fix(python): Fix max_colname_length formatting in glimpse()

[jacksonthall22]

It seems there was an arbitrary limit of 100 characters for the length of the string showing the column name and dtype in glimpse(). If you have long column names and pass a max_colname_length that is too long, you will get a formatting error because the format width specifier for the last line item becomes negative. Specifically, If you have a df with a column named a = 'some_long_col_name_xxxxxxxxx...', then df.glimpse(max_colname_length=b) throws a formatting error for any b > len(a) + len(f"<{_dtype_str_repr(df[a].dtype)}>") (see below). This PR just removes the width specifier on the last line item which was redundant anyway.

def test(max_colname_length):
    df = pl.DataFrame({'a' * 100: [1, 2, 3]})
    df.glimpse(max_colname_length=max_colname_length)

test(95)

Rows: 3
Columns: 1
$ aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa… <i64> 1, 2, 3

test(96)

ValueError                                Traceback (most recent call last)
[~\AppData\Local\Temp\ipykernel_30960\2962826816.py](https://file+.vscode-resource.vscode-cdn.net/c%3A/Users/user/.../~/AppData/Local/Temp/ipykernel_30960/2962826816.py) in ?()
      1 test = pl.DataFrame({'a' * 100: [1, 2, 3]})
----> 2 test.glimpse(max_colname_length=96)

[c:\Users\user\miniforge3\envs\env\Lib\site-packages\polars\dataframe\frame.py](file:///C:/Users/user/miniforge3/envs/env/Lib/site-packages/polars/dataframe/frame.py) in ?(self, max_items_per_column, max_colname_length, return_as_string)
   4338 
   4339         # print individual columns: one row per column
   4340         for col_name, dtype_str, val_str in data:
   4341             output.write(
-> 4342                 f"$ {col_name:<{max_col_name}}"
   4343                 f" {dtype_str:>{max_col_dtype}}"
   4344                 f" {val_str:<{min(len(val_str), max_col_values)}}\n"
   4345             )

ValueError: Sign not allowed in string format specifier

[stinodego]

I had to look into this a bit to fully understand it, but it is indeed a good fix! Now we properly respect the max_items_per_column / max_colname_length parameters.

The test failure is unrelated.

Let me add a test first