[python] TileDB-SOMA-Py should map None, pd.NA, and math.nan in string columns to null values

[johnkerl]

Split out from #2858.

Here is a repro script:
https://gist.github.com/johnkerl/de55bd74b146b19a7915c5aee9914752

Here is a readback script:
https://gist.github.com/johnkerl/20e0ad08701f5913f90be706ecd99b01

Notes:

• The input values are ["", "B cell", "T cell", None, pd.NA, math.nan]
• We write these all as strings, including "None", "<NA>", and "nan"
• I believe (but will seek feedback regarding) we should map None, pd.NA, and math.nan in string data to core nulls in TileDB storage
• For clarity: this issue refers stricly to string columns, not enumeration-of-string columns

Note: this is for Pandas data. For Arrow nulls -- tiledbsoma.Experiment.add_new_dataframe rather than tiledbsoma.io.from_anndata/tiledbsoma.io.from_h5ad -- please see #2858 where I'll create a separate issue to track.