame._get_cell() returns a pl.Series, but _make_bar_stack_svg expects list[float] and attempts to evaluate if not values on line 1944, which fails for Polars Series. In contrast, PdDataFrame._get_cell() passes a list[float] which works correctly.

Polars code returns a pl.Series (if cell entry is a container data type):

great-tables/great_tables/_tbl_data.py

Lines 221 to 224 in 140adee

	@_get_cell.register(PlDataFrame)
	def _(data: Any, row: int, column: str) -> Any:
	return data[column][row]

Pandas code returns a list of values:

great-tables/great_tables/_tbl_data.py

Lines 226 to 233 in 140adee

	@_get_cell.register(PdDataFrame)
	def _(data: Any, row: int, col: str) -> Any:
	col_ii = data.columns.get_loc(col)
	if not isinstance(col_ii, int):
	raise ValueError("Column named " + col + " matches multiple columns.")
	return data.iloc[row, col_ii]

Possible fixes:

1. Modify gte.plotting._make_bar_stack_svg to accept pl.Series as values, or
2. Update PlDataFrame._get_cell() upstream in great-tables to return a list instead of a Series (matching Pandas behavior)

I was not entirely sure which approach is best suited. I'm happy to work on a PR to address this issue if that would be helpful!

Reproducible example

This is basically the example from gt-extras docs. If this code is placed into test_bug.py and ran as python -m test_bug it should give the error above. The error should go away if Pandas is used instead of Polars.

import polars as pl
from great_tables import GT
import gt_extras as gte

df = pl.DataFrame({
    "x": ["Example A", "Example B", "Example C"],
    "col": [
        [10, 40, 50],
        [30, 30, 40],
        [50, 20, 30],
    ],
})

gt = (
    GT(df)
    .pipe(
        gte.gt_plt_bar_stack,
        column="col",
        palette=["red", "grey", "black"],
        labels=["Group 1", "Group 2", "Group 3"],
        width=200,
    )
)
gt.save(file="./plt_bar_stack_table_pl.png")

Expected result

The table is rendered and saved to disk.

Development environment

• Operating System: macOS Sequoia 15.5 (M1 Mac)
• great_tables version: 0.18.1.dev13+g140adeea2.d20251006 (also reproduced with stable 0.18.0)
• gt-extras version: 0.0.9.dev28+g704e7b13f (also reproduced with stable 0.0.9)
• Polars: 1.34.0
• Pandas: 2.3.3
• Python: 3.12